From 36decab9095d7a01b642139e6cc3f8c48bc85f7a Mon Sep 17 00:00:00 2001
From: Morten Hjorth-Jensen <morten.hjorth-jensen@fys.uio.no>
Date: Sun, 12 Nov 2023 18:58:36 +0100
Subject: [PATCH] update week 46

---
 doc/pub/week46/html/._week46-bs000.html       |  309 +-
 doc/pub/week46/html/._week46-bs001.html       |  353 +-
 doc/pub/week46/html/._week46-bs002.html       |  342 +-
 doc/pub/week46/html/._week46-bs003.html       |  410 +-
 doc/pub/week46/html/._week46-bs004.html       |  331 +-
 doc/pub/week46/html/._week46-bs005.html       |  344 +-
 doc/pub/week46/html/._week46-bs006.html       |  333 +-
 doc/pub/week46/html/._week46-bs007.html       |  328 +-
 doc/pub/week46/html/._week46-bs008.html       |  331 +-
 doc/pub/week46/html/._week46-bs009.html       |  421 +-
 doc/pub/week46/html/._week46-bs010.html       |  346 +-
 doc/pub/week46/html/._week46-bs011.html       |  322 +-
 doc/pub/week46/html/._week46-bs012.html       |  376 +-
 doc/pub/week46/html/._week46-bs013.html       |  360 +-
 doc/pub/week46/html/._week46-bs014.html       |  359 +-
 doc/pub/week46/html/._week46-bs015.html       |  350 +-
 doc/pub/week46/html/._week46-bs016.html       |  334 +-
 doc/pub/week46/html/._week46-bs017.html       |  347 +-
 doc/pub/week46/html/._week46-bs018.html       |  364 +-
 doc/pub/week46/html/._week46-bs019.html       |  397 +-
 doc/pub/week46/html/._week46-bs020.html       |  386 +-
 doc/pub/week46/html/._week46-bs021.html       |  369 +-
 doc/pub/week46/html/._week46-bs022.html       |  366 +-
 doc/pub/week46/html/._week46-bs023.html       |  341 +-
 doc/pub/week46/html/._week46-bs024.html       |  534 +-
 doc/pub/week46/html/._week46-bs025.html       |  340 +-
 doc/pub/week46/html/._week46-bs026.html       |  350 +-
 doc/pub/week46/html/._week46-bs027.html       |  405 +-
 doc/pub/week46/html/._week46-bs028.html       |  350 +-
 doc/pub/week46/html/._week46-bs029.html       |  328 +-
 doc/pub/week46/html/week46-bs.html            |  309 +-
 doc/pub/week46/html/week46-reveal.html        | 3579 +++++++-----
 doc/pub/week46/html/week46-solarized.html     | 3487 ++++++++----
 doc/pub/week46/html/week46.html               | 3487 ++++++++----
 doc/pub/week46/ipynb/ipynb-week46-src.tar.gz  |  Bin 192 -> 294343 bytes
 doc/pub/week46/ipynb/week46.ipynb             | 4858 ++++++++++-------
 doc/src/week46/DataFiles/bank.csv             | 1372 +++++
 doc/src/week46/DataFiles/cancer.dot           |   57 +
 doc/src/week46/DataFiles/cancer.png           |  Bin 0 -> 221561 bytes
 doc/src/week46/DataFiles/ensembleoverview.png |  Bin 0 -> 81932 bytes
 doc/src/week46/DataFiles/ride.csv             |   15 +
 doc/src/week46/DataFiles/ride.dot             |   13 +
 doc/src/week46/DataFiles/rideclass.csv        |   15 +
 doc/src/week46/DataFiles/zoo.csv              |  101 +
 doc/src/week46/backup2022.do.txt              | 1171 ++++
 doc/src/week46/week46.do.txt                  | 2683 +++++----
 46 files changed, 22068 insertions(+), 9905 deletions(-)
 create mode 100644 doc/src/week46/DataFiles/bank.csv
 create mode 100644 doc/src/week46/DataFiles/cancer.dot
 create mode 100644 doc/src/week46/DataFiles/cancer.png
 create mode 100644 doc/src/week46/DataFiles/ensembleoverview.png
 create mode 100644 doc/src/week46/DataFiles/ride.csv
 create mode 100644 doc/src/week46/DataFiles/ride.dot
 create mode 100644 doc/src/week46/DataFiles/rideclass.csv
 create mode 100644 doc/src/week46/DataFiles/zoo.csv
 create mode 100644 doc/src/week46/backup2022.do.txt

diff --git a/doc/pub/week46/html/._week46-bs000.html b/doc/pub/week46/html/._week46-bs000.html
index ced4225cc..39a24a971 100644
--- a/doc/pub/week46/html/._week46-bs000.html
+++ b/doc/pub/week46/html/._week46-bs000.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -174,7 +331,7 @@ MathJax.Hub.Config({
 <!-- ------------------- main content ---------------------- -->
 <div class="jumbotron">
 <center>
-<h1>Week 46: Support Vector Machines and Project 3. </h1>
+<h1>Week 46: Decision Trees, Ensemble methods  and Random Forests</h1>
 </center>  <!-- document title -->
 
 <!-- author(s): Morten Hjorth-Jensen -->
@@ -190,7 +347,7 @@ MathJax.Hub.Config({
 </center>
 <br>
 <center>
-<h4>Nov 19, 2022</h4>
+<h4>Week 46, November 13-17</h4>
 </center> <!-- date -->
 <br>
 
@@ -215,7 +372,7 @@ MathJax.Hub.Config({
   <li><a href="._week46-bs008.html">9</a></li>
   <li><a href="._week46-bs009.html">10</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs001.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
@@ -229,7 +386,7 @@ MathJax.Hub.Config({
 </footer>
 -->
 <center style="font-size:80%">
-<!-- copyright --> &copy; 1999-2022, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
+<!-- copyright --> &copy; 1999-2023, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
 </center>
 </body>
 </html>
diff --git a/doc/pub/week46/html/._week46-bs001.html b/doc/pub/week46/html/._week46-bs001.html
index 22f5149bb..2eb5a8a1c 100644
--- a/doc/pub/week46/html/._week46-bs001.html
+++ b/doc/pub/week46/html/._week46-bs001.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,38 +329,32 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0001"></a>
 <!-- !split -->
-<h2 id="overview-of-week-46" class="anchor">Overview of week 46 </h2>
-
-<ul>
-<li> <b>Thursday</b>: Support Vector Machines, discussion of project 3</li>
-<ul>
-  <li> <a href="https://youtu.be/F3CkH-opbdY" target="_self">Video of lecture</a></li>
-</ul>
-<li> <b>Friday</b>: Support vector machines and start Principal Component Analysis</li>
-<ul>
-  <li> <a href="https://youtu.be/BbupEEvMXtg" target="_self">Video of lecture</a></li>
-</ul>
-</ul>
-<div class="panel panel-default">
-<div class="panel-body">
-<!-- subsequent paragraphs come in larger fonts, so start with a paragraph -->
-<p>Reading recommendations:</p>
-<ol>
- <li> See lecture notes for week 46 at <a href="https://compphysics.github.io/MachineLearning/doc/web/course.html." target="_self"><tt>https://compphysics.github.io/MachineLearning/doc/web/course.html.</tt></a></li>
- <li> Hastie et al chapter 12</li>
- <li> Bishop chapter 7.1 and 7.2</li>
-</ol>
-</div>
-</div>
-
+<h2 id="plan-for-week-46" class="anchor">Plan for week 46 </h2>
 
 <div class="panel panel-default">
 <div class="panel-body">
 <!-- subsequent paragraphs come in larger fonts, so start with a paragraph -->
-<ol>
-<li> <a href="https://www.youtube.com/watch?v=efR1C6CvhmE&ab_channel=StatQuestwithJoshStarmer" target="_self">Overview video on Support Vector Machines</a></li>
-<li> See also <a href="https://www.youtube.com/watch?v=N1vOgolbjSc&ab_channel=AliceZhao" target="_self">this video</a>.</li>
-</ol>
+<ul>
+  <li> Work and Discussion of project 2</li>
+  <li> Discussion of project 3 as well</li>
+</ul>
+</div>
+</div>
+  
+
+<div class="panel panel-default">
+<div class="panel-body">
+<!-- subsequent paragraphs come in larger fonts, so start with a paragraph -->
+<ul>
+  <li> Thursday: Basics of decision trees, classification and regression algorithms and ensemble models</li> 
+  <li> Readings and Videos:</li>
+<ul>
+    <li> These lecture notes</li>
+    <li> <a href="https://youtu.be/" target="_self">Video of lecture to be added</a></li>
+    <li> <a href="https://www.youtube.com/watch?v=RmajweUFKvM&ab_channel=Simplilearn" target="_self">Video on Decision trees</a></li>
+    <li> Decision Trees: Geron's chapter 6 covers decision trees while ensemble models, voting and bagging are discussed in chapter 7. See also lecture from <a href="https://www.uio.no/studier/emner/matnat/math/STK-IN4300/h20/slides/lecture_7.pdf" target="_self">STK-IN4300, lecture 7</a>. Chapter 9.2 of Hastie et al contains also a good discussion.</li>
+</ul>
+</ul>
 </div>
 </div>
 
@@ -224,7 +375,7 @@ MathJax.Hub.Config({
   <li><a href="._week46-bs009.html">10</a></li>
   <li><a href="._week46-bs010.html">11</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs002.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs002.html b/doc/pub/week46/html/._week46-bs002.html
index e67372972..01c341185 100644
--- a/doc/pub/week46/html/._week46-bs002.html
+++ b/doc/pub/week46/html/._week46-bs002.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,34 +329,31 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0002"></a>
 <!-- !split -->
-<h2 id="support-vector-machines-overarching-aims" class="anchor">Support Vector Machines, overarching aims  </h2>
+<h2 id="decision-trees-overarching-aims" class="anchor">Decision trees, overarching aims  </h2>
 
-<p>A Support Vector Machine (SVM) is a very powerful and versatile
-Machine Learning method, capable of performing linear or nonlinear
-classification, regression, and even outlier detection. It is one of
-the most popular models in Machine Learning, and anyone interested in
-Machine Learning should have it in their toolbox. SVMs are
-particularly well suited for classification of complex but small-sized or
-medium-sized datasets.  
+<p>We start here with the most basic algorithm, the so-called decision
+tree. With this basic algorithm we can in turn build more complex
+networks, spanning from homogeneous and heterogenous forests (bagging,
+random forests and more) to one of the most popular supervised
+algorithms nowadays, the extreme gradient boosting, or just
+XGBoost. But let us start with the simplest possible ingredient.
 </p>
 
-<p>The case with two well-separated classes only can be understood in an
-intuitive way in terms of lines in a two-dimensional space separating
-the two classes (see figure below).
+<p>Decision trees are supervised learning algorithms used for both,
+classification and regression tasks.
 </p>
 
-<p>The basic mathematics behind the SVM is however less familiar to most of us. 
-It relies on the definition of hyperplanes and the
-definition of a <b>margin</b> which separates classes (in case of
-classification problems) of variables. It is also used for regression
-problems.
+<p>The main idea of decision trees
+is to find those descriptive features which contain the most
+<b>information</b> regarding the target feature and then split the dataset
+along the values of these features such that the target feature values
+for the resulting underlying datasets are as pure as possible.
 </p>
 
-<p>With SVMs we distinguish between hard margin and soft margins. The
-latter introduces a so-called softening parameter to be discussed
-below.  We distinguish also between linear and non-linear
-approaches. The latter are the most frequent ones since it is rather
-unlikely that we can separate classes easily by say straight lines.
+<p>The descriptive features which reproduce best the target/output features are normally  said
+to be the most informative ones. The process of finding the <b>most
+informative</b> feature is done until we accomplish a stopping criteria
+where we then finally end up in so called <b>leaf nodes</b>. 
 </p>
 
 <p>
@@ -219,7 +373,7 @@ unlikely that we can separate classes easily by say straight lines.
   <li><a href="._week46-bs010.html">11</a></li>
   <li><a href="._week46-bs011.html">12</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs003.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs003.html b/doc/pub/week46/html/._week46-bs003.html
index 360fadfa5..439dec21c 100644
--- a/doc/pub/week46/html/._week46-bs003.html
+++ b/doc/pub/week46/html/._week46-bs003.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,109 +329,20 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0003"></a>
 <!-- !split -->
-<h2 id="hyperplanes-and-all-that" class="anchor">Hyperplanes and all that </h2>
+<h2 id="basics-of-a-tree" class="anchor">Basics of a tree </h2>
 
-<p>The theory behind support vector machines (SVM hereafter) is based on
-the mathematical description of so-called hyperplanes. Let us start
-with a two-dimensional case. This will also allow us to introduce our
-first SVM examples. These will be tailored to the case of two specific
-classes, as displayed in the figure here based on the usage of the petal data.
+<p>A decision tree is typically divided into a <b>root node</b>, the <b>interior nodes</b>,
+and the final <b>leaf nodes</b> or just <b>leaves</b>. These entities are then connected by so-called <b>branches</b>.
 </p>
 
-<p>We assume here that our data set can be well separated into two
-domains, where a straight line does the job in the separating the two
-classes. Here the two classes are represented by either squares or
-circles.
+<p>The leaf nodes
+contain the predictions we will make for new query instances presented
+to our trained model. This is possible since the model has 
+learned the underlying structure of the training data and hence can,
+given some assumptions, make predictions about the target feature value
+(class) of unseen query instances.
 </p>
 
-<!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #f8f8f8">
-  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn</span> <span style="color: #008000; font-weight: bold">import</span> datasets
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC, LinearSVC
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.linear_model</span> <span style="color: #008000; font-weight: bold">import</span> SGDClassifier
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> StandardScaler
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib</span>
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;axes.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">14</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;xtick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;ytick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
-
-
-iris <span style="color: #666666">=</span> datasets<span style="color: #666666">.</span>load_iris()
-X <span style="color: #666666">=</span> iris[<span style="color: #BA2121">&quot;data&quot;</span>][:, (<span style="color: #666666">2</span>, <span style="color: #666666">3</span>)]  <span style="color: #408080; font-style: italic"># petal length, petal width</span>
-y <span style="color: #666666">=</span> iris[<span style="color: #BA2121">&quot;target&quot;</span>]
-
-setosa_or_versicolor <span style="color: #666666">=</span> (y <span style="color: #666666">==</span> <span style="color: #666666">0</span>) <span style="color: #666666">|</span> (y <span style="color: #666666">==</span> <span style="color: #666666">1</span>)
-X <span style="color: #666666">=</span> X[setosa_or_versicolor]
-y <span style="color: #666666">=</span> y[setosa_or_versicolor]
-
-
-
-C <span style="color: #666666">=</span> <span style="color: #666666">5</span>
-alpha <span style="color: #666666">=</span> <span style="color: #666666">1</span> <span style="color: #666666">/</span> (C <span style="color: #666666">*</span> <span style="color: #008000">len</span>(X))
-
-lin_clf <span style="color: #666666">=</span> LinearSVC(loss<span style="color: #666666">=</span><span style="color: #BA2121">&quot;hinge&quot;</span>, C<span style="color: #666666">=</span>C, random_state<span style="color: #666666">=42</span>)
-svm_clf <span style="color: #666666">=</span> SVC(kernel<span style="color: #666666">=</span><span style="color: #BA2121">&quot;linear&quot;</span>, C<span style="color: #666666">=</span>C)
-sgd_clf <span style="color: #666666">=</span> SGDClassifier(loss<span style="color: #666666">=</span><span style="color: #BA2121">&quot;hinge&quot;</span>, learning_rate<span style="color: #666666">=</span><span style="color: #BA2121">&quot;constant&quot;</span>, eta0<span style="color: #666666">=0.001</span>, alpha<span style="color: #666666">=</span>alpha,
-                        max_iter<span style="color: #666666">=100000</span>, random_state<span style="color: #666666">=42</span>)
-
-scaler <span style="color: #666666">=</span> StandardScaler()
-X_scaled <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>fit_transform(X)
-
-lin_clf<span style="color: #666666">.</span>fit(X_scaled, y)
-svm_clf<span style="color: #666666">.</span>fit(X_scaled, y)
-sgd_clf<span style="color: #666666">.</span>fit(X_scaled, y)
-
-<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;LinearSVC:                   &quot;</span>, lin_clf<span style="color: #666666">.</span>intercept_, lin_clf<span style="color: #666666">.</span>coef_)
-<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;SVC:                         &quot;</span>, svm_clf<span style="color: #666666">.</span>intercept_, svm_clf<span style="color: #666666">.</span>coef_)
-<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;SGDClassifier(alpha=</span><span style="color: #BB6688; font-weight: bold">{:.5f}</span><span style="color: #BA2121">):&quot;</span><span style="color: #666666">.</span>format(sgd_clf<span style="color: #666666">.</span>alpha), sgd_clf<span style="color: #666666">.</span>intercept_, sgd_clf<span style="color: #666666">.</span>coef_)
-
-<span style="color: #408080; font-style: italic"># Compute the slope and bias of each decision boundary</span>
-w1 <span style="color: #666666">=</span> <span style="color: #666666">-</span>lin_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">0</span>]<span style="color: #666666">/</span>lin_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-b1 <span style="color: #666666">=</span> <span style="color: #666666">-</span>lin_clf<span style="color: #666666">.</span>intercept_[<span style="color: #666666">0</span>]<span style="color: #666666">/</span>lin_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-w2 <span style="color: #666666">=</span> <span style="color: #666666">-</span>svm_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">0</span>]<span style="color: #666666">/</span>svm_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-b2 <span style="color: #666666">=</span> <span style="color: #666666">-</span>svm_clf<span style="color: #666666">.</span>intercept_[<span style="color: #666666">0</span>]<span style="color: #666666">/</span>svm_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-w3 <span style="color: #666666">=</span> <span style="color: #666666">-</span>sgd_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">0</span>]<span style="color: #666666">/</span>sgd_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-b3 <span style="color: #666666">=</span> <span style="color: #666666">-</span>sgd_clf<span style="color: #666666">.</span>intercept_[<span style="color: #666666">0</span>]<span style="color: #666666">/</span>sgd_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-
-<span style="color: #408080; font-style: italic"># Transform the decision boundary lines back to the original scale</span>
-line1 <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>inverse_transform([[<span style="color: #666666">-10</span>, <span style="color: #666666">-10</span> <span style="color: #666666">*</span> w1 <span style="color: #666666">+</span> b1], [<span style="color: #666666">10</span>, <span style="color: #666666">10</span> <span style="color: #666666">*</span> w1 <span style="color: #666666">+</span> b1]])
-line2 <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>inverse_transform([[<span style="color: #666666">-10</span>, <span style="color: #666666">-10</span> <span style="color: #666666">*</span> w2 <span style="color: #666666">+</span> b2], [<span style="color: #666666">10</span>, <span style="color: #666666">10</span> <span style="color: #666666">*</span> w2 <span style="color: #666666">+</span> b2]])
-line3 <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>inverse_transform([[<span style="color: #666666">-10</span>, <span style="color: #666666">-10</span> <span style="color: #666666">*</span> w3 <span style="color: #666666">+</span> b3], [<span style="color: #666666">10</span>, <span style="color: #666666">10</span> <span style="color: #666666">*</span> w3 <span style="color: #666666">+</span> b3]])
-
-<span style="color: #408080; font-style: italic"># Plot all three decision boundaries</span>
-plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">4</span>))
-plt<span style="color: #666666">.</span>plot(line1[:, <span style="color: #666666">0</span>], line1[:, <span style="color: #666666">1</span>], <span style="color: #BA2121">&quot;k:&quot;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;LinearSVC&quot;</span>)
-plt<span style="color: #666666">.</span>plot(line2[:, <span style="color: #666666">0</span>], line2[:, <span style="color: #666666">1</span>], <span style="color: #BA2121">&quot;b--&quot;</span>, linewidth<span style="color: #666666">=2</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;SVC&quot;</span>)
-plt<span style="color: #666666">.</span>plot(line3[:, <span style="color: #666666">0</span>], line3[:, <span style="color: #666666">1</span>], <span style="color: #BA2121">&quot;r-&quot;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;SGDClassifier&quot;</span>)
-plt<span style="color: #666666">.</span>plot(X[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==1</span>], X[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==1</span>], <span style="color: #BA2121">&quot;bs&quot;</span>) <span style="color: #408080; font-style: italic"># label=&quot;Iris-Versicolor&quot;</span>
-plt<span style="color: #666666">.</span>plot(X[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==0</span>], X[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==0</span>], <span style="color: #BA2121">&quot;yo&quot;</span>) <span style="color: #408080; font-style: italic"># label=&quot;Iris-Setosa&quot;</span>
-plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">&quot;Petal length&quot;</span>, fontsize<span style="color: #666666">=14</span>)
-plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">&quot;Petal width&quot;</span>, fontsize<span style="color: #666666">=14</span>)
-plt<span style="color: #666666">.</span>legend(loc<span style="color: #666666">=</span><span style="color: #BA2121">&quot;upper center&quot;</span>, fontsize<span style="color: #666666">=14</span>)
-plt<span style="color: #666666">.</span>axis([<span style="color: #666666">0</span>, <span style="color: #666666">5.5</span>, <span style="color: #666666">0</span>, <span style="color: #666666">2</span>])
-
-plt<span style="color: #666666">.</span>show()
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
-
-
 <p>
 <!-- navigation buttons at the bottom of the page -->
 <ul class="pagination">
@@ -293,7 +361,7 @@ plt<span style="color: #666666">.</span>show()
   <li><a href="._week46-bs011.html">12</a></li>
   <li><a href="._week46-bs012.html">13</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs004.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs004.html b/doc/pub/week46/html/._week46-bs004.html
index 0497c9d79..5fb656ca1 100644
--- a/doc/pub/week46/html/._week46-bs004.html
+++ b/doc/pub/week46/html/._week46-bs004.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,33 +329,11 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0004"></a>
 <!-- !split -->
-<h2 id="what-is-a-hyperplane" class="anchor">What is a hyperplane? </h2>
+<h2 id="a-sketch-of-a-tree-regression-problem" class="anchor">A Sketch of a Tree, Regression problem  </h2>
 
-<p>The aim of the SVM algorithm is to find a hyperplane in a
-\( p \)-dimensional space, where \( p \) is the number of features that
-distinctly classifies the data points.
-</p>
-
-<p>In a \( p \)-dimensional space, a hyperplane is what we call an affine subspace of dimension of \( p-1 \).
-As an example, in two dimension, a hyperplane is simply as straight line while in three dimensions it is 
-a two-dimensional subspace, or stated simply, a plane. 
-</p>
-
-<p>In two dimensions, with the variables \( x_1 \) and \( x_2 \), the hyperplane is defined as</p>
-$$
-b+w_1x_1+w_2x_2=0,
-$$
-
-<p>where \( b \) is the intercept and \( w_1 \) and \( w_2 \) define the elements of a vector orthogonal to the line 
-\( b+w_1x_1+w_2x_2=0 \). 
-In two dimensions we define the vectors \( \boldsymbol{x} =[x1,x2] \) and \( \boldsymbol{w}=[w1,w2] \). 
-We can then rewrite the above equation as 
-</p>
-
-$$
-\boldsymbol{x}^T\boldsymbol{w}+b=0.
-$$
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_self">See handwritten notes November 3</a>
 
+<!-- FIGURE: [DataFiles/Regsimpletree.png, width=600 frac=0.8] -->
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -219,7 +354,7 @@ $$
   <li><a href="._week46-bs012.html">13</a></li>
   <li><a href="._week46-bs013.html">14</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs005.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs005.html b/doc/pub/week46/html/._week46-bs005.html
index c906c999e..7cf82a557 100644
--- a/doc/pub/week46/html/._week46-bs005.html
+++ b/doc/pub/week46/html/._week46-bs005.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,45 +329,10 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0005"></a>
 <!-- !split -->
-<h2 id="a-p-dimensional-space-of-features" class="anchor">A \( p \)-dimensional space of features </h2>
+<h2 id="a-sketch-of-a-tree-classification-problem" class="anchor">A Sketch of a Tree, Classification  problem  </h2>
 
-<p>We limit ourselves to two classes of outputs \( y_i \) and assign these classes the values \( y_i = \pm 1 \). 
-In a \( p \)-dimensional space of say \( p \) features we have a hyperplane defines as 
-</p>
-$$
-b+wx_1+w_2x_2+\dots +w_px_p=0.
-$$
-
-<p>If we define a 
-matrix \( \boldsymbol{X}=\left[\boldsymbol{x}_1,\boldsymbol{x}_2,\dots, \boldsymbol{x}_p\right] \)
-of dimension \( n\times p \), where \( n \) represents the observations for each feature and each vector \( x_i \) is a column vector of the matrix \( \boldsymbol{X} \), 
-</p>
-$$
-\boldsymbol{x}_i = \begin{bmatrix} x_{i1} \\ x_{i2} \\ \dots \\ \dots \\ x_{ip} \end{bmatrix}.
-$$
-
-<p>If the above condition is not met for a given vector \( \boldsymbol{x}_i \) we have </p>
-$$
-b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} >0,
-$$
-
-<p>if our output \( y_i=1 \).
-In this case we say that \( \boldsymbol{x}_i \) lies on one of the sides of the hyperplane and if 
-</p>
-$$
-b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} < 0,
-$$
-
-<p>for the class of observations \( y_i=-1 \), 
-then \( \boldsymbol{x}_i \) lies on the other side. 
-</p>
-
-<p>Equivalently, for the two classes of observations we have </p>
-$$
-y_i\left(b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip}\right) > 0. 
-$$
-
-<p>When we try to separate hyperplanes, if it exists, we can use it to construct a natural classifier: a test observation is assigned a given class depending on which side of the hyperplane it is located.</p>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_self">See handwritten notes November 3</a>
+<!-- FIGURE: [DataFiles/Classimpletree.png, width=600 frac=0.8] -->
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -232,7 +354,7 @@ $$
   <li><a href="._week46-bs013.html">14</a></li>
   <li><a href="._week46-bs014.html">15</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs006.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs006.html b/doc/pub/week46/html/._week46-bs006.html
index f8fd2cf94..584e8a6ab 100644
--- a/doc/pub/week46/html/._week46-bs006.html
+++ b/doc/pub/week46/html/._week46-bs006.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -171,30 +328,16 @@ MathJax.Hub.Config({
 <div class="container">
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0006"></a>
-<!-- !split  -->
-<h2 id="the-two-dimensional-case" class="anchor">The two-dimensional case </h2>
+<!-- !split -->
+<h2 id="a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" class="anchor">A typical Decision Tree with its pertinent Jargon, Classification Problem </h2>
 
-<p>Let us try to develop our intuition about SVMs by limiting ourselves to a two-dimensional
-plane.  To separate the two classes of data points, there are many
-possible lines (hyperplanes if you prefer a more strict naming)  
-that could be chosen. Our objective is to find a
-plane that has the maximum margin, i.e the maximum distance between
-data points of both classes. Maximizing the margin distance provides
-some reinforcement so that future data points can be classified with
-more confidence.
-</p>
+<br/><br/>
+<center>
+<p><img src="DataFiles/cancer.png" width="600" align="bottom"></p>
+</center>
+<br/><br/>
 
-<p>What a linear classifier attempts to accomplish is to split the
-feature space into two half spaces by placing a hyperplane between the
-data points.  This hyperplane will be our decision boundary.  All
-points on one side of the plane will belong to class one and all points
-on the other side of the plane will belong to the second class two.
-</p>
-
-<p>Unfortunately there are many ways in which we can place a hyperplane
-to divide the data.  Below is an example of two candidate hyperplanes
-for our data sample.
-</p>
+<p>This tree was produced using the Wisconsin cancer data (discussed here as well, see code examples below) using <b>Scikit-Learn</b>'s decision tree classifier. Here we have used the so-called <b>gini</b> index (see below) to split the various branches.</p>
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -217,7 +360,7 @@ for our data sample.
   <li><a href="._week46-bs014.html">15</a></li>
   <li><a href="._week46-bs015.html">16</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs007.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs007.html b/doc/pub/week46/html/._week46-bs007.html
index bdb1ad7ce..42c8de6ec 100644
--- a/doc/pub/week46/html/._week46-bs007.html
+++ b/doc/pub/week46/html/._week46-bs007.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,22 +329,19 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0007"></a>
 <!-- !split -->
-<h2 id="getting-into-the-details" class="anchor">Getting into the details </h2>
+<h2 id="general-features" class="anchor">General Features </h2>
 
-<p>Let us define the function</p>
-$$
-f(x) = \boldsymbol{w}^T\boldsymbol{x}+b = 0,
-$$
-
-<p>as the function that determines the line \( L \) that separates two classes (our two features), see the figure here. </p>
-
-<p>Any point defined by \( \boldsymbol{x}_i \) and \( \boldsymbol{x}_2 \) on the line \( L \) will satisfy \( \boldsymbol{w}^T(\boldsymbol{x}_1-\boldsymbol{x}_2)=0 \). </p>
-
-<p>The signed distance \( \delta \) from any point defined by a vector \( \boldsymbol{x} \) and a point \( \boldsymbol{x}_0 \) on the line \( L \) is then</p>
-$$
-\delta = \frac{1}{\vert\vert \boldsymbol{w}\vert\vert}(\boldsymbol{w}^T\boldsymbol{x}+b).
-$$
+<p>The overarching approach to decision trees is a top-down approach.</p>
 
+<ul>
+<li> A leaf provides the classification of a given instance.</li>
+<li> A node specifies a test of some attribute of the instance.</li>
+<li> A branch corresponds to a possible values of an attribute.</li>
+<li> An instance is classified by starting at the root node of the tree, testing the attribute specified by this node, then moving down the tree branch corresponding to the value of the attribute in the given example.</li>
+</ul>
+<p>This process is then repeated for the subtree rooted at the new
+node.
+</p>
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -211,7 +365,7 @@ $$
   <li><a href="._week46-bs015.html">16</a></li>
   <li><a href="._week46-bs016.html">17</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs008.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs008.html b/doc/pub/week46/html/._week46-bs008.html
index 8464e209f..6abaa583f 100644
--- a/doc/pub/week46/html/._week46-bs008.html
+++ b/doc/pub/week46/html/._week46-bs008.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,27 +329,19 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0008"></a>
 <!-- !split -->
-<h2 id="first-attempt-at-a-minimization-approach" class="anchor">First attempt at a minimization approach </h2>
+<h2 id="how-do-we-set-it-up" class="anchor">How do we set it up? </h2>
 
-<p>How do we find the parameter \( b \) and the vector \( \boldsymbol{w} \)? What we could
-do is to define a cost function which now contains the set of all
-misclassified points \( M \) and attempt to minimize this function
+<p>In simplified terms, the process of training a decision tree and
+predicting the target features of query instances is as follows:
 </p>
 
-$$
-C(\boldsymbol{w},b) = -\sum_{i\in M} y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b).
-$$
-
-<p>We could now for example define all values \( y_i =1 \) as misclassified in case we have \( \boldsymbol{w}^T\boldsymbol{x}_i+b < 0 \) and the opposite if we have \( y_i=-1 \). Taking the derivatives gives us</p>
-$$
-\frac{\partial C}{\partial b} = -\sum_{i\in M} y_i,
-$$
-
-<p>and </p>
-$$
-\frac{\partial C}{\partial \boldsymbol{w}} = -\sum_{i\in M} y_ix_i.
-$$
-
+<ol>
+<li> Present a dataset containing of a number of training instances characterized by a number of descriptive features and a target feature</li>
+<li> Train the decision tree model by continuously splitting the target feature along the values of the descriptive features using a measure of information gain during the training process</li>
+<li> Grow the tree until we accomplish a stopping criteria create leaf nodes which represent the <em>predictions</em> we want to make for new query instances</li>
+<li> Show query instances to the tree and run down the tree until we arrive at leaf nodes</li>
+</ol>
+<p>Then we are essentially done!</p>
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -217,7 +366,7 @@ $$
   <li><a href="._week46-bs016.html">17</a></li>
   <li><a href="._week46-bs017.html">18</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs009.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs009.html b/doc/pub/week46/html/._week46-bs009.html
index b23b024e4..1afb983f3 100644
--- a/doc/pub/week46/html/._week46-bs009.html
+++ b/doc/pub/week46/html/._week46-bs009.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,19 +329,117 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0009"></a>
 <!-- !split -->
-<h2 id="solving-the-equations" class="anchor">Solving the equations </h2>
+<h2 id="decision-trees-and-regression" class="anchor">Decision trees and Regression  </h2>
 
-<p>We can now use the Newton-Raphson method or different variants of the gradient descent family (from plain gradient descent to various stochastic gradient descent approaches) to solve the equations</p>
-$$
-b \leftarrow b +\eta \frac{\partial C}{\partial b},
-$$
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> PolynomialFeatures
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.linear_model</span> <span style="color: #008000; font-weight: bold">import</span> LinearRegression
 
-<p>and</p>
-$$
-\boldsymbol{w} \leftarrow \boldsymbol{w} +\eta \frac{\partial C}{\partial \boldsymbol{w}},
-$$
+steps<span style="color: #666666">=250</span>
+
+distance<span style="color: #666666">=0</span>
+x<span style="color: #666666">=0</span>
+distance_list<span style="color: #666666">=</span>[]
+steps_list<span style="color: #666666">=</span>[]
+<span style="color: #008000; font-weight: bold">while</span> x<span style="color: #666666">&lt;</span>steps:
+    distance<span style="color: #666666">+=</span>np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>randint(<span style="color: #666666">-1</span>,<span style="color: #666666">2</span>)
+    distance_list<span style="color: #666666">.</span>append(distance)
+    x<span style="color: #666666">+=1</span>
+    steps_list<span style="color: #666666">.</span>append(x)
+plt<span style="color: #666666">.</span>plot(steps_list,distance_list, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;green&#39;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;Random Walk Data&quot;</span>)
+
+steps_list<span style="color: #666666">=</span>np<span style="color: #666666">.</span>asarray(steps_list)
+distance_list<span style="color: #666666">=</span>np<span style="color: #666666">.</span>asarray(distance_list)
+
+X<span style="color: #666666">=</span>steps_list[:,np<span style="color: #666666">.</span>newaxis]
+
+<span style="color: #408080; font-style: italic">#Polynomial fits</span>
+
+<span style="color: #408080; font-style: italic">#Degree 2</span>
+poly_features<span style="color: #666666">=</span>PolynomialFeatures(degree<span style="color: #666666">=2</span>, include_bias<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">False</span>)
+X_poly<span style="color: #666666">=</span>poly_features<span style="color: #666666">.</span>fit_transform(X)
+
+lin_reg<span style="color: #666666">=</span>LinearRegression()
+poly_fit<span style="color: #666666">=</span>lin_reg<span style="color: #666666">.</span>fit(X_poly,distance_list)
+b<span style="color: #666666">=</span>lin_reg<span style="color: #666666">.</span>coef_
+c<span style="color: #666666">=</span>lin_reg<span style="color: #666666">.</span>intercept_
+<span style="color: #008000">print</span> (<span style="color: #BA2121">&quot;2nd degree coefficients:&quot;</span>)
+<span style="color: #008000">print</span> (<span style="color: #BA2121">&quot;zero power: &quot;</span>,c)
+<span style="color: #008000">print</span> (<span style="color: #BA2121">&quot;first power: &quot;</span>, b[<span style="color: #666666">0</span>])
+<span style="color: #008000">print</span> (<span style="color: #BA2121">&quot;second power: &quot;</span>,b[<span style="color: #666666">1</span>])
+
+z <span style="color: #666666">=</span> np<span style="color: #666666">.</span>arange(<span style="color: #666666">0</span>, steps, <span style="color: #666666">.01</span>)
+z_mod<span style="color: #666666">=</span>b[<span style="color: #666666">1</span>]<span style="color: #666666">*</span>z<span style="color: #666666">**2+</span>b[<span style="color: #666666">0</span>]<span style="color: #666666">*</span>z<span style="color: #666666">+</span>c
+
+fit_mod<span style="color: #666666">=</span>b[<span style="color: #666666">1</span>]<span style="color: #666666">*</span>X<span style="color: #666666">**2+</span>b[<span style="color: #666666">0</span>]<span style="color: #666666">*</span>X<span style="color: #666666">+</span>c
+plt<span style="color: #666666">.</span>plot(z, z_mod, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;r&#39;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;2nd Degree Fit&quot;</span>)
+plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">&quot;Polynomial Regression&quot;</span>)
+
+plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">&quot;Steps&quot;</span>)
+plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">&quot;Distance&quot;</span>)
+
+<span style="color: #408080; font-style: italic">#Degree 10</span>
+poly_features10<span style="color: #666666">=</span>PolynomialFeatures(degree<span style="color: #666666">=10</span>, include_bias<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">False</span>)
+X_poly10<span style="color: #666666">=</span>poly_features10<span style="color: #666666">.</span>fit_transform(X)
+
+poly_fit10<span style="color: #666666">=</span>lin_reg<span style="color: #666666">.</span>fit(X_poly10,distance_list)
+
+y_plot<span style="color: #666666">=</span>poly_fit10<span style="color: #666666">.</span>predict(X_poly10)
+plt<span style="color: #666666">.</span>plot(X, y_plot, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;black&#39;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;10th Degree Fit&quot;</span>)
+
+plt<span style="color: #666666">.</span>legend()
+plt<span style="color: #666666">.</span>show()
+
+
+<span style="color: #408080; font-style: italic">#Decision Tree Regression</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeRegressor
+regr_1<span style="color: #666666">=</span>DecisionTreeRegressor(max_depth<span style="color: #666666">=2</span>)
+regr_2<span style="color: #666666">=</span>DecisionTreeRegressor(max_depth<span style="color: #666666">=5</span>)
+regr_3<span style="color: #666666">=</span>DecisionTreeRegressor(max_depth<span style="color: #666666">=7</span>)
+regr_1<span style="color: #666666">.</span>fit(X, distance_list)
+regr_2<span style="color: #666666">.</span>fit(X, distance_list)
+regr_3<span style="color: #666666">.</span>fit(X, distance_list)
+
+X_test <span style="color: #666666">=</span> np<span style="color: #666666">.</span>arange(<span style="color: #666666">0.0</span>, steps, <span style="color: #666666">0.01</span>)[:, np<span style="color: #666666">.</span>newaxis]
+y_1 <span style="color: #666666">=</span> regr_1<span style="color: #666666">.</span>predict(X_test)
+y_2 <span style="color: #666666">=</span> regr_2<span style="color: #666666">.</span>predict(X_test)
+y_3<span style="color: #666666">=</span>regr_3<span style="color: #666666">.</span>predict(X_test)
+
+<span style="color: #408080; font-style: italic"># Plot the results</span>
+plt<span style="color: #666666">.</span>figure()
+plt<span style="color: #666666">.</span>scatter(X, distance_list, s<span style="color: #666666">=2.5</span>, c<span style="color: #666666">=</span><span style="color: #BA2121">&quot;black&quot;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;data&quot;</span>)
+plt<span style="color: #666666">.</span>plot(X_test, y_1, color<span style="color: #666666">=</span><span style="color: #BA2121">&quot;red&quot;</span>,
+         label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;max_depth=2&quot;</span>, linewidth<span style="color: #666666">=2</span>)
+plt<span style="color: #666666">.</span>plot(X_test, y_2, color<span style="color: #666666">=</span><span style="color: #BA2121">&quot;green&quot;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;max_depth=5&quot;</span>, linewidth<span style="color: #666666">=2</span>)
+plt<span style="color: #666666">.</span>plot(X_test, y_3, color<span style="color: #666666">=</span><span style="color: #BA2121">&quot;m&quot;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;max_depth=7&quot;</span>, linewidth<span style="color: #666666">=2</span>)
+
+plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">&quot;Data&quot;</span>)
+plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">&quot;Darget&quot;</span>)
+plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">&quot;Decision Tree Regression&quot;</span>)
+plt<span style="color: #666666">.</span>legend()
+plt<span style="color: #666666">.</span>show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
 
-<p>where \( \eta \) is our by now well-known learning rate. </p>
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -210,7 +465,7 @@ $$
   <li><a href="._week46-bs017.html">18</a></li>
   <li><a href="._week46-bs018.html">19</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs010.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs010.html b/doc/pub/week46/html/._week46-bs010.html
index 9e5252bbc..1748a8188 100644
--- a/doc/pub/week46/html/._week46-bs010.html
+++ b/doc/pub/week46/html/._week46-bs010.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,35 +329,28 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0010"></a>
 <!-- !split -->
-<h2 id="code-example" class="anchor">Code Example </h2>
+<h2 id="building-a-tree-regression" class="anchor">Building a tree, regression </h2>
 
-<p>The equations we discussed above can be coded rather easily (the
-framework is similar to what we developed for logistic
-regression). We are going to set up a simple case with two classes only and we want to find a line which separates them the best possible way.
+<p>There are mainly two steps</p>
+<ol>
+<li> We split the predictor space (the set of possible values \( x_1,x_2,\dots, x_p \)) into \( J \) distinct and non-non-overlapping regions, \( R_1,R_2,\dots,R_J \).</li>  
+<li> For every observation that falls into the region \( R_j \) , we make the same prediction, which is simply the mean of the response values for the training observations in \( R_j \).</li>
+</ol>
+<p>How do we construct the regions \( R_1,\dots,R_J \)?  In theory, the
+regions could have any shape. However, we choose to divide the
+predictor space into high-dimensional rectangles, or boxes, for
+simplicity and for ease of interpretation of the resulting predictive
+model. The goal is to find boxes \( R_1,\dots,R_J \) that minimize the
+MSE, given by
 </p>
 
-<!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #f8f8f8">
-  <pre style="line-height: 125%;">
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
+$$
+\sum_{j=1}^J\sum_{i\in R_j}(y_i-\overline{y}_{R_j})^2,
+$$
 
+<p>where \( \overline{y}_{R_j} \)  is the mean response for the training observations 
+within box \( j \). 
+</p>
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -227,7 +377,7 @@ regression). We are going to set up a simple case with two classes only and we w
   <li><a href="._week46-bs018.html">19</a></li>
   <li><a href="._week46-bs019.html">20</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs011.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs011.html b/doc/pub/week46/html/._week46-bs011.html
index 2f1909e64..1f7b790d4 100644
--- a/doc/pub/week46/html/._week46-bs011.html
+++ b/doc/pub/week46/html/._week46-bs011.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,17 +329,20 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0011"></a>
 <!-- !split -->
-<h2 id="problems-with-the-simpler-approach" class="anchor">Problems with the Simpler Approach </h2>
+<h2 id="a-top-down-approach-recursive-binary-splitting" class="anchor">A top-down approach, recursive binary splitting </h2>
 
-<p>There are however problems with this approach, although it looks
-pretty straightforward to implement. When running the above code, we see that we can easily end up with many diffeent lines which separate the two classes.
+<p>Unfortunately, it is computationally infeasible to consider every
+possible partition of the feature space into \( J \) boxes.  The common
+strategy is to take a top-down approach
 </p>
 
-<p>For small
-gaps between the entries, we may also end up needing many iterations
-before the solutions converge and if the data cannot be separated
-properly into two distinct classes, we may not experience a converge
-at all.
+<p>The approach is top-down because it begins at the top of the tree (all
+observations belong to a single region) and then successively splits
+the predictor space; each split is indicated via two new branches
+further down on the tree. It is greedy because at each step of the
+tree-building process, the best split is made at that particular step,
+rather than looking ahead and picking a split that will lead to a
+better tree in some future step.
 </p>
 
 <p>
@@ -210,7 +370,7 @@ at all.
   <li><a href="._week46-bs019.html">20</a></li>
   <li><a href="._week46-bs020.html">21</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs012.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs012.html b/doc/pub/week46/html/._week46-bs012.html
index 72cf772d4..7d9a465e7 100644
--- a/doc/pub/week46/html/._week46-bs012.html
+++ b/doc/pub/week46/html/._week46-bs012.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,43 +329,52 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0012"></a>
 <!-- !split -->
-<h2 id="a-better-approach" class="anchor">A better approach </h2>
+<h2 id="making-a-tree" class="anchor">Making a tree </h2>
 
-<p>A better approach is rather to try to define a large margin between
-the two classes (if they are well separated from the beginning).
-</p>
-
-<p>Thus, we wish to find a margin \( M \) with \( \boldsymbol{w} \) normalized to
-\( \vert\vert \boldsymbol{w}\vert\vert =1 \) subject to the condition
-</p>
-
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, p. 
-$$
-
-<p>All points are thus at a signed distance from the decision boundary defined by the line \( L \). The parameters \( b \) and \( w_1 \) and \( w_2 \) define this line. </p>
-
-<p>We seek thus the largest value \( M \) defined by</p>
-$$
-\frac{1}{\vert \vert \boldsymbol{w}\vert\vert}y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, n, 
-$$
-
-<p>or just </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M\vert \vert \boldsymbol{w}\vert\vert \hspace{0.1cm}\forall i. 
-$$
-
-<p>If we scale the equation so that \( \vert \vert \boldsymbol{w}\vert\vert = 1/M \), we have to find the minimum of 
-\( \boldsymbol{w}^T\boldsymbol{w}=\vert \vert \boldsymbol{w}\vert\vert \) (the norm) subject to the condition
+<p>In order to implement the recursive binary splitting we start by selecting
+the predictor \( x_j \) and a cutpoint \( s \) that splits the predictor space into two regions \( R_1 \) and \( R_2 \)
 </p>
 $$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq 1 \hspace{0.1cm}\forall i.
+\left\{X\vert x_j < s\right\},
 $$
 
-<p>We have thus defined our margin as the invers of the norm of
-\( \boldsymbol{w} \). We want to minimize the norm in order to have a as large as
-possible margin \( M \). Before we proceed, we need to remind ourselves
-about Lagrangian multipliers.
+<p>and</p>
+$$
+\left\{X\vert x_j \geq s\right\},
+$$
+
+<p>so that we obtain the lowest MSE, that is</p>
+$$
+\sum_{i:x_i\in R_j}(y_i-\overline{y}_{R_1})^2+\sum_{i:x_i\in R_2}(y_i-\overline{y}_{R_2})^2,
+$$
+
+<p>which we want to minimize by considering all predictors
+\( x_1,x_2,\dots,x_p \).  We consider also all possible values of \( s \) for
+each predictor. These values could be determined by randomly assigned
+numbers or by starting at the midpoint and then proceed till we find
+an optimal value.
+</p>
+
+<p>For any \( j \) and \( s \), we define the pair of half-planes where
+\( \overline{y}_{R_1} \) is the mean response for the training
+observations in \( R_1(j,s) \), and \( \overline{y}_{R_2} \) is the mean
+response for the training observations in \( R_2(j,s) \).
+</p>
+
+<p>Finding the values of \( j \) and \( s \) that minimize the above equation can be
+done quite quickly, especially when the number of features \( p \) is not
+too large.
+</p>
+
+<p>Next, we repeat the process, looking
+for the best predictor and best cutpoint in order to split the data
+further so as to minimize the MSE within each of the resulting
+regions. However, this time, instead of splitting the entire predictor
+space, we split one of the two previously identified regions. We now
+have three regions. Again, we look to split one of these three regions
+further, so as to minimize the MSE. The process continues until a
+stopping criterion is reached; for instance, we may continue until no
+region contains more than five observations.
 </p>
 
 <p>
@@ -236,7 +402,7 @@ about Lagrangian multipliers.
   <li><a href="._week46-bs020.html">21</a></li>
   <li><a href="._week46-bs021.html">22</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs013.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs013.html b/doc/pub/week46/html/._week46-bs013.html
index 75da5b8f3..040ddab41 100644
--- a/doc/pub/week46/html/._week46-bs013.html
+++ b/doc/pub/week46/html/._week46-bs013.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -171,55 +328,24 @@ MathJax.Hub.Config({
 <div class="container">
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0013"></a>
-<!-- !split -->
-<h2 id="a-quick-reminder-on-lagrangian-multipliers" class="anchor">A quick Reminder on Lagrangian Multipliers </h2>
+<!-- !split  -->
+<h2 id="pruning-the-tree" class="anchor">Pruning the tree </h2>
 
-<p>Consider a function of three independent variables \( f(x,y,z) \) . For the function \( f \) to be an
-extreme we have
-</p>
-$$
-df=0.
-$$
-
-<p>A necessary and sufficient condition is</p>
-$$
-\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
-$$
-
-<p>due to</p>
-$$
-df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz.
-$$
-
-<p>In many problems the variables \( x,y,z \) are often subject to constraints (such as those above for the margin)
-so that they are no longer all independent. It is possible at least in principle to use each 
-constraint to eliminate one variable
-and to proceed with a new and smaller set of independent varables.
+<p>The above procedure is rather straightforward, but leads often to
+overfitting and unnecessarily large and complicated trees. The basic
+idea is to grow a large tree \( T_0 \) and then prune it back in order to
+obtain a subtree. A smaller tree with fewer splits (fewer regions) can
+lead to smaller variance and better interpretation at the cost of a
+little more bias.
 </p>
 
-<p>The use of so-called Lagrangian  multipliers is an alternative technique  when the elimination
-of variables is incovenient or undesirable.  Assume that we have an equation of constraint on 
-the variables \( x,y,z \)
+<p>The so-called Cost complexity pruning algorithm gives us a
+way to do just this. Rather than considering every possible subtree,
+we consider a sequence of trees indexed by a nonnegative tuning
+parameter \( \alpha \).
 </p>
-$$
-\phi(x,y,z) = 0,
-$$
 
-<p> resulting in</p>
-$$
-d\phi = \frac{\partial \phi}{\partial x}dx+\frac{\partial \phi}{\partial y}dy+\frac{\partial \phi}{\partial z}dz =0.
-$$
-
-<p>Now we cannot set anymore</p>
-$$
-\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
-$$
-
-<p>if \( df=0 \) is wanted
-because there are now only two independent variables!  Assume \( x \) and \( y \) are the independent 
-variables.
-Then \( dz \) is no longer arbitrary.
-</p>
+<p>Read more at the following <a href="https://scikit-learn.org/stable/auto_examples/tree/plot_cost_complexity_pruning.html#sphx-glr-auto-examples-tree-plot-cost-complexity-pruning-py" target="_self">Scikit-Learn link on pruning</a>.</p>
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -246,7 +372,7 @@ Then \( dz \) is no longer arbitrary.
   <li><a href="._week46-bs021.html">22</a></li>
   <li><a href="._week46-bs022.html">23</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs014.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs014.html b/doc/pub/week46/html/._week46-bs014.html
index 3daacbf58..746560848 100644
--- a/doc/pub/week46/html/._week46-bs014.html
+++ b/doc/pub/week46/html/._week46-bs014.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,45 +329,35 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0014"></a>
 <!-- !split -->
-<h2 id="adding-the-multiplier" class="anchor">Adding the Multiplier </h2>
+<h2 id="cost-complexity-pruning" class="anchor">Cost complexity pruning </h2>
 
-<p>However, we can add to</p>
+<p>For each value of \( \alpha \)  there corresponds a subtree \( T \in T_0 \) such that</p>
 $$
-df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz,
+\sum_{m=1}^{\overline{T}}\sum_{i:x_i\in R_m}(y_i-\overline{y}_{R_m})^2+\alpha\overline{T},
 $$
 
-<p>a multiplum of \( d\phi \), viz. \( \lambda d\phi \), resulting  in</p>
-$$
-df+\lambda d\phi = (\frac{\partial f}{\partial z}+\lambda
-\frac{\partial \phi}{\partial x})dx+(\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y})dy+
-(\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z})dz =0.
-$$
-
-<p>Our multiplier is chosen so that</p>
-$$
-\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z} =0.
-$$
-
-<p>We need to remember that we took \( dx \) and \( dy \) to be arbitrary and thus we must have</p>
-$$
-\frac{\partial f}{\partial x}+\lambda\frac{\partial \phi}{\partial x} =0,
-$$
-
-<p>and</p>
-$$
-\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y} =0.
-$$
-
-<p>When all these equations are satisfied, \( df=0 \).  We have four unknowns, \( x,y,z \) and
-\( \lambda \). Actually we want only \( x,y,z \), \( \lambda \) needs not to be determined, 
-it is therefore often called
-Lagrange's undetermined multiplier.
-If we have a set of constraints \( \phi_k \) we have the equations
+<p>is as small as possible. Here \( \overline{T} \) is 
+the number of terminal nodes of the tree \( T \) , \( R_m \) is the
+rectangle (i.e. the subset of predictor space)  corresponding to the \( m \)-th terminal node.
 </p>
-$$
-\frac{\partial f}{\partial x_i}+\sum_k\lambda_k\frac{\partial \phi_k}{\partial x_i} =0.
-$$
 
+<p>The tuning parameter \( \alpha \) controls a trade-off between the subtree&#8217;s
+complexity and its fit to the training data. When \( \alpha = 0 \), then the
+subtree \( T \) will simply equal \( T_0 \), 
+because then the above equation just measures the
+training error. 
+However, as \( \alpha \) increases, there is a price to pay for
+having a tree with many terminal nodes. The above equation will
+tend to be minimized for a smaller subtree. 
+</p>
+
+<p>It turns out that as we increase \( \alpha \) from zero
+branches get pruned from the tree in a nested and predictable fashion,
+so obtaining the whole sequence of subtrees as a function of \( \alpha \) is
+easy. We can select a value of \( \alpha \) using a validation set or using
+cross-validation. We then return to the full data set and obtain the
+subtree corresponding to \( \alpha \). 
+</p>
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -237,7 +384,7 @@ $$
   <li><a href="._week46-bs022.html">23</a></li>
   <li><a href="._week46-bs023.html">24</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs015.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs015.html b/doc/pub/week46/html/._week46-bs015.html
index 6004a4d15..167fa006c 100644
--- a/doc/pub/week46/html/._week46-bs015.html
+++ b/doc/pub/week46/html/._week46-bs015.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,41 +329,26 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0015"></a>
 <!-- !split -->
-<h2 id="setting-up-the-problem" class="anchor">Setting up the Problem </h2>
-<p>In order to solve the above problem, we define the following Lagrangian function to be minimized </p>
-$$
-{\cal L}(\lambda,b,\boldsymbol{w})=\frac{1}{2}\boldsymbol{w}^T\boldsymbol{w}-\sum_{i=1}^n\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)-1\right],
-$$
+<h2 id="schematic-regression-procedure" class="anchor">Schematic Regression Procedure </h2>
 
-<p>where \( \lambda_i \) is a so-called Lagrange multiplier subject to the condition \( \lambda_i \geq 0 \).</p>
-
-<p>Taking the derivatives  with respect to \( b \) and \( \boldsymbol{w} \) we obtain </p>
-$$
-\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
-$$
-
-<p>and </p>
-$$
-\frac{\partial {\cal L}}{\partial \boldsymbol{w}} = 0 = \boldsymbol{w}-\sum_{i} \lambda_iy_i\boldsymbol{x}_i.
-$$
-
-<p>Inserting these constraints into the equation for \( {\cal L} \) we obtain</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-
-<p>subject to the constraints \( \lambda_i\geq 0 \) and \( \sum_i\lambda_iy_i=0 \). 
-We must in addition satisfy the <a href="https://en.wikipedia.org/wiki/Karush%E2%80%93Kuhn%E2%80%93Tucker_conditions" target="_self">Karush-Kuhn-Tucker</a> (KKT) condition
-</p>
-$$
-\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -1\right] \hspace{0.1cm}\forall i.
-$$
+<div class="panel panel-default">
+<div class="panel-body">
+<!-- subsequent paragraphs come in larger fonts, so start with a paragraph -->
 
 <ol>
-<li> If \( \lambda_i > 0 \), then \( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1 \) and we say that \( x_i \) is on the boundary.</li>
-<li> If \( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)> 1 \), we say \( x_i \) is not on the boundary and we set \( \lambda_i=0 \).</li> 
+<li> Use recursive binary splitting to grow a large tree on the training data, stopping only when each terminal node has fewer than some minimum number of observations.</li>
+<li> Apply cost complexity pruning to the large tree in order to obtain a sequence of best subtrees, as a function of \( \alpha \).</li>
+<li> Use for example \( K \)-fold cross-validation to choose \( \alpha \). Divide the training observations into \( K \) folds. For each \( k=1,2,\dots,K \) we:</li> 
+<ul>
+  <li> repeat steps 1 and 2 on all but the \( k \)-th fold of the training data.</li> 
+  <li> Then we valuate the mean squared prediction error on the data in the left-out \( k \)-th fold, as a function of \( \alpha \).</li>
+  <li> Finally  we average the results for each value of \( \alpha \), and pick \( \alpha \) to minimize the average error.</li>
+</ul>
+<li> Return the subtree from Step 2 that corresponds to the chosen value of \( \alpha \).</li> 
 </ol>
-<p>When \( \lambda_i > 0 \), the vectors \( \boldsymbol{x}_i \) are called support vectors. They are the vectors closest to the line (or hyperplane) and define the margin \( M \). </p>
+</div>
+</div>
+
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -233,7 +375,7 @@ $$
   <li><a href="._week46-bs023.html">24</a></li>
   <li><a href="._week46-bs024.html">25</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs016.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs016.html b/doc/pub/week46/html/._week46-bs016.html
index 72c91d0b8..041c404a1 100644
--- a/doc/pub/week46/html/._week46-bs016.html
+++ b/doc/pub/week46/html/._week46-bs016.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,25 +329,20 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0016"></a>
 <!-- !split -->
-<h2 id="the-problem-to-solve" class="anchor">The problem to solve </h2>
+<h2 id="a-classification-tree" class="anchor">A Classification Tree </h2>
 
-<p>We can rewrite </p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-
-<p>and its constraints in terms of a matrix-vector problem where we minimize w.r.t. \( \lambda \) the following problem</p>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1\boldsymbol{x}_1^T\boldsymbol{x}_1 & y_1y_2\boldsymbol{x}_1^T\boldsymbol{x}_2 & \dots & \dots & y_1y_n\boldsymbol{x}_1^T\boldsymbol{x}_n \\
-y_2y_1\boldsymbol{x}_2^T\boldsymbol{x}_1 & y_2y_2\boldsymbol{x}_2^T\boldsymbol{x}_2 & \dots & \dots & y_1y_n\boldsymbol{x}_2^T\boldsymbol{x}_n \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1\boldsymbol{x}_n^T\boldsymbol{x}_1 & y_ny_2\boldsymbol{x}_n^T\boldsymbol{x}_2 & \dots & \dots & y_ny_n\boldsymbol{x}_n^T\boldsymbol{x}_n \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{1}\boldsymbol{\lambda}, 
-$$
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
+<p>A classification tree is very similar to a regression tree, except
+that it is used to predict a qualitative response rather than a
+quantitative one. Recall that for a regression tree, the predicted
+response for an observation is given by the mean response of the
+training observations that belong to the same terminal node. In
+contrast, for a classification tree, we predict that each observation
+belongs to the most commonly occurring class of training observations
+in the region to which it belongs. In interpreting the results of a
+classification tree, we are often interested not only in the class
+prediction corresponding to a particular terminal node region, but
+also in the class proportions among the training observations that
+fall into that region.  
 </p>
 
 <p>
@@ -218,7 +370,7 @@ $$
   <li><a href="._week46-bs024.html">25</a></li>
   <li><a href="._week46-bs025.html">26</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs017.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs017.html b/doc/pub/week46/html/._week46-bs017.html
index 07a578373..0f27d8e2e 100644
--- a/doc/pub/week46/html/._week46-bs017.html
+++ b/doc/pub/week46/html/._week46-bs017.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,36 +329,26 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0017"></a>
 <!-- !split -->
-<h2 id="the-last-steps" class="anchor">The last steps </h2>
+<h2 id="growing-a-classification-tree" class="anchor">Growing a classification tree </h2>
 
-<p>Solving the above problem, yields the values of \( \lambda_i \).
-To find the coefficients of your hyperplane we need simply to compute 
+<p>The task of growing a
+classification tree is quite similar to the task of growing a
+regression tree. Just as in the regression setting, we use recursive
+binary splitting to grow a classification tree. However, in the
+classification setting, the MSE cannot be used as a criterion for making
+the binary splits.  A natural alternative to MSE is the <b>classification
+error rate</b>. Since we plan to assign an observation in a given region
+to the most commonly occurring error rate class of training
+observations in that region, the classification error rate is simply
+the fraction of the training observations in that region that do not
+belong to the most common class. 
 </p>
-$$
-\boldsymbol{w}=\sum_{i} \lambda_iy_i\boldsymbol{x}_i.
-$$
 
-<p>With our vector \( \boldsymbol{w} \) we can in turn find the value of the intercept \( b \) (here in two dimensions) via </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1,
-$$
-
-<p>resulting in</p>
-$$
-b = \frac{1}{y_i}-\boldsymbol{w}^T\boldsymbol{x}_i,
-$$
-
-<p>or if we write it out in terms of the support vectors only, with \( N_s \) being their number,  we have</p>
-$$
-b = \frac{1}{N_s}\sum_{j\in N_s}\left(y_j-\sum_{i=1}^n\lambda_iy_i\boldsymbol{x}_i^T\boldsymbol{x}_j\right).
-$$
-
-<p>With our hyperplane coefficients we can use our classifier to assign any observation by simply using </p>
-$$
-y_i = \mathrm{sign}(\boldsymbol{w}^T\boldsymbol{x}_i+b).
-$$
-
-<p>Below we discuss how to find the optimal values of \( \lambda_i \). Before we proceed however, we discuss now the so-called soft classifier. </p>
+<p>When building a classification tree, either the Gini index or the
+entropy are typically used to evaluate the quality of a particular
+split, since these two approaches are more sensitive to node purity
+than is the classification error rate. 
+</p>
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -228,7 +375,7 @@ $$
   <li><a href="._week46-bs025.html">26</a></li>
   <li><a href="._week46-bs026.html">27</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs018.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs018.html b/doc/pub/week46/html/._week46-bs018.html
index 562bcf972..862c9be13 100644
--- a/doc/pub/week46/html/._week46-bs018.html
+++ b/doc/pub/week46/html/._week46-bs018.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,38 +329,49 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0018"></a>
 <!-- !split -->
-<h2 id="a-soft-classifier" class="anchor">A soft classifier </h2>
+<h2 id="classification-tree-how-to-split-nodes" class="anchor">Classification tree, how to split nodes </h2>
 
-<p>Till now, the margin is strictly defined by the support vectors. This defines what is called a hard classifier, that is the margins are well defined.</p>
-
-<p>Suppose now that classes overlap in feature space, as shown in the
-figure here. One way to deal with this problem before we define the
-so-called <b>kernel approach</b>, is to allow a kind of slack in the sense
-that we allow some points to be on the wrong side of the margin.
+<p>If our targets are the outcome of a classification process that takes
+for example \( k=1,2,\dots,K \) values, the only thing we need to think of
+is to set up the splitting criteria for each node.
 </p>
 
-<p>We introduce thus the so-called <b>slack</b> variables \( \boldsymbol{\xi} =[\xi_1,x_2,\dots,x_n] \) and 
-modify our previous equation
-</p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1,
-$$
-
-<p>to </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1-\xi_i,
-$$
-
-<p>with the requirement \( \xi_i\geq 0 \). The total violation is now \( \sum_i\xi \). 
-The value \( \xi_i \) in the constraint the last constraint corresponds to the  amount by which the prediction
-\( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1 \) is on the wrong side of its margin. Hence by bounding the sum \( \sum_i \xi_i \),
-we bound the total amount by which predictions fall on the wrong side of their margins.
+<p>We define a PDF \( p_{mk} \) that represents the number of observations of
+a class \( k \) in a region \( R_m \) with \( N_m \) observations. We represent
+this likelihood function in terms of the proportion \( I(y_i=k) \) of
+observations of this class in the region \( R_m \) as
 </p>
 
-<p>Misclassifications occur when \( \xi_i > 1 \). Thus bounding the total sum by some value \( C \) bounds in turn the total number of
-misclassifications.
+$$
+p_{mk} = \frac{1}{N_m}\sum_{x_i\in R_m}I(y_i=k).
+$$
+
+<p>We let \( p_{mk} \) represent the majority class of observations in region
+\( m \). The three most common ways of splitting a node are given by
 </p>
 
+<ul>
+<li> Misclassification error</li> 
+</ul>
+$$
+p_{mk} = \frac{1}{N_m}\sum_{x_i\in R_m}I(y_i\ne k) = 1-p_{mk}.
+$$
+
+<ul>
+<li> Gini index \( g \)</li>
+</ul>
+$$
+g = \sum_{k=1}^K p_{mk}(1-p_{mk}).
+$$
+
+<ul>
+<li> Information entropy or just entropy \( s \)</li>
+</ul>
+$$
+s = -\sum_{k=1}^K p_{mk}\log{p_{mk}}.
+$$
+
+
 <p>
 <!-- navigation buttons at the bottom of the page -->
 <ul class="pagination">
@@ -229,7 +397,7 @@ misclassifications.
   <li><a href="._week46-bs026.html">27</a></li>
   <li><a href="._week46-bs027.html">28</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs019.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs019.html b/doc/pub/week46/html/._week46-bs019.html
index 84084c9d7..2847e9a98 100644
--- a/doc/pub/week46/html/._week46-bs019.html
+++ b/doc/pub/week46/html/._week46-bs019.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,55 +329,61 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0019"></a>
 <!-- !split -->
-<h2 id="soft-optmization-problem" class="anchor">Soft optmization problem </h2>
+<h2 id="visualizing-the-tree-classification" class="anchor">Visualizing the Tree, Classification </h2>
 
-<p>This has in turn the consequences that we change our optmization problem to finding the minimum of </p>
-$$
-{\cal L}=\frac{1}{2}\boldsymbol{w}^T\boldsymbol{w}-\sum_{i=1}^n\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)-(1-\xi_)\right]+C\sum_{i=1}^n\xi_i-\sum_{i=1}^n\gamma_i\xi_i,
-$$
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">os</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> load_breast_cancer
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span> train_test_split
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.metrics</span> <span style="color: #008000; font-weight: bold">import</span> confusion_matrix
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> export_graphviz
 
-<p>subject to </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1-\xi_i \hspace{0.1cm}\forall i,
-$$
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">IPython.display</span> <span style="color: #008000; font-weight: bold">import</span> Image 
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">pydot</span> <span style="color: #008000; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">pandas</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">pd</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
 
-<p>with the requirement \( \xi_i\geq 0 \).</p>
 
-<p>Taking the derivatives  with respect to \( b \) and \( \boldsymbol{w} \) we obtain </p>
-$$
-\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
-$$
+cancer <span style="color: #666666">=</span> load_breast_cancer()
+X <span style="color: #666666">=</span> pd<span style="color: #666666">.</span>DataFrame(cancer<span style="color: #666666">.</span>data, columns<span style="color: #666666">=</span>cancer<span style="color: #666666">.</span>feature_names)
+<span style="color: #008000">print</span>(X)
+y <span style="color: #666666">=</span> pd<span style="color: #666666">.</span>Categorical<span style="color: #666666">.</span>from_codes(cancer<span style="color: #666666">.</span>target, cancer<span style="color: #666666">.</span>target_names)
+y <span style="color: #666666">=</span> pd<span style="color: #666666">.</span>get_dummies(y)
+<span style="color: #008000">print</span>(y)
+X_train, X_test, y_train, y_test <span style="color: #666666">=</span> train_test_split(X, y, random_state<span style="color: #666666">=1</span>)
+tree_clf <span style="color: #666666">=</span> DecisionTreeClassifier(max_depth<span style="color: #666666">=5</span>)
+tree_clf<span style="color: #666666">.</span>fit(X_train, y_train)
 
-<p>and </p>
-$$
-\frac{\partial {\cal L}}{\partial \boldsymbol{w}} = 0 = \boldsymbol{w}-\sum_{i} \lambda_iy_i\boldsymbol{x}_i, 
-$$
-
-<p>and</p>
-$$
-\lambda_i = C-\gamma_i \hspace{0.1cm}\forall i.
-$$
-
-<p>Inserting these constraints into the equation for \( {\cal L} \) we obtain the same equation as before</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-
-<p>but now subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \) and \( 0\leq\lambda_i \leq C \). 
-We must in addition satisfy the Karush-Kuhn-Tucker condition which now reads
-</p>
-$$
-\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -(1-\xi_)\right]=0 \hspace{0.1cm}\forall i,
-$$
-
-$$
-\gamma_i\xi_i = 0,
-$$
-
-<p>and </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -(1-\xi_) \geq 0 \hspace{0.1cm}\forall i.
-$$
+export_graphviz(
+    tree_clf,
+    out_file<span style="color: #666666">=</span><span style="color: #BA2121">&quot;DataFiles/cancer.dot&quot;</span>,
+    feature_names<span style="color: #666666">=</span>cancer<span style="color: #666666">.</span>feature_names,
+    class_names<span style="color: #666666">=</span>cancer<span style="color: #666666">.</span>target_names,
+    rounded<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>,
+    filled<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>
+)
+cmd <span style="color: #666666">=</span> <span style="color: #BA2121">&#39;dot -Tpng DataFiles/cancer.dot -o DataFiles/cancer.png&#39;</span>
+os<span style="color: #666666">.</span>system(cmd)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
 
 
 <p>
@@ -248,7 +411,7 @@ $$
   <li><a href="._week46-bs027.html">28</a></li>
   <li><a href="._week46-bs028.html">29</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs020.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs020.html b/doc/pub/week46/html/._week46-bs020.html
index 78f4ad881..07343c15a 100644
--- a/doc/pub/week46/html/._week46-bs020.html
+++ b/doc/pub/week46/html/._week46-bs020.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,26 +329,7 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0020"></a>
 <!-- !split -->
-<h2 id="kernels-and-non-linearity" class="anchor">Kernels and non-linearity </h2>
-
-<p>The cases we have studied till now, were all characterized by two classes
-with a close to linear separability. The classifiers we have described
-so far find linear boundaries in our input feature space. It is
-possible to make our procedure more flexible by exploring the feature
-space using other basis expansions such as higher-order polynomials,
-wavelets, splines etc.
-</p>
-
-<p>If our feature space is not easy to separate, as shown in the figure
-here, we can achieve a better separation by introducing more complex
-basis functions. The ideal would be, as shown in the next figure, to, via a specific transformation to 
-obtain a separation between the classes which is almost linear. 
-</p>
-
-<p>The change of basis, from \( x\rightarrow z=\phi(x) \) leads to the same type of equations to be solved, except that
-we need to introduce for example a polynomial transformation to a two-dimensional training set.
-</p>
-
+<h2 id="visualizing-the-tree-the-moons" class="anchor">Visualizing the Tree, The Moons  </h2>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
 <div class="cell border-box-sizing code_cell rendered">
@@ -199,52 +337,30 @@ we need to introduce for example a polynomial transformation to a two-dimensiona
     <div class="inner_cell">
       <div class="input_area">
         <div class="highlight" style="background: #f8f8f8">
-  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
+  <pre style="line-height: 125%;"><span style="color: #408080; font-style: italic"># Common imports</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span>  train_test_split 
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> make_moons
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> export_graphviz
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">pydot</span> <span style="color: #008000; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">pandas</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">pd</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">os</span>
 
 np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>seed(<span style="color: #666666">42</span>)
+X, y <span style="color: #666666">=</span> make_moons(n_samples<span style="color: #666666">=100</span>, noise<span style="color: #666666">=0.25</span>, random_state<span style="color: #666666">=53</span>)
+X_train, X_test, y_train, y_test <span style="color: #666666">=</span> train_test_split(X,y,random_state<span style="color: #666666">=0</span>)
+tree_clf <span style="color: #666666">=</span> DecisionTreeClassifier(max_depth<span style="color: #666666">=5</span>)
+tree_clf<span style="color: #666666">.</span>fit(X_train, y_train)
 
-<span style="color: #408080; font-style: italic"># To plot pretty figures</span>
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib</span>
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;axes.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">14</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;xtick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;ytick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
-
-
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn</span> <span style="color: #008000; font-weight: bold">import</span> datasets
-
-
-
-X1D <span style="color: #666666">=</span> np<span style="color: #666666">.</span>linspace(<span style="color: #666666">-4</span>, <span style="color: #666666">4</span>, <span style="color: #666666">9</span>)<span style="color: #666666">.</span>reshape(<span style="color: #666666">-1</span>, <span style="color: #666666">1</span>)
-X2D <span style="color: #666666">=</span> np<span style="color: #666666">.</span>c_[X1D, X1D<span style="color: #666666">**2</span>]
-y <span style="color: #666666">=</span> np<span style="color: #666666">.</span>array([<span style="color: #666666">0</span>, <span style="color: #666666">0</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">0</span>, <span style="color: #666666">0</span>])
-
-plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">4</span>))
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">121</span>)
-plt<span style="color: #666666">.</span>grid(<span style="color: #008000; font-weight: bold">True</span>, which<span style="color: #666666">=</span><span style="color: #BA2121">&#39;both&#39;</span>)
-plt<span style="color: #666666">.</span>axhline(y<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>plot(X1D[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==0</span>], np<span style="color: #666666">.</span>zeros(<span style="color: #666666">4</span>), <span style="color: #BA2121">&quot;bs&quot;</span>)
-plt<span style="color: #666666">.</span>plot(X1D[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==1</span>], np<span style="color: #666666">.</span>zeros(<span style="color: #666666">5</span>), <span style="color: #BA2121">&quot;g^&quot;</span>)
-plt<span style="color: #666666">.</span>gca()<span style="color: #666666">.</span>get_yaxis()<span style="color: #666666">.</span>set_ticks([])
-plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">r&quot;$x_1$&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>axis([<span style="color: #666666">-4.5</span>, <span style="color: #666666">4.5</span>, <span style="color: #666666">-0.2</span>, <span style="color: #666666">0.2</span>])
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">122</span>)
-plt<span style="color: #666666">.</span>grid(<span style="color: #008000; font-weight: bold">True</span>, which<span style="color: #666666">=</span><span style="color: #BA2121">&#39;both&#39;</span>)
-plt<span style="color: #666666">.</span>axhline(y<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>axvline(x<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>plot(X2D[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==0</span>], X2D[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==0</span>], <span style="color: #BA2121">&quot;bs&quot;</span>)
-plt<span style="color: #666666">.</span>plot(X2D[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==1</span>], X2D[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==1</span>], <span style="color: #BA2121">&quot;g^&quot;</span>)
-plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">r&quot;$x_1$&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">r&quot;$x_2$&quot;</span>, fontsize<span style="color: #666666">=20</span>, rotation<span style="color: #666666">=0</span>)
-plt<span style="color: #666666">.</span>gca()<span style="color: #666666">.</span>get_yaxis()<span style="color: #666666">.</span>set_ticks([<span style="color: #666666">0</span>, <span style="color: #666666">4</span>, <span style="color: #666666">8</span>, <span style="color: #666666">12</span>, <span style="color: #666666">16</span>])
-plt<span style="color: #666666">.</span>plot([<span style="color: #666666">-4.5</span>, <span style="color: #666666">4.5</span>], [<span style="color: #666666">6.5</span>, <span style="color: #666666">6.5</span>], <span style="color: #BA2121">&quot;r--&quot;</span>, linewidth<span style="color: #666666">=3</span>)
-plt<span style="color: #666666">.</span>axis([<span style="color: #666666">-4.5</span>, <span style="color: #666666">4.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">17</span>])
-plt<span style="color: #666666">.</span>subplots_adjust(right<span style="color: #666666">=1</span>)
-plt<span style="color: #666666">.</span>show()
+export_graphviz(
+    tree_clf,
+    out_file<span style="color: #666666">=</span><span style="color: #BA2121">&quot;DataFiles/moons.dot&quot;</span>,
+    rounded<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>,
+    filled<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>
+)
+cmd <span style="color: #666666">=</span> <span style="color: #BA2121">&#39;dot -Tpng DataFiles/moons.dot -o DataFiles/moons.png&#39;</span>
+os<span style="color: #666666">.</span>system(cmd)
 </pre>
 </div>
       </div>
@@ -285,6 +401,8 @@ plt<span style="color: #666666">.</span>show()
   <li><a href="._week46-bs027.html">28</a></li>
   <li><a href="._week46-bs028.html">29</a></li>
   <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="">...</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs021.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs021.html b/doc/pub/week46/html/._week46-bs021.html
index c23177cb6..f24c735bc 100644
--- a/doc/pub/week46/html/._week46-bs021.html
+++ b/doc/pub/week46/html/._week46-bs021.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,46 +329,39 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0021"></a>
 <!-- !split -->
-<h2 id="the-equations" class="anchor">The equations </h2>
+<h2 id="other-ways-of-visualizing-the-trees" class="anchor">Other ways of visualizing the trees </h2>
 
-<p>Suppose we define a polynomial transformation of degree two only (we continue to live in a plane with \( x_i \) and \( y_i \) as variables)</p>
-$$
-z = \phi(x_i) =\left(x_i^2, y_i^2, \sqrt{2}x_iy_i\right).
-$$
+<p><b>Scikit-Learn</b> has also another way to visualize the trees which is very useful, here with the Iris data.</p>
 
-<p>With our new basis, the equations we solved earlier are basically the same, that is we have now (without the slack option for simplicity)</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{z}_i^T\boldsymbol{z}_j,
-$$
 
-<p>subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \), and for the support vectors</p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{z}_i+b)= 1 \hspace{0.1cm}\forall i,
-$$
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> load_iris
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn</span> <span style="color: #008000; font-weight: bold">import</span> tree
+X, y <span style="color: #666666">=</span> load_iris(return_X_y<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>)
+tree_clf <span style="color: #666666">=</span> tree<span style="color: #666666">.</span>DecisionTreeClassifier()
+tree_clf <span style="color: #666666">=</span> tree_clf<span style="color: #666666">.</span>fit(X, y)
+<span style="color: #408080; font-style: italic"># and then plot the tree</span>
+tree<span style="color: #666666">.</span>plot_tree(tree_clf) 
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
 
-<p>from which we also find \( b \).
-To compute \( \boldsymbol{z}_i^T\boldsymbol{z}_j \) we define the kernel \( K(\boldsymbol{x}_i,\boldsymbol{x}_j) \) as
-</p>
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=\boldsymbol{z}_i^T\boldsymbol{z}_j= \phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j).
-$$
-
-<p>For the above example, the kernel reads</p>
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=[x_i^2, y_i^2, \sqrt{2}x_iy_i]^T\begin{bmatrix} x_j^2 \\ y_j^2 \\ \sqrt{2}x_jy_j \end{bmatrix}=x_i^2x_j^2+2x_ix_jy_iy_j+y_i^2y_j^2.
-$$
-
-<p>We note that this is nothing but the dot product of the two original
-vectors \( (\boldsymbol{x}_i^T\boldsymbol{x}_j)^2 \). Instead of thus computing the
-product in the Lagrangian of \( \boldsymbol{z}_i^T\boldsymbol{z}_j \) we simply compute
-the dot product \( (\boldsymbol{x}_i^T\boldsymbol{x}_j)^2 \).
-</p>
-
-<p>This leads to the so-called
-kernel trick and the result leads to the same as if we went through
-the trouble of performing the transformation
-\( \phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j) \) during the SVM calculations.
-</p>
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -236,6 +386,9 @@ the trouble of performing the transformation
   <li><a href="._week46-bs027.html">28</a></li>
   <li><a href="._week46-bs028.html">29</a></li>
   <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs030.html">31</a></li>
+  <li><a href="">...</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs022.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs022.html b/doc/pub/week46/html/._week46-bs022.html
index 883b5eafa..d638d6cb1 100644
--- a/doc/pub/week46/html/._week46-bs022.html
+++ b/doc/pub/week46/html/._week46-bs022.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,39 +329,42 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0022"></a>
 <!-- !split -->
-<h2 id="the-problem-to-solve" class="anchor">The problem to solve </h2>
-<p>Using our definition of the kernel We can rewrite again the Lagrangian</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{z}_j,
-$$
+<h2 id="printing-out-as-text" class="anchor">Printing out as text </h2>
 
-<p>subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \) in terms of a convex optimization problem</p>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1K(\boldsymbol{x}_1,\boldsymbol{x}_1) & y_1y_2K(\boldsymbol{x}_1,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_1,\boldsymbol{x}_n) \\
-y_2y_1K(\boldsymbol{x}_2,\boldsymbol{x}_1) & y_2y_2(\boldsymbol{x}_2,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_2,\boldsymbol{x}_n) \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1K(\boldsymbol{x}_n,\boldsymbol{x}_1) & y_ny_2K(\boldsymbol{x}_n\boldsymbol{x}_2) & \dots & \dots & y_ny_nK(\boldsymbol{x}_n,\boldsymbol{x}_n) \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{1}\boldsymbol{\lambda}, 
-$$
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
-If we add the slack constants this leads to the additional constraint \( 0\leq \lambda_i \leq C \).
+<p>Alternatively, the tree can also be exported in textual format with the function exporttext.
+This method doesn&#8217;t require the installation of external libraries and is more compact:
 </p>
 
-<p>We can rewrite this (see the solutions below) in terms of a convex optimization problem of the type</p>
-$$
-\begin{align*}
-    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\boldsymbol{\lambda}^T\boldsymbol{P}\boldsymbol{\lambda}+\boldsymbol{q}^T\boldsymbol{\lambda},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{\lambda} \preceq \boldsymbol{h} \hspace{0.2cm} \wedge \boldsymbol{A}\boldsymbol{\lambda}=f.
-\end{align*}
-$$
 
-<p>Below we discuss how to solve these equations. Here we note that the matrix \( \boldsymbol{P} \) has matrix elements \( p_{ij}=y_iy_jK(\boldsymbol{x}_i,\boldsymbol{x}_j) \).
-Given a kernel \( K \) and the targets \( y_i \) this matrix is easy to set up. The constraint \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \) leads to \( f=0 \) and \( \boldsymbol{A}=\boldsymbol{y} \). How to set up the matrix \( \boldsymbol{G} \) is discussed later. Here note that the inequalities \( 0\leq \lambda_i \leq C \) can be split up into
-\( 0\leq \lambda_i \) and \( \lambda_i \leq C \). These two inequalities define then the matrix \( \boldsymbol{G} \) and the vector \( \boldsymbol{h} \).
-</p>
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> load_iris
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> export_text
+iris <span style="color: #666666">=</span> load_iris()
+decision_tree <span style="color: #666666">=</span> DecisionTreeClassifier(random_state<span style="color: #666666">=0</span>, max_depth<span style="color: #666666">=2</span>)
+decision_tree <span style="color: #666666">=</span> decision_tree<span style="color: #666666">.</span>fit(iris<span style="color: #666666">.</span>data, iris<span style="color: #666666">.</span>target)
+r <span style="color: #666666">=</span> export_text(decision_tree, feature_names<span style="color: #666666">=</span>iris[<span style="color: #BA2121">&#39;feature_names&#39;</span>])
+<span style="color: #008000">print</span>(r)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -228,6 +388,10 @@ Given a kernel \( K \) and the targets \( y_i \) this matrix is easy to set up.
   <li><a href="._week46-bs027.html">28</a></li>
   <li><a href="._week46-bs028.html">29</a></li>
   <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs030.html">31</a></li>
+  <li><a href="._week46-bs031.html">32</a></li>
+  <li><a href="">...</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs023.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs023.html b/doc/pub/week46/html/._week46-bs023.html
index 3687a89d8..32687de86 100644
--- a/doc/pub/week46/html/._week46-bs023.html
+++ b/doc/pub/week46/html/._week46-bs023.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,36 +329,17 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0023"></a>
 <!-- !split -->
-<h2 id="different-kernels-and-mercer-s-theorem" class="anchor">Different kernels and Mercer's theorem </h2>
+<h2 id="algorithms-for-setting-up-decision-trees" class="anchor">Algorithms for Setting up Decision Trees </h2>
 
-<p>There are several popular kernels being used. These are</p>
+<p>Two algorithms stand out in the set up of decision trees:</p>
 <ol>
-<li> Linear: \( K(\boldsymbol{x},\boldsymbol{y})=\boldsymbol{x}^T\boldsymbol{y} \),</li>
-<li> Polynomial: \( K(\boldsymbol{x},\boldsymbol{y})=(\boldsymbol{x}^T\boldsymbol{y}+\gamma)^d \),</li>
-<li> Gaussian Radial Basis Function: \( K(\boldsymbol{x},\boldsymbol{y})=\exp{\left(-\gamma\vert\vert\boldsymbol{x}-\boldsymbol{y}\vert\vert^2\right)} \),</li>
-<li> Tanh: \( K(\boldsymbol{x},\boldsymbol{y})=\tanh{(\boldsymbol{x}^T\boldsymbol{y}+\gamma)} \),</li>
+<li> The CART (Classification And Regression Tree) algorithm for both classification and regression</li>
+<li> The ID3 algorithm based on the computation of the information gain for classification</li>
 </ol>
-<p>and many other ones.</p>
-
-<p>An important theorem for us is <a href="https://en.wikipedia.org/wiki/Mercer%27s_theorem" target="_self">Mercer's
-theorem</a>.  The
-theorem states that if a kernel function \( K \) is symmetric, continuous
-and leads to a positive semi-definite matrix \( \boldsymbol{P} \) then there
-exists a function \( \phi \) that maps \( \boldsymbol{x}_i \) and \( \boldsymbol{x}_j \) into
-another space (possibly with much higher dimensions) such that
-</p>
-
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=\phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j).
-$$
-
-<p>So you can use \( K \) as a kernel since you know \( \phi \) exists, even if
-you don&#8217;t know what \( \phi \) is. 
-</p>
-
-<p>Note that some frequently used kernels (such as the Sigmoid kernel)
-don&#8217;t respect all of Mercer&#8217;s conditions, yet they generally work well
-in practice.
+<p>We discuss both algorithms with applications here. The popular library
+<b>Scikit-Learn</b> uses the CART algorithm. For classification problems
+you can use either the <b>gini</b> index or the <b>entropy</b> to split a tree
+in two branches.
 </p>
 
 <p>
@@ -225,6 +363,11 @@ in practice.
   <li><a href="._week46-bs027.html">28</a></li>
   <li><a href="._week46-bs028.html">29</a></li>
   <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs030.html">31</a></li>
+  <li><a href="._week46-bs031.html">32</a></li>
+  <li><a href="._week46-bs032.html">33</a></li>
+  <li><a href="">...</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs024.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs024.html b/doc/pub/week46/html/._week46-bs024.html
index 0da1afa7e..089f86464 100644
--- a/doc/pub/week46/html/._week46-bs024.html
+++ b/doc/pub/week46/html/._week46-bs024.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,217 +329,30 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0024"></a>
 <!-- !split -->
-<h2 id="the-moons-example" class="anchor">The moons example </h2>
+<h2 id="the-cart-algorithm-for-classification" class="anchor">The CART algorithm for Classification </h2>
 
-<!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #f8f8f8">
-  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">__future__</span> <span style="color: #008000; font-weight: bold">import</span> division, print_function, unicode_literals
+<p>For classification, the CART algorithm splits the data set in two subsets using a single feature \( k \) and a threshold \( t_k \).
+This could be for example a threshold set by a number below a certain circumference of a malign tumor.
+</p>
 
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
-np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>seed(<span style="color: #666666">42</span>)
+<p>How do we find these two quantities?
+We search for the pair \( (k,t_k) \) that produces the purest subset using for example the <b>gini</b> factor \( G \).
+The cost function it tries to minimize is then
+</p>
+$$
+C(k,t_k) = \frac{m_{\mathrm{left}}}{m}G_{\mathrm{left}}+ \frac{m_{\mathrm{right}}}{m}G_{\mathrm{right}},
+$$
 
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib</span>
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;axes.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">14</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;xtick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;ytick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
-
-
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn</span> <span style="color: #008000; font-weight: bold">import</span> datasets
-
-
-
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.pipeline</span> <span style="color: #008000; font-weight: bold">import</span> Pipeline
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> StandardScaler
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> LinearSVC
-
-
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> make_moons
-X, y <span style="color: #666666">=</span> make_moons(n_samples<span style="color: #666666">=100</span>, noise<span style="color: #666666">=0.15</span>, random_state<span style="color: #666666">=42</span>)
-
-<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">plot_dataset</span>(X, y, axes):
-    plt<span style="color: #666666">.</span>plot(X[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==0</span>], X[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==0</span>], <span style="color: #BA2121">&quot;bs&quot;</span>)
-    plt<span style="color: #666666">.</span>plot(X[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==1</span>], X[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==1</span>], <span style="color: #BA2121">&quot;g^&quot;</span>)
-    plt<span style="color: #666666">.</span>axis(axes)
-    plt<span style="color: #666666">.</span>grid(<span style="color: #008000; font-weight: bold">True</span>, which<span style="color: #666666">=</span><span style="color: #BA2121">&#39;both&#39;</span>)
-    plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">r&quot;$x_1$&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-    plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">r&quot;$x_2$&quot;</span>, fontsize<span style="color: #666666">=20</span>, rotation<span style="color: #666666">=0</span>)
-
-plot_dataset(X, y, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plt<span style="color: #666666">.</span>show()
-
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> make_moons
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.pipeline</span> <span style="color: #008000; font-weight: bold">import</span> Pipeline
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> PolynomialFeatures
-
-polynomial_svm_clf <span style="color: #666666">=</span> Pipeline([
-        (<span style="color: #BA2121">&quot;poly_features&quot;</span>, PolynomialFeatures(degree<span style="color: #666666">=3</span>)),
-        (<span style="color: #BA2121">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #BA2121">&quot;svm_clf&quot;</span>, LinearSVC(C<span style="color: #666666">=10</span>, loss<span style="color: #666666">=</span><span style="color: #BA2121">&quot;hinge&quot;</span>, random_state<span style="color: #666666">=42</span>))
-    ])
-
-polynomial_svm_clf<span style="color: #666666">.</span>fit(X, y)
-
-<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">plot_predictions</span>(clf, axes):
-    x0s <span style="color: #666666">=</span> np<span style="color: #666666">.</span>linspace(axes[<span style="color: #666666">0</span>], axes[<span style="color: #666666">1</span>], <span style="color: #666666">100</span>)
-    x1s <span style="color: #666666">=</span> np<span style="color: #666666">.</span>linspace(axes[<span style="color: #666666">2</span>], axes[<span style="color: #666666">3</span>], <span style="color: #666666">100</span>)
-    x0, x1 <span style="color: #666666">=</span> np<span style="color: #666666">.</span>meshgrid(x0s, x1s)
-    X <span style="color: #666666">=</span> np<span style="color: #666666">.</span>c_[x0<span style="color: #666666">.</span>ravel(), x1<span style="color: #666666">.</span>ravel()]
-    y_pred <span style="color: #666666">=</span> clf<span style="color: #666666">.</span>predict(X)<span style="color: #666666">.</span>reshape(x0<span style="color: #666666">.</span>shape)
-    y_decision <span style="color: #666666">=</span> clf<span style="color: #666666">.</span>decision_function(X)<span style="color: #666666">.</span>reshape(x0<span style="color: #666666">.</span>shape)
-    plt<span style="color: #666666">.</span>contourf(x0, x1, y_pred, cmap<span style="color: #666666">=</span>plt<span style="color: #666666">.</span>cm<span style="color: #666666">.</span>brg, alpha<span style="color: #666666">=0.2</span>)
-    plt<span style="color: #666666">.</span>contourf(x0, x1, y_decision, cmap<span style="color: #666666">=</span>plt<span style="color: #666666">.</span>cm<span style="color: #666666">.</span>brg, alpha<span style="color: #666666">=0.1</span>)
-
-plot_predictions(polynomial_svm_clf, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plot_dataset(X, y, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-
-plt<span style="color: #666666">.</span>show()
-
-
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC
-
-poly_kernel_svm_clf <span style="color: #666666">=</span> Pipeline([
-        (<span style="color: #BA2121">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #BA2121">&quot;svm_clf&quot;</span>, SVC(kernel<span style="color: #666666">=</span><span style="color: #BA2121">&quot;poly&quot;</span>, degree<span style="color: #666666">=3</span>, coef0<span style="color: #666666">=1</span>, C<span style="color: #666666">=5</span>))
-    ])
-poly_kernel_svm_clf<span style="color: #666666">.</span>fit(X, y)
-
-poly100_kernel_svm_clf <span style="color: #666666">=</span> Pipeline([
-        (<span style="color: #BA2121">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #BA2121">&quot;svm_clf&quot;</span>, SVC(kernel<span style="color: #666666">=</span><span style="color: #BA2121">&quot;poly&quot;</span>, degree<span style="color: #666666">=10</span>, coef0<span style="color: #666666">=100</span>, C<span style="color: #666666">=5</span>))
-    ])
-poly100_kernel_svm_clf<span style="color: #666666">.</span>fit(X, y)
-
-plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">4</span>))
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">121</span>)
-plot_predictions(poly_kernel_svm_clf, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plot_dataset(X, y, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">r&quot;$d=3, r=1, C=5$&quot;</span>, fontsize<span style="color: #666666">=18</span>)
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">122</span>)
-plot_predictions(poly100_kernel_svm_clf, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plot_dataset(X, y, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">r&quot;$d=10, r=100, C=5$&quot;</span>, fontsize<span style="color: #666666">=18</span>)
-
-plt<span style="color: #666666">.</span>show()
-
-<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">gaussian_rbf</span>(x, landmark, gamma):
-    <span style="color: #008000; font-weight: bold">return</span> np<span style="color: #666666">.</span>exp(<span style="color: #666666">-</span>gamma <span style="color: #666666">*</span> np<span style="color: #666666">.</span>linalg<span style="color: #666666">.</span>norm(x <span style="color: #666666">-</span> landmark, axis<span style="color: #666666">=1</span>)<span style="color: #666666">**2</span>)
-
-gamma <span style="color: #666666">=</span> <span style="color: #666666">0.3</span>
-
-x1s <span style="color: #666666">=</span> np<span style="color: #666666">.</span>linspace(<span style="color: #666666">-4.5</span>, <span style="color: #666666">4.5</span>, <span style="color: #666666">200</span>)<span style="color: #666666">.</span>reshape(<span style="color: #666666">-1</span>, <span style="color: #666666">1</span>)
-x2s <span style="color: #666666">=</span> gaussian_rbf(x1s, <span style="color: #666666">-2</span>, gamma)
-x3s <span style="color: #666666">=</span> gaussian_rbf(x1s, <span style="color: #666666">1</span>, gamma)
-
-XK <span style="color: #666666">=</span> np<span style="color: #666666">.</span>c_[gaussian_rbf(X1D, <span style="color: #666666">-2</span>, gamma), gaussian_rbf(X1D, <span style="color: #666666">1</span>, gamma)]
-yk <span style="color: #666666">=</span> np<span style="color: #666666">.</span>array([<span style="color: #666666">0</span>, <span style="color: #666666">0</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">0</span>, <span style="color: #666666">0</span>])
-
-plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">4</span>))
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">121</span>)
-plt<span style="color: #666666">.</span>grid(<span style="color: #008000; font-weight: bold">True</span>, which<span style="color: #666666">=</span><span style="color: #BA2121">&#39;both&#39;</span>)
-plt<span style="color: #666666">.</span>axhline(y<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>scatter(x<span style="color: #666666">=</span>[<span style="color: #666666">-2</span>, <span style="color: #666666">1</span>], y<span style="color: #666666">=</span>[<span style="color: #666666">0</span>, <span style="color: #666666">0</span>], s<span style="color: #666666">=150</span>, alpha<span style="color: #666666">=0.5</span>, c<span style="color: #666666">=</span><span style="color: #BA2121">&quot;red&quot;</span>)
-plt<span style="color: #666666">.</span>plot(X1D[:, <span style="color: #666666">0</span>][yk<span style="color: #666666">==0</span>], np<span style="color: #666666">.</span>zeros(<span style="color: #666666">4</span>), <span style="color: #BA2121">&quot;bs&quot;</span>)
-plt<span style="color: #666666">.</span>plot(X1D[:, <span style="color: #666666">0</span>][yk<span style="color: #666666">==1</span>], np<span style="color: #666666">.</span>zeros(<span style="color: #666666">5</span>), <span style="color: #BA2121">&quot;g^&quot;</span>)
-plt<span style="color: #666666">.</span>plot(x1s, x2s, <span style="color: #BA2121">&quot;g--&quot;</span>)
-plt<span style="color: #666666">.</span>plot(x1s, x3s, <span style="color: #BA2121">&quot;b:&quot;</span>)
-plt<span style="color: #666666">.</span>gca()<span style="color: #666666">.</span>get_yaxis()<span style="color: #666666">.</span>set_ticks([<span style="color: #666666">0</span>, <span style="color: #666666">0.25</span>, <span style="color: #666666">0.5</span>, <span style="color: #666666">0.75</span>, <span style="color: #666666">1</span>])
-plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">r&quot;$x_1$&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">r&quot;Similarity&quot;</span>, fontsize<span style="color: #666666">=14</span>)
-plt<span style="color: #666666">.</span>annotate(<span style="color: #BA2121">r&#39;$\mathbf</span><span style="color: #BB6688; font-weight: bold">{x}</span><span style="color: #BA2121">$&#39;</span>,
-             xy<span style="color: #666666">=</span>(X1D[<span style="color: #666666">3</span>, <span style="color: #666666">0</span>], <span style="color: #666666">0</span>),
-             xytext<span style="color: #666666">=</span>(<span style="color: #666666">-0.5</span>, <span style="color: #666666">0.20</span>),
-             ha<span style="color: #666666">=</span><span style="color: #BA2121">&quot;center&quot;</span>,
-             arrowprops<span style="color: #666666">=</span><span style="color: #008000">dict</span>(facecolor<span style="color: #666666">=</span><span style="color: #BA2121">&#39;black&#39;</span>, shrink<span style="color: #666666">=0.1</span>),
-             fontsize<span style="color: #666666">=18</span>,
-            )
-plt<span style="color: #666666">.</span>text(<span style="color: #666666">-2</span>, <span style="color: #666666">0.9</span>, <span style="color: #BA2121">&quot;$x_2$&quot;</span>, ha<span style="color: #666666">=</span><span style="color: #BA2121">&quot;center&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>text(<span style="color: #666666">1</span>, <span style="color: #666666">0.9</span>, <span style="color: #BA2121">&quot;$x_3$&quot;</span>, ha<span style="color: #666666">=</span><span style="color: #BA2121">&quot;center&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>axis([<span style="color: #666666">-4.5</span>, <span style="color: #666666">4.5</span>, <span style="color: #666666">-0.1</span>, <span style="color: #666666">1.1</span>])
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">122</span>)
-plt<span style="color: #666666">.</span>grid(<span style="color: #008000; font-weight: bold">True</span>, which<span style="color: #666666">=</span><span style="color: #BA2121">&#39;both&#39;</span>)
-plt<span style="color: #666666">.</span>axhline(y<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>axvline(x<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>plot(XK[:, <span style="color: #666666">0</span>][yk<span style="color: #666666">==0</span>], XK[:, <span style="color: #666666">1</span>][yk<span style="color: #666666">==0</span>], <span style="color: #BA2121">&quot;bs&quot;</span>)
-plt<span style="color: #666666">.</span>plot(XK[:, <span style="color: #666666">0</span>][yk<span style="color: #666666">==1</span>], XK[:, <span style="color: #666666">1</span>][yk<span style="color: #666666">==1</span>], <span style="color: #BA2121">&quot;g^&quot;</span>)
-plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">r&quot;$x_2$&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">r&quot;$x_3$  &quot;</span>, fontsize<span style="color: #666666">=20</span>, rotation<span style="color: #666666">=0</span>)
-plt<span style="color: #666666">.</span>annotate(<span style="color: #BA2121">r&#39;$\phi\left(\mathbf</span><span style="color: #BB6688; font-weight: bold">{x}</span><span style="color: #BA2121">\right)$&#39;</span>,
-             xy<span style="color: #666666">=</span>(XK[<span style="color: #666666">3</span>, <span style="color: #666666">0</span>], XK[<span style="color: #666666">3</span>, <span style="color: #666666">1</span>]),
-             xytext<span style="color: #666666">=</span>(<span style="color: #666666">0.65</span>, <span style="color: #666666">0.50</span>),
-             ha<span style="color: #666666">=</span><span style="color: #BA2121">&quot;center&quot;</span>,
-             arrowprops<span style="color: #666666">=</span><span style="color: #008000">dict</span>(facecolor<span style="color: #666666">=</span><span style="color: #BA2121">&#39;black&#39;</span>, shrink<span style="color: #666666">=0.1</span>),
-             fontsize<span style="color: #666666">=18</span>,
-            )
-plt<span style="color: #666666">.</span>plot([<span style="color: #666666">-0.1</span>, <span style="color: #666666">1.1</span>], [<span style="color: #666666">0.57</span>, <span style="color: #666666">-0.1</span>], <span style="color: #BA2121">&quot;r--&quot;</span>, linewidth<span style="color: #666666">=3</span>)
-plt<span style="color: #666666">.</span>axis([<span style="color: #666666">-0.1</span>, <span style="color: #666666">1.1</span>, <span style="color: #666666">-0.1</span>, <span style="color: #666666">1.1</span>])
-    
-plt<span style="color: #666666">.</span>subplots_adjust(right<span style="color: #666666">=1</span>)
-
-plt<span style="color: #666666">.</span>show()
-
-
-x1_example <span style="color: #666666">=</span> X1D[<span style="color: #666666">3</span>, <span style="color: #666666">0</span>]
-<span style="color: #008000; font-weight: bold">for</span> landmark <span style="color: #AA22FF; font-weight: bold">in</span> (<span style="color: #666666">-2</span>, <span style="color: #666666">1</span>):
-    k <span style="color: #666666">=</span> gaussian_rbf(np<span style="color: #666666">.</span>array([[x1_example]]), np<span style="color: #666666">.</span>array([[landmark]]), gamma)
-    <span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;Phi(</span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">, </span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">) = </span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">&quot;</span><span style="color: #666666">.</span>format(x1_example, landmark, k))
-
-rbf_kernel_svm_clf <span style="color: #666666">=</span> Pipeline([
-        (<span style="color: #BA2121">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #BA2121">&quot;svm_clf&quot;</span>, SVC(kernel<span style="color: #666666">=</span><span style="color: #BA2121">&quot;rbf&quot;</span>, gamma<span style="color: #666666">=5</span>, C<span style="color: #666666">=0.001</span>))
-    ])
-rbf_kernel_svm_clf<span style="color: #666666">.</span>fit(X, y)
-
-
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC
-
-gamma1, gamma2 <span style="color: #666666">=</span> <span style="color: #666666">0.1</span>, <span style="color: #666666">5</span>
-C1, C2 <span style="color: #666666">=</span> <span style="color: #666666">0.001</span>, <span style="color: #666666">1000</span>
-hyperparams <span style="color: #666666">=</span> (gamma1, C1), (gamma1, C2), (gamma2, C1), (gamma2, C2)
-
-svm_clfs <span style="color: #666666">=</span> []
-<span style="color: #008000; font-weight: bold">for</span> gamma, C <span style="color: #AA22FF; font-weight: bold">in</span> hyperparams:
-    rbf_kernel_svm_clf <span style="color: #666666">=</span> Pipeline([
-            (<span style="color: #BA2121">&quot;scaler&quot;</span>, StandardScaler()),
-            (<span style="color: #BA2121">&quot;svm_clf&quot;</span>, SVC(kernel<span style="color: #666666">=</span><span style="color: #BA2121">&quot;rbf&quot;</span>, gamma<span style="color: #666666">=</span>gamma, C<span style="color: #666666">=</span>C))
-        ])
-    rbf_kernel_svm_clf<span style="color: #666666">.</span>fit(X, y)
-    svm_clfs<span style="color: #666666">.</span>append(rbf_kernel_svm_clf)
-
-plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">7</span>))
-
-<span style="color: #008000; font-weight: bold">for</span> i, svm_clf <span style="color: #AA22FF; font-weight: bold">in</span> <span style="color: #008000">enumerate</span>(svm_clfs):
-    plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">221</span> <span style="color: #666666">+</span> i)
-    plot_predictions(svm_clf, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-    plot_dataset(X, y, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-    gamma, C <span style="color: #666666">=</span> hyperparams[i]
-    plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">r&quot;$\gamma = </span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">, C = </span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">$&quot;</span><span style="color: #666666">.</span>format(gamma, C), fontsize<span style="color: #666666">=16</span>)
-
-plt<span style="color: #666666">.</span>show()
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
+<p>where \( G_{\mathrm{left/right}} \) measures the impurity of the left/right subset  and \( m_{\mathrm{left/right}} \)
+ is the number of instances in the left/right subset
+</p>
 
+<p>Once it has successfully split the training set in two, it splits the subsets using the same logic, then the subsubsets
+and so on, recursively. It stops recursing once it reaches the maximum depth (defined by the
+\( max\_depth \) hyperparameter), or if it cannot find a split that will reduce impurity. A few other
+hyperparameters control additional stopping conditions such as the \( min\_samples\_split \),
+\( min\_samples\_leaf \), \( min\_weight\_fraction\_leaf \), and \( max\_leaf\_nodes \).
+</p>
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -404,6 +374,12 @@ plt<span style="color: #666666">.</span>show()
   <li><a href="._week46-bs027.html">28</a></li>
   <li><a href="._week46-bs028.html">29</a></li>
   <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs030.html">31</a></li>
+  <li><a href="._week46-bs031.html">32</a></li>
+  <li><a href="._week46-bs032.html">33</a></li>
+  <li><a href="._week46-bs033.html">34</a></li>
+  <li><a href="">...</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs025.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs025.html b/doc/pub/week46/html/._week46-bs025.html
index 12c5e55e8..1c6fd8a01 100644
--- a/doc/pub/week46/html/._week46-bs025.html
+++ b/doc/pub/week46/html/._week46-bs025.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,26 +329,30 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0025"></a>
 <!-- !split -->
-<h2 id="mathematical-optimization-of-convex-functions" class="anchor">Mathematical optimization of convex functions </h2>
+<h2 id="the-cart-algorithm-for-regression" class="anchor">The CART algorithm for Regression </h2>
 
-<p>A mathematical (quadratic) optimization problem, or just optimization problem, has the form</p>
+<p>The CART algorithm for regression works is similar to the one for classification except that instead of trying to split the
+training set in a way that minimizes say the <b>gini</b> or <b>entropy</b> impurity, it now tries to split the training set in a way that minimizes our well-known mean-squared error (MSE). The cost function is now
+</p>
 $$
-\begin{align*}
-    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\boldsymbol{\lambda}^T\boldsymbol{P}\boldsymbol{\lambda}+\boldsymbol{q}^T\boldsymbol{\lambda},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{\lambda} \preceq \boldsymbol{h} \wedge  \boldsymbol{A}\boldsymbol{\lambda}=f.
-\end{align*}
+C(k,t_k) = \frac{m_{\mathrm{left}}}{m}\mathrm{MSE}_{\mathrm{left}}+ \frac{m_{\mathrm{right}}}{m}\mathrm{MSE}_{\mathrm{right}}.
 $$
 
-<p>subject to some constraints for say a selected set \( i=1,2,\dots, n \).
-In our case we are optimizing with respect to the Lagrangian multipliers \( \lambda_i \), and the
-vector \( \boldsymbol{\lambda}=[\lambda_1, \lambda_2,\dots, \lambda_n] \) is the optimization variable we are dealing with.
-</p>
+<p>Here the MSE for a specific node is defined as</p>
+$$
+\mathrm{MSE}_{\mathrm{node}}=\frac{1}{m_\mathrm{node}}\sum_{i\in \mathrm{node}}(\overline{y}_{\mathrm{node}}-y_i)^2,
+$$
 
-<p>In our case we are particularly interested in a class of optimization problems called convex optmization problems. 
-In our discussion on gradient descent methods we discussed at length the definition of a convex function. 
-</p>
+<p>with</p>
+$$
+\overline{y}_{\mathrm{node}}=\frac{1}{m_\mathrm{node}}\sum_{i\in \mathrm{node}}y_i,
+$$
 
-<p>Convex optimization problems play a central role in applied mathematics and we recommend strongly <a href="http://web.stanford.edu/~boyd/cvxbook/" target="_self">Boyd and Vandenberghe's text on the topics</a>.</p>
+<p>the mean value of all observations in a specific node.</p>
+
+<p>Without any regularization, the regression task for decision trees, 
+just like for classification tasks, is  prone to overfitting.
+</p>
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -212,6 +373,13 @@ In our discussion on gradient descent methods we discussed at length the definit
   <li><a href="._week46-bs027.html">28</a></li>
   <li><a href="._week46-bs028.html">29</a></li>
   <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs030.html">31</a></li>
+  <li><a href="._week46-bs031.html">32</a></li>
+  <li><a href="._week46-bs032.html">33</a></li>
+  <li><a href="._week46-bs033.html">34</a></li>
+  <li><a href="._week46-bs034.html">35</a></li>
+  <li><a href="">...</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs026.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs026.html b/doc/pub/week46/html/._week46-bs026.html
index f6b707cf8..50a6296c6 100644
--- a/doc/pub/week46/html/._week46-bs026.html
+++ b/doc/pub/week46/html/._week46-bs026.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,45 +329,14 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0026"></a>
 <!-- !split -->
-<h2 id="how-do-we-solve-these-problems" class="anchor">How do we solve these problems? </h2>
+<h2 id="why-binary-splits" class="anchor">Why binary splits? </h2>
 
-<p>If we use Python as programming language and wish to venture beyond
-<b>scikit-learn</b>, <b>tensorflow</b> and similar software which makes our
-lives so much easier, we need to dive into the wonderful world of
-quadratic programming. We can, if we wish, solve the minimization
-problem using say standard gradient methods or conjugate gradient
-methods. However, these methods tend to exhibit a rather slow
-converge. So, welcome to the promised land of quadratic programming.
+<p>It is custom to split to a tree uising binary splits. The reason is
+that multiway splits fragment the data too quickly, leaving
+insufficient data at the next level down.  Multiway splits can be
+achieved by a series of binary split and this is normally preferred.
 </p>
 
-<p>The functions we need are contained in the <a href="https://cvxopt.org/userguide/coneprog.html" target="_self">quadratic programming library CVXOPT</a> and we need to import it together with <b>numpy</b> as</p>
-
-
-<!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #f8f8f8">
-  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span>
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">cvxopt</span>
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
-
-<p>This will make our life much easier. You don't need t0 write your own optimizer.</p>
-
 <p>
 <!-- navigation buttons at the bottom of the page -->
 <ul class="pagination">
@@ -229,6 +355,14 @@ converge. So, welcome to the promised land of quadratic programming.
   <li><a href="._week46-bs027.html">28</a></li>
   <li><a href="._week46-bs028.html">29</a></li>
   <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs030.html">31</a></li>
+  <li><a href="._week46-bs031.html">32</a></li>
+  <li><a href="._week46-bs032.html">33</a></li>
+  <li><a href="._week46-bs033.html">34</a></li>
+  <li><a href="._week46-bs034.html">35</a></li>
+  <li><a href="._week46-bs035.html">36</a></li>
+  <li><a href="">...</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs027.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs027.html b/doc/pub/week46/html/._week46-bs027.html
index ad0081bfc..30f9e315f 100644
--- a/doc/pub/week46/html/._week46-bs027.html
+++ b/doc/pub/week46/html/._week46-bs027.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,90 +329,23 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0027"></a>
 <!-- !split -->
-<h2 id="a-simple-example" class="anchor">A simple example </h2>
+<h2 id="computing-a-tree-using-the-gini-index" class="anchor">Computing a Tree using the Gini Index </h2>
 
-<p>We remind ourselves about the general problem we want to solve</p>
-$$
-\begin{align*}
-    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}\boldsymbol{x}^T\boldsymbol{P}\boldsymbol{x}+\boldsymbol{q}^T\boldsymbol{x},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm} to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{x} \preceq \boldsymbol{h} \wedge  \boldsymbol{A}\boldsymbol{x}=f.
-\end{align*}
-$$
-
-<p>Let us show how to perform the optmization using a simple case. Assume we want to optimize the following problem</p>
-$$
-\begin{align*}
-    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}x^2+5x+3y \\ \nonumber
-    &\mathrm{subject}\hspace{0.5cm} \mathrm{to} \\ \nonumber
-    &x, y \geq 0 \\ \nonumber
-    &x+3y  \geq 15 \\ \nonumber
-    &2x+5y  \leq  100 \\ \nonumber
-    &3x+4y  \leq  80.  \\ \nonumber
-\end{align*}
-$$
-
-<p>The minimization problem can be rewritten in terms of vectors and matrices as (with \( x \) and \( y \) being the unknowns)</p>
-$$
-\frac{1}{2}\begin{bmatrix} x\\ y \end{bmatrix}^T   \begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}  \begin{bmatrix} x \\ y \end{bmatrix}  + \begin{bmatrix}3\\ 4  \end{bmatrix}^T \begin{bmatrix}x \\ y  \end{bmatrix}.  
-$$
-
-<p>Similarly, we can now set up the inequalities (we need to change \( \geq \) to \( \leq \) by multiplying with \( -1 \) on bot sides) as the following matrix-vector equation</p>
-$$
-\begin{bmatrix} -1 & 0 \\ 0 & -1 \\ -1 & -3 \\ 2 & 5 \\ 3 & 4\end{bmatrix}\begin{bmatrix} x \\ y\end{bmatrix} \preceq \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
-$$
-
-<p>We have collapsed all the inequalities into a single matrix \( \boldsymbol{G} \). We see also that our matrix </p>
-$$
-\boldsymbol{P} =\begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}
-$$
-
-<p>is clearly positive semi-definite (all eigenvalues larger or equal zero). 
-Finally, the vector \( \boldsymbol{h} \) is defined as 
-</p>
-$$
-\boldsymbol{h} = \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
-$$
-
-<p>Since we don't have any equalities the matrix \( \boldsymbol{A} \) is set to zero
-The following code solves the equations for us
+<p>Consider the following example with attributes/features and two
+possible outcomes (classes)  for each attribute.  Assume we wish to find some
+correlations between the average grade of a student as function of the
+number of hours studied and hours slept. We want also to correlate the
+grade in a given course with the general trend, whether the students
+recently has gotten grades below average or above.
 </p>
 
-<!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #f8f8f8">
-  <pre style="line-height: 125%;"><span style="color: #408080; font-style: italic"># Import the necessary packages</span>
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span>
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">cvxopt</span> <span style="color: #008000; font-weight: bold">import</span> matrix
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">cvxopt</span> <span style="color: #008000; font-weight: bold">import</span> solvers
-P <span style="color: #666666">=</span> matrix(numpy<span style="color: #666666">.</span>diag([<span style="color: #666666">1</span>,<span style="color: #666666">0</span>]), tc<span style="color: #666666">=</span><span style="color: #BA2121">&#39;d&#39;</span>)
-q <span style="color: #666666">=</span> matrix(numpy<span style="color: #666666">.</span>array([<span style="color: #666666">3</span>,<span style="color: #666666">4</span>]), tc<span style="color: #666666">=</span><span style="color: #BA2121">&#39;d&#39;</span>)
-G <span style="color: #666666">=</span> matrix(numpy<span style="color: #666666">.</span>array([[<span style="color: #666666">-1</span>,<span style="color: #666666">0</span>],[<span style="color: #666666">0</span>,<span style="color: #666666">-1</span>],[<span style="color: #666666">-1</span>,<span style="color: #666666">-3</span>],[<span style="color: #666666">2</span>,<span style="color: #666666">5</span>],[<span style="color: #666666">3</span>,<span style="color: #666666">4</span>]]), tc<span style="color: #666666">=</span><span style="color: #BA2121">&#39;d&#39;</span>)
-h <span style="color: #666666">=</span> matrix(numpy<span style="color: #666666">.</span>array([<span style="color: #666666">0</span>,<span style="color: #666666">0</span>,<span style="color: #666666">-15</span>,<span style="color: #666666">100</span>,<span style="color: #666666">80</span>]), tc<span style="color: #666666">=</span><span style="color: #BA2121">&#39;d&#39;</span>)
-<span style="color: #408080; font-style: italic"># Construct the QP, invoke solver</span>
-sol <span style="color: #666666">=</span> solvers<span style="color: #666666">.</span>qp(P,q,G,h)
-<span style="color: #408080; font-style: italic"># Extract optimal value and solution</span>
-sol[<span style="color: #BA2121">&#39;x&#39;</span>] 
-sol[<span style="color: #BA2121">&#39;primal objective&#39;</span>]
-<span style="color: #008000">print</span>(sol[<span style="color: #BA2121">&#39;x&#39;</span>] )
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
-
-
+<p>We have three features/attributes</p>
+<ol>
+<li> Trend of average grades before present course, classified as either below  or above  the average grade of the whole class</li>  
+<li> The number of hours studies, classified again as either higher (more than 3 hours per day) or lower . Here we have used a standard for one \( ECTS \) which is scaled to 25-30 hours of work for a semester which lasts 18 weeks, with 15 weeks of lectures and 3 weeks for exams, assuming a total of 30 ECTS per semester.</li> 
+<li> The number of hours slept as high for more than \( 8 \) hours and below  for less than 8 hours of sleep, classified again as either high or low</li>
+<li> The final grade whether it is above or below average</li>
+</ol>
 <p>
 <!-- navigation buttons at the bottom of the page -->
 <ul class="pagination">
@@ -273,6 +363,15 @@ sol[<span style="color: #BA2121">&#39;primal objective&#39;</span>]
   <li class="active"><a href="._week46-bs027.html">28</a></li>
   <li><a href="._week46-bs028.html">29</a></li>
   <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs030.html">31</a></li>
+  <li><a href="._week46-bs031.html">32</a></li>
+  <li><a href="._week46-bs032.html">33</a></li>
+  <li><a href="._week46-bs033.html">34</a></li>
+  <li><a href="._week46-bs034.html">35</a></li>
+  <li><a href="._week46-bs035.html">36</a></li>
+  <li><a href="._week46-bs036.html">37</a></li>
+  <li><a href="">...</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs028.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs028.html b/doc/pub/week46/html/._week46-bs028.html
index c2b2679e9..230585a1f 100644
--- a/doc/pub/week46/html/._week46-bs028.html
+++ b/doc/pub/week46/html/._week46-bs028.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,24 +329,29 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0028"></a>
 <!-- !split -->
-<h2 id="back-to-the-more-realistic-cases" class="anchor">Back to the more realistic cases </h2>
+<h2 id="the-table" class="anchor">The Table </h2>
 
-<p>We are now ready to return to our setup of the optmization problem for a more realistic case. Introducing the <b>slack</b> parameter \( C \) we have</p>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1K(\boldsymbol{x}_1,\boldsymbol{x}_1) & y_1y_2K(\boldsymbol{x}_1,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_1,\boldsymbol{x}_n) \\
-y_2y_1K(\boldsymbol{x}_2,\boldsymbol{x}_1) & y_2y_2K(\boldsymbol{x}_2,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_2,\boldsymbol{x}_n) \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1K(\boldsymbol{x}_n,\boldsymbol{x}_1) & y_ny_2K(\boldsymbol{x}_n\boldsymbol{x}_2) & \dots & \dots & y_ny_nK(\boldsymbol{x}_n,\boldsymbol{x}_n) \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{I}\boldsymbol{\lambda}, 
-$$
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
-With  the slack constants this leads to the additional constraint \( 0\leq \lambda_i \leq C \).
-</p>
-
-<p>Using the <b>CVXOPT</b> library, the matrix \( P \) would then be defined by the above matrix while the KKT conditions would all be collected by the matrix \( G \).</p>
+<div class="row">
+  <div class="col-xs-4">
+    <table class="dotable table-striped table-hover table-condensed">
+<thead>
+<tr><td align="center"><b>Grade Trend</b></td> <td align="center"><b>Hours slept</b></td> <td align="center"><b>Hours Studied</b></td> <td align="center"><b>Grade</b></td> </tr>
+</thead>
+<tbody>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   High           </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+</tbody>
+    </table>
+  </div> <!-- col-xs-4 -->
+</div> <!-- cell row -->
 
 <p>
 <!-- navigation buttons at the bottom of the page -->
@@ -207,6 +369,16 @@ With  the slack constants this leads to the additional constraint \( 0\leq \lamb
   <li><a href="._week46-bs027.html">28</a></li>
   <li class="active"><a href="._week46-bs028.html">29</a></li>
   <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs030.html">31</a></li>
+  <li><a href="._week46-bs031.html">32</a></li>
+  <li><a href="._week46-bs032.html">33</a></li>
+  <li><a href="._week46-bs033.html">34</a></li>
+  <li><a href="._week46-bs034.html">35</a></li>
+  <li><a href="._week46-bs035.html">36</a></li>
+  <li><a href="._week46-bs036.html">37</a></li>
+  <li><a href="._week46-bs037.html">38</a></li>
+  <li><a href="">...</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs029.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
diff --git a/doc/pub/week46/html/._week46-bs029.html b/doc/pub/week46/html/._week46-bs029.html
index 0f389e466..602ec4d5c 100644
--- a/doc/pub/week46/html/._week46-bs029.html
+++ b/doc/pub/week46/html/._week46-bs029.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -172,9 +329,20 @@ MathJax.Hub.Config({
 <p>&nbsp;</p><p>&nbsp;</p><p>&nbsp;</p> <!-- add vertical space -->
 <a name="part0029"></a>
 <!-- !split -->
-<h2 id="support-vector-machines-for-regression" class="anchor">Support vector machines for regression </h2>
+<h2 id="computing-the-various-gini-indices" class="anchor">Computing the various Gini Indices </h2>
+
+<p>In computations we will translate all classes into numbers. Being
+these binary classes, they can easily be split into ones and zeros.
+</p>
+
+<div class="panel panel-default">
+<div class="panel-body">
+<!-- subsequent paragraphs come in larger fonts, so start with a paragraph -->
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_self">See handwritten notes November 3</a>
+</div>
+</div>
+
 
-<p>Material will be added here.</p>
 <p>
 <!-- navigation buttons at the bottom of the page -->
 <ul class="pagination">
@@ -190,6 +358,18 @@ MathJax.Hub.Config({
   <li><a href="._week46-bs027.html">28</a></li>
   <li><a href="._week46-bs028.html">29</a></li>
   <li class="active"><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs030.html">31</a></li>
+  <li><a href="._week46-bs031.html">32</a></li>
+  <li><a href="._week46-bs032.html">33</a></li>
+  <li><a href="._week46-bs033.html">34</a></li>
+  <li><a href="._week46-bs034.html">35</a></li>
+  <li><a href="._week46-bs035.html">36</a></li>
+  <li><a href="._week46-bs036.html">37</a></li>
+  <li><a href="._week46-bs037.html">38</a></li>
+  <li><a href="._week46-bs038.html">39</a></li>
+  <li><a href="">...</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
+  <li><a href="._week46-bs030.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
 </div>  <!-- end container -->
diff --git a/doc/pub/week46/html/week46-bs.html b/doc/pub/week46/html/week46-bs.html
index ced4225cc..39a24a971 100644
--- a/doc/pub/week46/html/week46-bs.html
+++ b/doc/pub/week46/html/week46-bs.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <!-- Bootstrap style: bootstrap -->
 <!-- doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=default --html_admon=bootstrap_panel --html_output=week46-bs --no_mako -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -36,68 +36,190 @@ doconce format html week46.do.txt --html_style=bootstrap --pygments_html_style=d
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -125,42 +247,77 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week46-bs.html">Week 46: Support Vector Machines and Project 3.</a>
+    <a class="navbar-brand" href="week46-bs.html">Week 46: Decision Trees, Ensemble methods  and Random Forests</a>
   </div>
   <div class="navbar-collapse collapse navbar-responsive-collapse">
     <ul class="nav navbar-nav navbar-right">
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week46-bs001.html#overview-of-week-46" style="font-size: 80%;">Overview of week 46</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs002.html#support-vector-machines-overarching-aims" style="font-size: 80%;">Support Vector Machines, overarching aims</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs003.html#hyperplanes-and-all-that" style="font-size: 80%;">Hyperplanes and all that</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs004.html#what-is-a-hyperplane" style="font-size: 80%;">What is a hyperplane?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-p-dimensional-space-of-features" style="font-size: 80%;">A \( p \)-dimensional space of features</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs006.html#the-two-dimensional-case" style="font-size: 80%;">The two-dimensional case</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs007.html#getting-into-the-details" style="font-size: 80%;">Getting into the details</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs008.html#first-attempt-at-a-minimization-approach" style="font-size: 80%;">First attempt at a minimization approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs009.html#solving-the-equations" style="font-size: 80%;">Solving the equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs010.html#code-example" style="font-size: 80%;">Code Example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs011.html#problems-with-the-simpler-approach" style="font-size: 80%;">Problems with the Simpler Approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs012.html#a-better-approach" style="font-size: 80%;">A better approach</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs013.html#a-quick-reminder-on-lagrangian-multipliers" style="font-size: 80%;">A quick Reminder on Lagrangian Multipliers</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs014.html#adding-the-multiplier" style="font-size: 80%;">Adding the Multiplier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs015.html#setting-up-the-problem" style="font-size: 80%;">Setting up the Problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs017.html#the-last-steps" style="font-size: 80%;">The last steps</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs018.html#a-soft-classifier" style="font-size: 80%;">A soft classifier</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs019.html#soft-optmization-problem" style="font-size: 80%;">Soft optmization problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs020.html#kernels-and-non-linearity" style="font-size: 80%;">Kernels and non-linearity</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs021.html#the-equations" style="font-size: 80%;">The equations</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs022.html#the-problem-to-solve" style="font-size: 80%;">The problem to solve</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs023.html#different-kernels-and-mercer-s-theorem" style="font-size: 80%;">Different kernels and Mercer's theorem</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-moons-example" style="font-size: 80%;">The moons example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs025.html#mathematical-optimization-of-convex-functions" style="font-size: 80%;">Mathematical optimization of convex functions</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs026.html#how-do-we-solve-these-problems" style="font-size: 80%;">How do we solve these problems?</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs027.html#a-simple-example" style="font-size: 80%;">A simple example</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs028.html#back-to-the-more-realistic-cases" style="font-size: 80%;">Back to the more realistic cases</a></li>
-     <!-- navigation toc: --> <li><a href="._week46-bs029.html#support-vector-machines-for-regression" style="font-size: 80%;">Support vector machines for regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs001.html#plan-for-week-46" style="font-size: 80%;">Plan for week 46</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs002.html#decision-trees-overarching-aims" style="font-size: 80%;">Decision trees, overarching aims</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs003.html#basics-of-a-tree" style="font-size: 80%;">Basics of a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs004.html#a-sketch-of-a-tree-regression-problem" style="font-size: 80%;">A Sketch of a Tree, Regression problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs005.html#a-sketch-of-a-tree-classification-problem" style="font-size: 80%;">A Sketch of a Tree, Classification  problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs006.html#a-typical-decision-tree-with-its-pertinent-jargon-classification-problem" style="font-size: 80%;">A typical Decision Tree with its pertinent Jargon, Classification Problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs007.html#general-features" style="font-size: 80%;">General Features</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs008.html#how-do-we-set-it-up" style="font-size: 80%;">How do we set it up?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs009.html#decision-trees-and-regression" style="font-size: 80%;">Decision trees and Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs010.html#building-a-tree-regression" style="font-size: 80%;">Building a tree, regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs011.html#a-top-down-approach-recursive-binary-splitting" style="font-size: 80%;">A top-down approach, recursive binary splitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs012.html#making-a-tree" style="font-size: 80%;">Making a tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs013.html#pruning-the-tree" style="font-size: 80%;">Pruning the tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs014.html#cost-complexity-pruning" style="font-size: 80%;">Cost complexity pruning</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs015.html#schematic-regression-procedure" style="font-size: 80%;">Schematic Regression Procedure</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs016.html#a-classification-tree" style="font-size: 80%;">A Classification Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs017.html#growing-a-classification-tree" style="font-size: 80%;">Growing a classification tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs018.html#classification-tree-how-to-split-nodes" style="font-size: 80%;">Classification tree, how to split nodes</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs019.html#visualizing-the-tree-classification" style="font-size: 80%;">Visualizing the Tree, Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs020.html#visualizing-the-tree-the-moons" style="font-size: 80%;">Visualizing the Tree, The Moons</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs021.html#other-ways-of-visualizing-the-trees" style="font-size: 80%;">Other ways of visualizing the trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs022.html#printing-out-as-text" style="font-size: 80%;">Printing out as text</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs023.html#algorithms-for-setting-up-decision-trees" style="font-size: 80%;">Algorithms for Setting up Decision Trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs024.html#the-cart-algorithm-for-classification" style="font-size: 80%;">The CART algorithm for Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs025.html#the-cart-algorithm-for-regression" style="font-size: 80%;">The CART algorithm for Regression</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs026.html#why-binary-splits" style="font-size: 80%;">Why binary splits?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs027.html#computing-a-tree-using-the-gini-index" style="font-size: 80%;">Computing a Tree using the Gini Index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs028.html#the-table" style="font-size: 80%;">The Table</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs029.html#computing-the-various-gini-indices" style="font-size: 80%;">Computing the various Gini Indices</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs030.html#computing-the-various-gini-indices-hours-slept" style="font-size: 80%;">Computing the various Gini Indices, Hours slept</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs031.html#computing-the-various-gini-indices-hours-studied" style="font-size: 80%;">Computing the various Gini Indices, Hours studied</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs032.html#a-possible-code-using-scikit-learn" style="font-size: 80%;">A possible code using Scikit-Learn</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs033.html#further-example-computing-the-gini-index" style="font-size: 80%;">Further example: Computing the Gini index</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs034.html#simple-python-code-to-read-in-data-and-perform-classification" style="font-size: 80%;">Simple Python Code to read in Data and perform Classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs035.html#computing-the-gini-factor" style="font-size: 80%;">Computing the Gini Factor</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs036.html#regression-trees" style="font-size: 80%;">Regression trees</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs037.html#final-regressor-code" style="font-size: 80%;">Final regressor code</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs038.html#pros-and-cons-of-trees-pros" style="font-size: 80%;">Pros and cons of trees, pros</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs039.html#disadvantages" style="font-size: 80%;">Disadvantages</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs040.html#ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods" style="font-size: 80%;">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs041.html#an-overview-of-ensemble-methods" style="font-size: 80%;">An Overview of Ensemble Methods</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs042.html#why-voting" style="font-size: 80%;">Why Voting?</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs043.html#tossing-coins" style="font-size: 80%;">Tossing coins</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs044.html#standard-imports-first" style="font-size: 80%;">Standard imports first</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs045.html#simple-voting-example-head-or-tail" style="font-size: 80%;">Simple Voting Example, head or tail</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs046.html#using-the-voting-classifier" style="font-size: 80%;">Using the Voting Classifier</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs047.html#voting-and-bagging" style="font-size: 80%;">Voting and Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs048.html#bagging" style="font-size: 80%;">Bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs049.html#more-bagging" style="font-size: 80%;">More bagging</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs050.html#making-your-own-bootstrap-changing-the-level-of-the-decision-tree" style="font-size: 80%;">Making your own Bootstrap: Changing the Level of the Decision Tree</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs051.html#random-forests" style="font-size: 80%;">Random forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs052.html#random-forest-algorithm" style="font-size: 80%;">Random Forest Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs053.html#random-forests-compared-with-other-methods-on-the-cancer-data" style="font-size: 80%;">Random Forests Compared with other Methods on the Cancer Data</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs054.html#compare-bagging-on-trees-with-random-forests" style="font-size: 80%;">Compare  Bagging on Trees with Random Forests</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs055.html#boosting-a-bird-s-eye-view" style="font-size: 80%;">Boosting, a Bird's Eye View</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs056.html#what-is-boosting-additive-modelling-iterative-fitting" style="font-size: 80%;">What is boosting? Additive Modelling/Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs057.html#iterative-fitting-regression-and-squared-error-cost-function" style="font-size: 80%;">Iterative Fitting, Regression and Squared-error Cost Function</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs058.html#squared-error-example-and-iterative-fitting" style="font-size: 80%;">Squared-Error Example and Iterative Fitting</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs059.html#iterative-fitting-classification-and-adaboost" style="font-size: 80%;">Iterative Fitting, Classification and AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs060.html#adaptive-boosting-adaboost" style="font-size: 80%;">Adaptive Boosting, AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs061.html#building-up-adaboost" style="font-size: 80%;">Building up AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs062.html#adaptive-boosting-adaboost-basic-algorithm" style="font-size: 80%;">Adaptive boosting: AdaBoost, Basic Algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs063.html#basic-steps-of-adaboost" style="font-size: 80%;">Basic Steps of AdaBoost</a></li>
+     <!-- navigation toc: --> <li><a href="._week46-bs064.html#adaboost-examples" style="font-size: 80%;">AdaBoost Examples</a></li>
 
         </ul>
       </li>
@@ -174,7 +331,7 @@ MathJax.Hub.Config({
 <!-- ------------------- main content ---------------------- -->
 <div class="jumbotron">
 <center>
-<h1>Week 46: Support Vector Machines and Project 3. </h1>
+<h1>Week 46: Decision Trees, Ensemble methods  and Random Forests</h1>
 </center>  <!-- document title -->
 
 <!-- author(s): Morten Hjorth-Jensen -->
@@ -190,7 +347,7 @@ MathJax.Hub.Config({
 </center>
 <br>
 <center>
-<h4>Nov 19, 2022</h4>
+<h4>Week 46, November 13-17</h4>
 </center> <!-- date -->
 <br>
 
@@ -215,7 +372,7 @@ MathJax.Hub.Config({
   <li><a href="._week46-bs008.html">9</a></li>
   <li><a href="._week46-bs009.html">10</a></li>
   <li><a href="">...</a></li>
-  <li><a href="._week46-bs029.html">30</a></li>
+  <li><a href="._week46-bs064.html">65</a></li>
   <li><a href="._week46-bs001.html">&raquo;</a></li>
 </ul>
 <!-- ------------------- end of main content --------------- -->
@@ -229,7 +386,7 @@ MathJax.Hub.Config({
 </footer>
 -->
 <center style="font-size:80%">
-<!-- copyright --> &copy; 1999-2022, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
+<!-- copyright --> &copy; 1999-2023, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
 </center>
 </body>
 </html>
diff --git a/doc/pub/week46/html/week46-reveal.html b/doc/pub/week46/html/week46-reveal.html
index 2a371fec1..5a792f4c7 100644
--- a/doc/pub/week46/html/week46-reveal.html
+++ b/doc/pub/week46/html/week46-reveal.html
@@ -9,8 +9,8 @@ doconce format html week46-reveal.html week46-reveal reveal --html_slide_theme=b
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 
 <!-- reveal.js: https://lab.hakim.se/reveal-js/ -->
 
@@ -168,7 +168,7 @@ MathJax.Hub.Config({
 <section>
 <!-- ------------------- main content ---------------------- -->
 <center>
-<h1 style="text-align: center;">Week 46: Support Vector Machines and Project 3. </h1>
+<h1 style="text-align: center;">Week 46: Decision Trees, Ensemble methods  and Random Forests</h1>
 </center>  <!-- document title -->
 
 <!-- author(s): Morten Hjorth-Jensen -->
@@ -184,833 +184,162 @@ MathJax.Hub.Config({
 </center>
 <br>
 <center>
-<h4>Nov 19, 2022</h4>
+<h4>Week 46, November 13-17</h4>
 </center> <!-- date -->
 <br>
 
 
 <center style="font-size:80%">
-<!-- copyright --> &copy; 1999-2022, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
+<!-- copyright --> &copy; 1999-2023, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
 </center>
 </section>
 
 <section>
-<h2 id="overview-of-week-46">Overview of week 46 </h2>
-
-<ul>
-<p><li> <b>Thursday</b>: Support Vector Machines, discussion of project 3</li>
-<ul>
-
-<p><li> <a href="https://youtu.be/F3CkH-opbdY" target="_blank">Video of lecture</a></li>
-</ul>
-<p>
-<p><li> <b>Friday</b>: Support vector machines and start Principal Component Analysis</li>
-<ul>
-
-<p><li> <a href="https://youtu.be/BbupEEvMXtg" target="_blank">Video of lecture</a></li>
-</ul>
-<p>
-</ul>
-<p>
-<div class="alert alert-block alert-block alert-text-normal">
-<b>Reading</b>
-<p>
-<p>Reading recommendations:</p>
-<ol>
- <p><li> See lecture notes for week 46 at <a href="https://compphysics.github.io/MachineLearning/doc/web/course.html." target="_blank"><tt>https://compphysics.github.io/MachineLearning/doc/web/course.html.</tt></a></li>
- <p><li> Hastie et al chapter 12</li>
- <p><li> Bishop chapter 7.1 and 7.2</li>
-</ol>
-</div>
-
+<h2 id="plan-for-week-46">Plan for week 46 </h2>
 
 <div class="alert alert-block alert-block alert-text-normal">
-<b>Videos</b>
+<b>Active learning sessions on Tuesday and Wednesday</b>
 <p>
-<ol>
-<p><li> <a href="https://www.youtube.com/watch?v=efR1C6CvhmE&ab_channel=StatQuestwithJoshStarmer" target="_blank">Overview video on Support Vector Machines</a></li>
-<p><li> See also <a href="https://www.youtube.com/watch?v=N1vOgolbjSc&ab_channel=AliceZhao" target="_blank">this video</a>.</li>
-</ol>
+<ul>
+
+<p><li> Work and Discussion of project 2</li>
+
+<p><li> Discussion of project 3 as well</li>
+</ul>
+</div>
+  
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Material for the lecture on Thursday November 16, 2023</b>
+<p>
+<ul>
+
+<p><li> Thursday: Basics of decision trees, classification and regression algorithms and ensemble models</li>
+
+<p><li> Readings and Videos:</li>
+<ul>
+
+<p><li> These lecture notes</li>
+
+<p><li> <a href="https://youtu.be/" target="_blank">Video of lecture to be added</a></li>
+
+<p><li> <a href="https://www.youtube.com/watch?v=RmajweUFKvM&ab_channel=Simplilearn" target="_blank">Video on Decision trees</a></li>
+
+<p><li> Decision Trees: Geron's chapter 6 covers decision trees while ensemble models, voting and bagging are discussed in chapter 7. See also lecture from <a href="https://www.uio.no/studier/emner/matnat/math/STK-IN4300/h20/slides/lecture_7.pdf" target="_blank">STK-IN4300, lecture 7</a>. Chapter 9.2 of Hastie et al contains also a good discussion.</li>
+</ul>
+<p>
+</ul>
 </div>
 </section>
 
 <section>
-<h2 id="support-vector-machines-overarching-aims">Support Vector Machines, overarching aims  </h2>
+<h2 id="decision-trees-overarching-aims">Decision trees, overarching aims  </h2>
 
-<p>A Support Vector Machine (SVM) is a very powerful and versatile
-Machine Learning method, capable of performing linear or nonlinear
-classification, regression, and even outlier detection. It is one of
-the most popular models in Machine Learning, and anyone interested in
-Machine Learning should have it in their toolbox. SVMs are
-particularly well suited for classification of complex but small-sized or
-medium-sized datasets.  
+<p>We start here with the most basic algorithm, the so-called decision
+tree. With this basic algorithm we can in turn build more complex
+networks, spanning from homogeneous and heterogenous forests (bagging,
+random forests and more) to one of the most popular supervised
+algorithms nowadays, the extreme gradient boosting, or just
+XGBoost. But let us start with the simplest possible ingredient.
 </p>
 
-<p>The case with two well-separated classes only can be understood in an
-intuitive way in terms of lines in a two-dimensional space separating
-the two classes (see figure below).
+<p>Decision trees are supervised learning algorithms used for both,
+classification and regression tasks.
 </p>
 
-<p>The basic mathematics behind the SVM is however less familiar to most of us. 
-It relies on the definition of hyperplanes and the
-definition of a <b>margin</b> which separates classes (in case of
-classification problems) of variables. It is also used for regression
-problems.
+<p>The main idea of decision trees
+is to find those descriptive features which contain the most
+<b>information</b> regarding the target feature and then split the dataset
+along the values of these features such that the target feature values
+for the resulting underlying datasets are as pure as possible.
 </p>
 
-<p>With SVMs we distinguish between hard margin and soft margins. The
-latter introduces a so-called softening parameter to be discussed
-below.  We distinguish also between linear and non-linear
-approaches. The latter are the most frequent ones since it is rather
-unlikely that we can separate classes easily by say straight lines.
+<p>The descriptive features which reproduce best the target/output features are normally  said
+to be the most informative ones. The process of finding the <b>most
+informative</b> feature is done until we accomplish a stopping criteria
+where we then finally end up in so called <b>leaf nodes</b>. 
 </p>
 </section>
 
 <section>
-<h2 id="hyperplanes-and-all-that">Hyperplanes and all that </h2>
+<h2 id="basics-of-a-tree">Basics of a tree </h2>
 
-<p>The theory behind support vector machines (SVM hereafter) is based on
-the mathematical description of so-called hyperplanes. Let us start
-with a two-dimensional case. This will also allow us to introduce our
-first SVM examples. These will be tailored to the case of two specific
-classes, as displayed in the figure here based on the usage of the petal data.
+<p>A decision tree is typically divided into a <b>root node</b>, the <b>interior nodes</b>,
+and the final <b>leaf nodes</b> or just <b>leaves</b>. These entities are then connected by so-called <b>branches</b>.
 </p>
 
-<p>We assume here that our data set can be well separated into two
-domains, where a straight line does the job in the separating the two
-classes. Here the two classes are represented by either squares or
-circles.
-</p>
-
-<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #eeeedd">
-  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC, LinearSVC
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> SGDClassifier
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib</span>
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
-plt.rcParams[<span style="color: #CD5555">&#39;axes.labelsize&#39;</span>] = <span style="color: #B452CD">14</span>
-plt.rcParams[<span style="color: #CD5555">&#39;xtick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
-plt.rcParams[<span style="color: #CD5555">&#39;ytick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
-
-
-iris = datasets.load_iris()
-X = iris[<span style="color: #CD5555">&quot;data&quot;</span>][:, (<span style="color: #B452CD">2</span>, <span style="color: #B452CD">3</span>)]  <span style="color: #228B22"># petal length, petal width</span>
-y = iris[<span style="color: #CD5555">&quot;target&quot;</span>]
-
-setosa_or_versicolor = (y == <span style="color: #B452CD">0</span>) | (y == <span style="color: #B452CD">1</span>)
-X = X[setosa_or_versicolor]
-y = y[setosa_or_versicolor]
-
-
-
-C = <span style="color: #B452CD">5</span>
-alpha = <span style="color: #B452CD">1</span> / (C * <span style="color: #658b00">len</span>(X))
-
-lin_clf = LinearSVC(loss=<span style="color: #CD5555">&quot;hinge&quot;</span>, C=C, random_state=<span style="color: #B452CD">42</span>)
-svm_clf = SVC(kernel=<span style="color: #CD5555">&quot;linear&quot;</span>, C=C)
-sgd_clf = SGDClassifier(loss=<span style="color: #CD5555">&quot;hinge&quot;</span>, learning_rate=<span style="color: #CD5555">&quot;constant&quot;</span>, eta0=<span style="color: #B452CD">0.001</span>, alpha=alpha,
-                        max_iter=<span style="color: #B452CD">100000</span>, random_state=<span style="color: #B452CD">42</span>)
-
-scaler = StandardScaler()
-X_scaled = scaler.fit_transform(X)
-
-lin_clf.fit(X_scaled, y)
-svm_clf.fit(X_scaled, y)
-sgd_clf.fit(X_scaled, y)
-
-<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;LinearSVC:                   &quot;</span>, lin_clf.intercept_, lin_clf.coef_)
-<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;SVC:                         &quot;</span>, svm_clf.intercept_, svm_clf.coef_)
-<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;SGDClassifier(alpha={:.5f}):&quot;</span>.format(sgd_clf.alpha), sgd_clf.intercept_, sgd_clf.coef_)
-
-<span style="color: #228B22"># Compute the slope and bias of each decision boundary</span>
-w1 = -lin_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>]/lin_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-b1 = -lin_clf.intercept_[<span style="color: #B452CD">0</span>]/lin_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-w2 = -svm_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>]/svm_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-b2 = -svm_clf.intercept_[<span style="color: #B452CD">0</span>]/svm_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-w3 = -sgd_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>]/sgd_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-b3 = -sgd_clf.intercept_[<span style="color: #B452CD">0</span>]/sgd_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-
-<span style="color: #228B22"># Transform the decision boundary lines back to the original scale</span>
-line1 = scaler.inverse_transform([[-<span style="color: #B452CD">10</span>, -<span style="color: #B452CD">10</span> * w1 + b1], [<span style="color: #B452CD">10</span>, <span style="color: #B452CD">10</span> * w1 + b1]])
-line2 = scaler.inverse_transform([[-<span style="color: #B452CD">10</span>, -<span style="color: #B452CD">10</span> * w2 + b2], [<span style="color: #B452CD">10</span>, <span style="color: #B452CD">10</span> * w2 + b2]])
-line3 = scaler.inverse_transform([[-<span style="color: #B452CD">10</span>, -<span style="color: #B452CD">10</span> * w3 + b3], [<span style="color: #B452CD">10</span>, <span style="color: #B452CD">10</span> * w3 + b3]])
-
-<span style="color: #228B22"># Plot all three decision boundaries</span>
-plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
-plt.plot(line1[:, <span style="color: #B452CD">0</span>], line1[:, <span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;k:&quot;</span>, label=<span style="color: #CD5555">&quot;LinearSVC&quot;</span>)
-plt.plot(line2[:, <span style="color: #B452CD">0</span>], line2[:, <span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;b--&quot;</span>, linewidth=<span style="color: #B452CD">2</span>, label=<span style="color: #CD5555">&quot;SVC&quot;</span>)
-plt.plot(line3[:, <span style="color: #B452CD">0</span>], line3[:, <span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;r-&quot;</span>, label=<span style="color: #CD5555">&quot;SGDClassifier&quot;</span>)
-plt.plot(X[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">1</span>], X[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;bs&quot;</span>) <span style="color: #228B22"># label=&quot;Iris-Versicolor&quot;</span>
-plt.plot(X[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">0</span>], X[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">0</span>], <span style="color: #CD5555">&quot;yo&quot;</span>) <span style="color: #228B22"># label=&quot;Iris-Setosa&quot;</span>
-plt.xlabel(<span style="color: #CD5555">&quot;Petal length&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
-plt.ylabel(<span style="color: #CD5555">&quot;Petal width&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
-plt.legend(loc=<span style="color: #CD5555">&quot;upper center&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
-plt.axis([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">5.5</span>, <span style="color: #B452CD">0</span>, <span style="color: #B452CD">2</span>])
-
-plt.show()
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
-</section>
-
-<section>
-<h2 id="what-is-a-hyperplane">What is a hyperplane? </h2>
-
-<p>The aim of the SVM algorithm is to find a hyperplane in a
-\( p \)-dimensional space, where \( p \) is the number of features that
-distinctly classifies the data points.
-</p>
-
-<p>In a \( p \)-dimensional space, a hyperplane is what we call an affine subspace of dimension of \( p-1 \).
-As an example, in two dimension, a hyperplane is simply as straight line while in three dimensions it is 
-a two-dimensional subspace, or stated simply, a plane. 
-</p>
-
-<p>In two dimensions, with the variables \( x_1 \) and \( x_2 \), the hyperplane is defined as</p>
-<p>&nbsp;<br>
-$$
-b+w_1x_1+w_2x_2=0,
-$$
-<p>&nbsp;<br>
-
-<p>where \( b \) is the intercept and \( w_1 \) and \( w_2 \) define the elements of a vector orthogonal to the line 
-\( b+w_1x_1+w_2x_2=0 \). 
-In two dimensions we define the vectors \( \boldsymbol{x} =[x1,x2] \) and \( \boldsymbol{w}=[w1,w2] \). 
-We can then rewrite the above equation as 
-</p>
-
-<p>&nbsp;<br>
-$$
-\boldsymbol{x}^T\boldsymbol{w}+b=0.
-$$
-<p>&nbsp;<br>
-</section>
-
-<section>
-<h2 id="a-p-dimensional-space-of-features">A \( p \)-dimensional space of features </h2>
-
-<p>We limit ourselves to two classes of outputs \( y_i \) and assign these classes the values \( y_i = \pm 1 \). 
-In a \( p \)-dimensional space of say \( p \) features we have a hyperplane defines as 
-</p>
-<p>&nbsp;<br>
-$$
-b+wx_1+w_2x_2+\dots +w_px_p=0.
-$$
-<p>&nbsp;<br>
-
-<p>If we define a 
-matrix \( \boldsymbol{X}=\left[\boldsymbol{x}_1,\boldsymbol{x}_2,\dots, \boldsymbol{x}_p\right] \)
-of dimension \( n\times p \), where \( n \) represents the observations for each feature and each vector \( x_i \) is a column vector of the matrix \( \boldsymbol{X} \), 
-</p>
-<p>&nbsp;<br>
-$$
-\boldsymbol{x}_i = \begin{bmatrix} x_{i1} \\ x_{i2} \\ \dots \\ \dots \\ x_{ip} \end{bmatrix}.
-$$
-<p>&nbsp;<br>
-
-<p>If the above condition is not met for a given vector \( \boldsymbol{x}_i \) we have </p>
-<p>&nbsp;<br>
-$$
-b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} >0,
-$$
-<p>&nbsp;<br>
-
-<p>if our output \( y_i=1 \).
-In this case we say that \( \boldsymbol{x}_i \) lies on one of the sides of the hyperplane and if 
-</p>
-<p>&nbsp;<br>
-$$
-b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} < 0,
-$$
-<p>&nbsp;<br>
-
-<p>for the class of observations \( y_i=-1 \), 
-then \( \boldsymbol{x}_i \) lies on the other side. 
-</p>
-
-<p>Equivalently, for the two classes of observations we have </p>
-<p>&nbsp;<br>
-$$
-y_i\left(b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip}\right) > 0. 
-$$
-<p>&nbsp;<br>
-
-<p>When we try to separate hyperplanes, if it exists, we can use it to construct a natural classifier: a test observation is assigned a given class depending on which side of the hyperplane it is located.</p>
-</section>
-
-<section>
-<h2 id="the-two-dimensional-case">The two-dimensional case </h2>
-
-<p>Let us try to develop our intuition about SVMs by limiting ourselves to a two-dimensional
-plane.  To separate the two classes of data points, there are many
-possible lines (hyperplanes if you prefer a more strict naming)  
-that could be chosen. Our objective is to find a
-plane that has the maximum margin, i.e the maximum distance between
-data points of both classes. Maximizing the margin distance provides
-some reinforcement so that future data points can be classified with
-more confidence.
-</p>
-
-<p>What a linear classifier attempts to accomplish is to split the
-feature space into two half spaces by placing a hyperplane between the
-data points.  This hyperplane will be our decision boundary.  All
-points on one side of the plane will belong to class one and all points
-on the other side of the plane will belong to the second class two.
-</p>
-
-<p>Unfortunately there are many ways in which we can place a hyperplane
-to divide the data.  Below is an example of two candidate hyperplanes
-for our data sample.
+<p>The leaf nodes
+contain the predictions we will make for new query instances presented
+to our trained model. This is possible since the model has 
+learned the underlying structure of the training data and hence can,
+given some assumptions, make predictions about the target feature value
+(class) of unseen query instances.
 </p>
 </section>
 
 <section>
-<h2 id="getting-into-the-details">Getting into the details </h2>
+<h2 id="a-sketch-of-a-tree-regression-problem">A Sketch of a Tree, Regression problem  </h2>
 
-<p>Let us define the function</p>
-<p>&nbsp;<br>
-$$
-f(x) = \boldsymbol{w}^T\boldsymbol{x}+b = 0,
-$$
-<p>&nbsp;<br>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
 
-<p>as the function that determines the line \( L \) that separates two classes (our two features), see the figure here. </p>
-
-<p>Any point defined by \( \boldsymbol{x}_i \) and \( \boldsymbol{x}_2 \) on the line \( L \) will satisfy \( \boldsymbol{w}^T(\boldsymbol{x}_1-\boldsymbol{x}_2)=0 \). </p>
-
-<p>The signed distance \( \delta \) from any point defined by a vector \( \boldsymbol{x} \) and a point \( \boldsymbol{x}_0 \) on the line \( L \) is then</p>
-<p>&nbsp;<br>
-$$
-\delta = \frac{1}{\vert\vert \boldsymbol{w}\vert\vert}(\boldsymbol{w}^T\boldsymbol{x}+b).
-$$
-<p>&nbsp;<br>
+<!-- FIGURE: [DataFiles/Regsimpletree.png, width=600 frac=0.8] -->
 </section>
 
 <section>
-<h2 id="first-attempt-at-a-minimization-approach">First attempt at a minimization approach </h2>
+<h2 id="a-sketch-of-a-tree-classification-problem">A Sketch of a Tree, Classification  problem  </h2>
 
-<p>How do we find the parameter \( b \) and the vector \( \boldsymbol{w} \)? What we could
-do is to define a cost function which now contains the set of all
-misclassified points \( M \) and attempt to minimize this function
-</p>
-
-<p>&nbsp;<br>
-$$
-C(\boldsymbol{w},b) = -\sum_{i\in M} y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b).
-$$
-<p>&nbsp;<br>
-
-<p>We could now for example define all values \( y_i =1 \) as misclassified in case we have \( \boldsymbol{w}^T\boldsymbol{x}_i+b < 0 \) and the opposite if we have \( y_i=-1 \). Taking the derivatives gives us</p>
-<p>&nbsp;<br>
-$$
-\frac{\partial C}{\partial b} = -\sum_{i\in M} y_i,
-$$
-<p>&nbsp;<br>
-
-<p>and </p>
-<p>&nbsp;<br>
-$$
-\frac{\partial C}{\partial \boldsymbol{w}} = -\sum_{i\in M} y_ix_i.
-$$
-<p>&nbsp;<br>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+<!-- FIGURE: [DataFiles/Classimpletree.png, width=600 frac=0.8] -->
 </section>
 
 <section>
-<h2 id="solving-the-equations">Solving the equations </h2>
+<h2 id="a-typical-decision-tree-with-its-pertinent-jargon-classification-problem">A typical Decision Tree with its pertinent Jargon, Classification Problem </h2>
 
-<p>We can now use the Newton-Raphson method or different variants of the gradient descent family (from plain gradient descent to various stochastic gradient descent approaches) to solve the equations</p>
-<p>&nbsp;<br>
-$$
-b \leftarrow b +\eta \frac{\partial C}{\partial b},
-$$
-<p>&nbsp;<br>
+<br/><br/>
+<center>
+<p><img src="DataFiles/cancer.png" width="600" align="bottom"></p>
+</center>
+<br/><br/>
 
-<p>and</p>
-<p>&nbsp;<br>
-$$
-\boldsymbol{w} \leftarrow \boldsymbol{w} +\eta \frac{\partial C}{\partial \boldsymbol{w}},
-$$
-<p>&nbsp;<br>
-
-<p>where \( \eta \) is our by now well-known learning rate. </p>
+<p>This tree was produced using the Wisconsin cancer data (discussed here as well, see code examples below) using <b>Scikit-Learn</b>'s decision tree classifier. Here we have used the so-called <b>gini</b> index (see below) to split the various branches.</p>
 </section>
 
 <section>
-<h2 id="code-example">Code Example </h2>
+<h2 id="general-features">General Features </h2>
 
-<p>The equations we discussed above can be coded rather easily (the
-framework is similar to what we developed for logistic
-regression). We are going to set up a simple case with two classes only and we want to find a line which separates them the best possible way.
-</p>
+<p>The overarching approach to decision trees is a top-down approach.</p>
 
-<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #eeeedd">
-  <pre style="font-size: 80%; line-height: 125%;">
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
-</section>
-
-<section>
-<h2 id="problems-with-the-simpler-approach">Problems with the Simpler Approach </h2>
-
-<p>There are however problems with this approach, although it looks
-pretty straightforward to implement. When running the above code, we see that we can easily end up with many diffeent lines which separate the two classes.
-</p>
-
-<p>For small
-gaps between the entries, we may also end up needing many iterations
-before the solutions converge and if the data cannot be separated
-properly into two distinct classes, we may not experience a converge
-at all.
+<ul>
+<p><li> A leaf provides the classification of a given instance.</li>
+<p><li> A node specifies a test of some attribute of the instance.</li>
+<p><li> A branch corresponds to a possible values of an attribute.</li>
+<p><li> An instance is classified by starting at the root node of the tree, testing the attribute specified by this node, then moving down the tree branch corresponding to the value of the attribute in the given example.</li>
+</ul>
+<p>
+<p>This process is then repeated for the subtree rooted at the new
+node.
 </p>
 </section>
 
 <section>
-<h2 id="a-better-approach">A better approach </h2>
+<h2 id="how-do-we-set-it-up">How do we set it up? </h2>
 
-<p>A better approach is rather to try to define a large margin between
-the two classes (if they are well separated from the beginning).
+<p>In simplified terms, the process of training a decision tree and
+predicting the target features of query instances is as follows:
 </p>
 
-<p>Thus, we wish to find a margin \( M \) with \( \boldsymbol{w} \) normalized to
-\( \vert\vert \boldsymbol{w}\vert\vert =1 \) subject to the condition
-</p>
-
-<p>&nbsp;<br>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, p. 
-$$
-<p>&nbsp;<br>
-
-<p>All points are thus at a signed distance from the decision boundary defined by the line \( L \). The parameters \( b \) and \( w_1 \) and \( w_2 \) define this line. </p>
-
-<p>We seek thus the largest value \( M \) defined by</p>
-<p>&nbsp;<br>
-$$
-\frac{1}{\vert \vert \boldsymbol{w}\vert\vert}y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, n, 
-$$
-<p>&nbsp;<br>
-
-<p>or just </p>
-<p>&nbsp;<br>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M\vert \vert \boldsymbol{w}\vert\vert \hspace{0.1cm}\forall i. 
-$$
-<p>&nbsp;<br>
-
-<p>If we scale the equation so that \( \vert \vert \boldsymbol{w}\vert\vert = 1/M \), we have to find the minimum of 
-\( \boldsymbol{w}^T\boldsymbol{w}=\vert \vert \boldsymbol{w}\vert\vert \) (the norm) subject to the condition
-</p>
-<p>&nbsp;<br>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq 1 \hspace{0.1cm}\forall i.
-$$
-<p>&nbsp;<br>
-
-<p>We have thus defined our margin as the invers of the norm of
-\( \boldsymbol{w} \). We want to minimize the norm in order to have a as large as
-possible margin \( M \). Before we proceed, we need to remind ourselves
-about Lagrangian multipliers.
-</p>
-</section>
-
-<section>
-<h2 id="a-quick-reminder-on-lagrangian-multipliers">A quick Reminder on Lagrangian Multipliers </h2>
-
-<p>Consider a function of three independent variables \( f(x,y,z) \) . For the function \( f \) to be an
-extreme we have
-</p>
-<p>&nbsp;<br>
-$$
-df=0.
-$$
-<p>&nbsp;<br>
-
-<p>A necessary and sufficient condition is</p>
-<p>&nbsp;<br>
-$$
-\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
-$$
-<p>&nbsp;<br>
-
-<p>due to</p>
-<p>&nbsp;<br>
-$$
-df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz.
-$$
-<p>&nbsp;<br>
-
-<p>In many problems the variables \( x,y,z \) are often subject to constraints (such as those above for the margin)
-so that they are no longer all independent. It is possible at least in principle to use each 
-constraint to eliminate one variable
-and to proceed with a new and smaller set of independent varables.
-</p>
-
-<p>The use of so-called Lagrangian  multipliers is an alternative technique  when the elimination
-of variables is incovenient or undesirable.  Assume that we have an equation of constraint on 
-the variables \( x,y,z \)
-</p>
-<p>&nbsp;<br>
-$$
-\phi(x,y,z) = 0,
-$$
-<p>&nbsp;<br>
-
-<p> resulting in</p>
-<p>&nbsp;<br>
-$$
-d\phi = \frac{\partial \phi}{\partial x}dx+\frac{\partial \phi}{\partial y}dy+\frac{\partial \phi}{\partial z}dz =0.
-$$
-<p>&nbsp;<br>
-
-<p>Now we cannot set anymore</p>
-<p>&nbsp;<br>
-$$
-\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
-$$
-<p>&nbsp;<br>
-
-<p>if \( df=0 \) is wanted
-because there are now only two independent variables!  Assume \( x \) and \( y \) are the independent 
-variables.
-Then \( dz \) is no longer arbitrary.
-</p>
-</section>
-
-<section>
-<h2 id="adding-the-multiplier">Adding the Multiplier </h2>
-
-<p>However, we can add to</p>
-<p>&nbsp;<br>
-$$
-df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz,
-$$
-<p>&nbsp;<br>
-
-<p>a multiplum of \( d\phi \), viz. \( \lambda d\phi \), resulting  in</p>
-<p>&nbsp;<br>
-$$
-df+\lambda d\phi = (\frac{\partial f}{\partial z}+\lambda
-\frac{\partial \phi}{\partial x})dx+(\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y})dy+
-(\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z})dz =0.
-$$
-<p>&nbsp;<br>
-
-<p>Our multiplier is chosen so that</p>
-<p>&nbsp;<br>
-$$
-\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z} =0.
-$$
-<p>&nbsp;<br>
-
-<p>We need to remember that we took \( dx \) and \( dy \) to be arbitrary and thus we must have</p>
-<p>&nbsp;<br>
-$$
-\frac{\partial f}{\partial x}+\lambda\frac{\partial \phi}{\partial x} =0,
-$$
-<p>&nbsp;<br>
-
-<p>and</p>
-<p>&nbsp;<br>
-$$
-\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y} =0.
-$$
-<p>&nbsp;<br>
-
-<p>When all these equations are satisfied, \( df=0 \).  We have four unknowns, \( x,y,z \) and
-\( \lambda \). Actually we want only \( x,y,z \), \( \lambda \) needs not to be determined, 
-it is therefore often called
-Lagrange's undetermined multiplier.
-If we have a set of constraints \( \phi_k \) we have the equations
-</p>
-<p>&nbsp;<br>
-$$
-\frac{\partial f}{\partial x_i}+\sum_k\lambda_k\frac{\partial \phi_k}{\partial x_i} =0.
-$$
-<p>&nbsp;<br>
-</section>
-
-<section>
-<h2 id="setting-up-the-problem">Setting up the Problem </h2>
-<p>In order to solve the above problem, we define the following Lagrangian function to be minimized </p>
-<p>&nbsp;<br>
-$$
-{\cal L}(\lambda,b,\boldsymbol{w})=\frac{1}{2}\boldsymbol{w}^T\boldsymbol{w}-\sum_{i=1}^n\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)-1\right],
-$$
-<p>&nbsp;<br>
-
-<p>where \( \lambda_i \) is a so-called Lagrange multiplier subject to the condition \( \lambda_i \geq 0 \).</p>
-
-<p>Taking the derivatives  with respect to \( b \) and \( \boldsymbol{w} \) we obtain </p>
-<p>&nbsp;<br>
-$$
-\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
-$$
-<p>&nbsp;<br>
-
-<p>and </p>
-<p>&nbsp;<br>
-$$
-\frac{\partial {\cal L}}{\partial \boldsymbol{w}} = 0 = \boldsymbol{w}-\sum_{i} \lambda_iy_i\boldsymbol{x}_i.
-$$
-<p>&nbsp;<br>
-
-<p>Inserting these constraints into the equation for \( {\cal L} \) we obtain</p>
-<p>&nbsp;<br>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-<p>&nbsp;<br>
-
-<p>subject to the constraints \( \lambda_i\geq 0 \) and \( \sum_i\lambda_iy_i=0 \). 
-We must in addition satisfy the <a href="https://en.wikipedia.org/wiki/Karush%E2%80%93Kuhn%E2%80%93Tucker_conditions" target="_blank">Karush-Kuhn-Tucker</a> (KKT) condition
-</p>
-<p>&nbsp;<br>
-$$
-\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -1\right] \hspace{0.1cm}\forall i.
-$$
-<p>&nbsp;<br>
-
 <ol>
-<p><li> If \( \lambda_i > 0 \), then \( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1 \) and we say that \( x_i \) is on the boundary.</li>
-<p><li> If \( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)> 1 \), we say \( x_i \) is not on the boundary and we set \( \lambda_i=0 \).</li> 
+<p><li> Present a dataset containing of a number of training instances characterized by a number of descriptive features and a target feature</li>
+<p><li> Train the decision tree model by continuously splitting the target feature along the values of the descriptive features using a measure of information gain during the training process</li>
+<p><li> Grow the tree until we accomplish a stopping criteria create leaf nodes which represent the <em>predictions</em> we want to make for new query instances</li>
+<p><li> Show query instances to the tree and run down the tree until we arrive at leaf nodes</li>
 </ol>
 <p>
-<p>When \( \lambda_i > 0 \), the vectors \( \boldsymbol{x}_i \) are called support vectors. They are the vectors closest to the line (or hyperplane) and define the margin \( M \). </p>
+<p>Then we are essentially done!</p>
 </section>
 
 <section>
-<h2 id="the-problem-to-solve">The problem to solve </h2>
-
-<p>We can rewrite </p>
-<p>&nbsp;<br>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-<p>&nbsp;<br>
-
-<p>and its constraints in terms of a matrix-vector problem where we minimize w.r.t. \( \lambda \) the following problem</p>
-<p>&nbsp;<br>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1\boldsymbol{x}_1^T\boldsymbol{x}_1 & y_1y_2\boldsymbol{x}_1^T\boldsymbol{x}_2 & \dots & \dots & y_1y_n\boldsymbol{x}_1^T\boldsymbol{x}_n \\
-y_2y_1\boldsymbol{x}_2^T\boldsymbol{x}_1 & y_2y_2\boldsymbol{x}_2^T\boldsymbol{x}_2 & \dots & \dots & y_1y_n\boldsymbol{x}_2^T\boldsymbol{x}_n \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1\boldsymbol{x}_n^T\boldsymbol{x}_1 & y_ny_2\boldsymbol{x}_n^T\boldsymbol{x}_2 & \dots & \dots & y_ny_n\boldsymbol{x}_n^T\boldsymbol{x}_n \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{1}\boldsymbol{\lambda}, 
-$$
-<p>&nbsp;<br>
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
-</p>
-</section>
-
-<section>
-<h2 id="the-last-steps">The last steps </h2>
-
-<p>Solving the above problem, yields the values of \( \lambda_i \).
-To find the coefficients of your hyperplane we need simply to compute 
-</p>
-<p>&nbsp;<br>
-$$
-\boldsymbol{w}=\sum_{i} \lambda_iy_i\boldsymbol{x}_i.
-$$
-<p>&nbsp;<br>
-
-<p>With our vector \( \boldsymbol{w} \) we can in turn find the value of the intercept \( b \) (here in two dimensions) via </p>
-<p>&nbsp;<br>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1,
-$$
-<p>&nbsp;<br>
-
-<p>resulting in</p>
-<p>&nbsp;<br>
-$$
-b = \frac{1}{y_i}-\boldsymbol{w}^T\boldsymbol{x}_i,
-$$
-<p>&nbsp;<br>
-
-<p>or if we write it out in terms of the support vectors only, with \( N_s \) being their number,  we have</p>
-<p>&nbsp;<br>
-$$
-b = \frac{1}{N_s}\sum_{j\in N_s}\left(y_j-\sum_{i=1}^n\lambda_iy_i\boldsymbol{x}_i^T\boldsymbol{x}_j\right).
-$$
-<p>&nbsp;<br>
-
-<p>With our hyperplane coefficients we can use our classifier to assign any observation by simply using </p>
-<p>&nbsp;<br>
-$$
-y_i = \mathrm{sign}(\boldsymbol{w}^T\boldsymbol{x}_i+b).
-$$
-<p>&nbsp;<br>
-
-<p>Below we discuss how to find the optimal values of \( \lambda_i \). Before we proceed however, we discuss now the so-called soft classifier. </p>
-</section>
-
-<section>
-<h2 id="a-soft-classifier">A soft classifier </h2>
-
-<p>Till now, the margin is strictly defined by the support vectors. This defines what is called a hard classifier, that is the margins are well defined.</p>
-
-<p>Suppose now that classes overlap in feature space, as shown in the
-figure here. One way to deal with this problem before we define the
-so-called <b>kernel approach</b>, is to allow a kind of slack in the sense
-that we allow some points to be on the wrong side of the margin.
-</p>
-
-<p>We introduce thus the so-called <b>slack</b> variables \( \boldsymbol{\xi} =[\xi_1,x_2,\dots,x_n] \) and 
-modify our previous equation
-</p>
-<p>&nbsp;<br>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1,
-$$
-<p>&nbsp;<br>
-
-<p>to </p>
-<p>&nbsp;<br>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1-\xi_i,
-$$
-<p>&nbsp;<br>
-
-<p>with the requirement \( \xi_i\geq 0 \). The total violation is now \( \sum_i\xi \). 
-The value \( \xi_i \) in the constraint the last constraint corresponds to the  amount by which the prediction
-\( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1 \) is on the wrong side of its margin. Hence by bounding the sum \( \sum_i \xi_i \),
-we bound the total amount by which predictions fall on the wrong side of their margins.
-</p>
-
-<p>Misclassifications occur when \( \xi_i > 1 \). Thus bounding the total sum by some value \( C \) bounds in turn the total number of
-misclassifications.
-</p>
-</section>
-
-<section>
-<h2 id="soft-optmization-problem">Soft optmization problem </h2>
-
-<p>This has in turn the consequences that we change our optmization problem to finding the minimum of </p>
-<p>&nbsp;<br>
-$$
-{\cal L}=\frac{1}{2}\boldsymbol{w}^T\boldsymbol{w}-\sum_{i=1}^n\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)-(1-\xi_)\right]+C\sum_{i=1}^n\xi_i-\sum_{i=1}^n\gamma_i\xi_i,
-$$
-<p>&nbsp;<br>
-
-<p>subject to </p>
-<p>&nbsp;<br>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1-\xi_i \hspace{0.1cm}\forall i,
-$$
-<p>&nbsp;<br>
-
-<p>with the requirement \( \xi_i\geq 0 \).</p>
-
-<p>Taking the derivatives  with respect to \( b \) and \( \boldsymbol{w} \) we obtain </p>
-<p>&nbsp;<br>
-$$
-\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
-$$
-<p>&nbsp;<br>
-
-<p>and </p>
-<p>&nbsp;<br>
-$$
-\frac{\partial {\cal L}}{\partial \boldsymbol{w}} = 0 = \boldsymbol{w}-\sum_{i} \lambda_iy_i\boldsymbol{x}_i, 
-$$
-<p>&nbsp;<br>
-
-<p>and</p>
-<p>&nbsp;<br>
-$$
-\lambda_i = C-\gamma_i \hspace{0.1cm}\forall i.
-$$
-<p>&nbsp;<br>
-
-<p>Inserting these constraints into the equation for \( {\cal L} \) we obtain the same equation as before</p>
-<p>&nbsp;<br>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-<p>&nbsp;<br>
-
-<p>but now subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \) and \( 0\leq\lambda_i \leq C \). 
-We must in addition satisfy the Karush-Kuhn-Tucker condition which now reads
-</p>
-<p>&nbsp;<br>
-$$
-\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -(1-\xi_)\right]=0 \hspace{0.1cm}\forall i,
-$$
-<p>&nbsp;<br>
-
-<p>&nbsp;<br>
-$$
-\gamma_i\xi_i = 0,
-$$
-<p>&nbsp;<br>
-
-<p>and </p>
-<p>&nbsp;<br>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -(1-\xi_) \geq 0 \hspace{0.1cm}\forall i.
-$$
-<p>&nbsp;<br>
-</section>
-
-<section>
-<h2 id="kernels-and-non-linearity">Kernels and non-linearity </h2>
-
-<p>The cases we have studied till now, were all characterized by two classes
-with a close to linear separability. The classifiers we have described
-so far find linear boundaries in our input feature space. It is
-possible to make our procedure more flexible by exploring the feature
-space using other basis expansions such as higher-order polynomials,
-wavelets, splines etc.
-</p>
-
-<p>If our feature space is not easy to separate, as shown in the figure
-here, we can achieve a better separation by introducing more complex
-basis functions. The ideal would be, as shown in the next figure, to, via a specific transformation to 
-obtain a separation between the classes which is almost linear. 
-</p>
-
-<p>The change of basis, from \( x\rightarrow z=\phi(x) \) leads to the same type of equations to be solved, except that
-we need to introduce for example a polynomial transformation to a two-dimensional training set.
-</p>
-
+<h2 id="decision-trees-and-regression">Decision trees and Regression  </h2>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
 <div class="cell border-box-sizing code_cell rendered">
@@ -1019,51 +348,478 @@ we need to introduce for example a polynomial transformation to a two-dimensiona
       <div class="input_area">
         <div class="highlight" style="background: #eeeedd">
   <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> PolynomialFeatures
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> LinearRegression
+
+steps=<span style="color: #B452CD">250</span>
+
+distance=<span style="color: #B452CD">0</span>
+x=<span style="color: #B452CD">0</span>
+distance_list=[]
+steps_list=[]
+<span style="color: #8B008B; font-weight: bold">while</span> x&lt;steps:
+    distance+=np.random.randint(-<span style="color: #B452CD">1</span>,<span style="color: #B452CD">2</span>)
+    distance_list.append(distance)
+    x+=<span style="color: #B452CD">1</span>
+    steps_list.append(x)
+plt.plot(steps_list,distance_list, color=<span style="color: #CD5555">&#39;green&#39;</span>, label=<span style="color: #CD5555">&quot;Random Walk Data&quot;</span>)
+
+steps_list=np.asarray(steps_list)
+distance_list=np.asarray(distance_list)
+
+X=steps_list[:,np.newaxis]
+
+<span style="color: #228B22">#Polynomial fits</span>
+
+<span style="color: #228B22">#Degree 2</span>
+poly_features=PolynomialFeatures(degree=<span style="color: #B452CD">2</span>, include_bias=<span style="color: #8B008B; font-weight: bold">False</span>)
+X_poly=poly_features.fit_transform(X)
+
+lin_reg=LinearRegression()
+poly_fit=lin_reg.fit(X_poly,distance_list)
+b=lin_reg.coef_
+c=lin_reg.intercept_
+<span style="color: #658b00">print</span> (<span style="color: #CD5555">&quot;2nd degree coefficients:&quot;</span>)
+<span style="color: #658b00">print</span> (<span style="color: #CD5555">&quot;zero power: &quot;</span>,c)
+<span style="color: #658b00">print</span> (<span style="color: #CD5555">&quot;first power: &quot;</span>, b[<span style="color: #B452CD">0</span>])
+<span style="color: #658b00">print</span> (<span style="color: #CD5555">&quot;second power: &quot;</span>,b[<span style="color: #B452CD">1</span>])
+
+z = np.arange(<span style="color: #B452CD">0</span>, steps, <span style="color: #B452CD">.01</span>)
+z_mod=b[<span style="color: #B452CD">1</span>]*z**<span style="color: #B452CD">2</span>+b[<span style="color: #B452CD">0</span>]*z+c
+
+fit_mod=b[<span style="color: #B452CD">1</span>]*X**<span style="color: #B452CD">2</span>+b[<span style="color: #B452CD">0</span>]*X+c
+plt.plot(z, z_mod, color=<span style="color: #CD5555">&#39;r&#39;</span>, label=<span style="color: #CD5555">&quot;2nd Degree Fit&quot;</span>)
+plt.title(<span style="color: #CD5555">&quot;Polynomial Regression&quot;</span>)
+
+plt.xlabel(<span style="color: #CD5555">&quot;Steps&quot;</span>)
+plt.ylabel(<span style="color: #CD5555">&quot;Distance&quot;</span>)
+
+<span style="color: #228B22">#Degree 10</span>
+poly_features10=PolynomialFeatures(degree=<span style="color: #B452CD">10</span>, include_bias=<span style="color: #8B008B; font-weight: bold">False</span>)
+X_poly10=poly_features10.fit_transform(X)
+
+poly_fit10=lin_reg.fit(X_poly10,distance_list)
+
+y_plot=poly_fit10.predict(X_poly10)
+plt.plot(X, y_plot, color=<span style="color: #CD5555">&#39;black&#39;</span>, label=<span style="color: #CD5555">&quot;10th Degree Fit&quot;</span>)
+
+plt.legend()
+plt.show()
+
+
+<span style="color: #228B22">#Decision Tree Regression</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeRegressor
+regr_1=DecisionTreeRegressor(max_depth=<span style="color: #B452CD">2</span>)
+regr_2=DecisionTreeRegressor(max_depth=<span style="color: #B452CD">5</span>)
+regr_3=DecisionTreeRegressor(max_depth=<span style="color: #B452CD">7</span>)
+regr_1.fit(X, distance_list)
+regr_2.fit(X, distance_list)
+regr_3.fit(X, distance_list)
+
+X_test = np.arange(<span style="color: #B452CD">0.0</span>, steps, <span style="color: #B452CD">0.01</span>)[:, np.newaxis]
+y_1 = regr_1.predict(X_test)
+y_2 = regr_2.predict(X_test)
+y_3=regr_3.predict(X_test)
+
+<span style="color: #228B22"># Plot the results</span>
+plt.figure()
+plt.scatter(X, distance_list, s=<span style="color: #B452CD">2.5</span>, c=<span style="color: #CD5555">&quot;black&quot;</span>, label=<span style="color: #CD5555">&quot;data&quot;</span>)
+plt.plot(X_test, y_1, color=<span style="color: #CD5555">&quot;red&quot;</span>,
+         label=<span style="color: #CD5555">&quot;max_depth=2&quot;</span>, linewidth=<span style="color: #B452CD">2</span>)
+plt.plot(X_test, y_2, color=<span style="color: #CD5555">&quot;green&quot;</span>, label=<span style="color: #CD5555">&quot;max_depth=5&quot;</span>, linewidth=<span style="color: #B452CD">2</span>)
+plt.plot(X_test, y_3, color=<span style="color: #CD5555">&quot;m&quot;</span>, label=<span style="color: #CD5555">&quot;max_depth=7&quot;</span>, linewidth=<span style="color: #B452CD">2</span>)
+
+plt.xlabel(<span style="color: #CD5555">&quot;Data&quot;</span>)
+plt.ylabel(<span style="color: #CD5555">&quot;Darget&quot;</span>)
+plt.title(<span style="color: #CD5555">&quot;Decision Tree Regression&quot;</span>)
+plt.legend()
+plt.show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
+
+<section>
+<h2 id="building-a-tree-regression">Building a tree, regression </h2>
+
+<p>There are mainly two steps</p>
+<ol>
+<p><li> We split the predictor space (the set of possible values \( x_1,x_2,\dots, x_p \)) into \( J \) distinct and non-non-overlapping regions, \( R_1,R_2,\dots,R_J \).</li>
+
+<p><li> For every observation that falls into the region \( R_j \) , we make the same prediction, which is simply the mean of the response values for the training observations in \( R_j \).</li>
+</ol>
+<p>
+<p>How do we construct the regions \( R_1,\dots,R_J \)?  In theory, the
+regions could have any shape. However, we choose to divide the
+predictor space into high-dimensional rectangles, or boxes, for
+simplicity and for ease of interpretation of the resulting predictive
+model. The goal is to find boxes \( R_1,\dots,R_J \) that minimize the
+MSE, given by
+</p>
+
+<p>&nbsp;<br>
+$$
+\sum_{j=1}^J\sum_{i\in R_j}(y_i-\overline{y}_{R_j})^2,
+$$
+<p>&nbsp;<br>
+
+<p>where \( \overline{y}_{R_j} \)  is the mean response for the training observations 
+within box \( j \). 
+</p>
+</section>
+
+<section>
+<h2 id="a-top-down-approach-recursive-binary-splitting">A top-down approach, recursive binary splitting </h2>
+
+<p>Unfortunately, it is computationally infeasible to consider every
+possible partition of the feature space into \( J \) boxes.  The common
+strategy is to take a top-down approach
+</p>
+
+<p>The approach is top-down because it begins at the top of the tree (all
+observations belong to a single region) and then successively splits
+the predictor space; each split is indicated via two new branches
+further down on the tree. It is greedy because at each step of the
+tree-building process, the best split is made at that particular step,
+rather than looking ahead and picking a split that will lead to a
+better tree in some future step.
+</p>
+</section>
+
+<section>
+<h2 id="making-a-tree">Making a tree </h2>
+
+<p>In order to implement the recursive binary splitting we start by selecting
+the predictor \( x_j \) and a cutpoint \( s \) that splits the predictor space into two regions \( R_1 \) and \( R_2 \)
+</p>
+<p>&nbsp;<br>
+$$
+\left\{X\vert x_j < s\right\},
+$$
+<p>&nbsp;<br>
+
+<p>and</p>
+<p>&nbsp;<br>
+$$
+\left\{X\vert x_j \geq s\right\},
+$$
+<p>&nbsp;<br>
+
+<p>so that we obtain the lowest MSE, that is</p>
+<p>&nbsp;<br>
+$$
+\sum_{i:x_i\in R_j}(y_i-\overline{y}_{R_1})^2+\sum_{i:x_i\in R_2}(y_i-\overline{y}_{R_2})^2,
+$$
+<p>&nbsp;<br>
+
+<p>which we want to minimize by considering all predictors
+\( x_1,x_2,\dots,x_p \).  We consider also all possible values of \( s \) for
+each predictor. These values could be determined by randomly assigned
+numbers or by starting at the midpoint and then proceed till we find
+an optimal value.
+</p>
+
+<p>For any \( j \) and \( s \), we define the pair of half-planes where
+\( \overline{y}_{R_1} \) is the mean response for the training
+observations in \( R_1(j,s) \), and \( \overline{y}_{R_2} \) is the mean
+response for the training observations in \( R_2(j,s) \).
+</p>
+
+<p>Finding the values of \( j \) and \( s \) that minimize the above equation can be
+done quite quickly, especially when the number of features \( p \) is not
+too large.
+</p>
+
+<p>Next, we repeat the process, looking
+for the best predictor and best cutpoint in order to split the data
+further so as to minimize the MSE within each of the resulting
+regions. However, this time, instead of splitting the entire predictor
+space, we split one of the two previously identified regions. We now
+have three regions. Again, we look to split one of these three regions
+further, so as to minimize the MSE. The process continues until a
+stopping criterion is reached; for instance, we may continue until no
+region contains more than five observations.
+</p>
+</section>
+
+<section>
+<h2 id="pruning-the-tree">Pruning the tree </h2>
+
+<p>The above procedure is rather straightforward, but leads often to
+overfitting and unnecessarily large and complicated trees. The basic
+idea is to grow a large tree \( T_0 \) and then prune it back in order to
+obtain a subtree. A smaller tree with fewer splits (fewer regions) can
+lead to smaller variance and better interpretation at the cost of a
+little more bias.
+</p>
+
+<p>The so-called Cost complexity pruning algorithm gives us a
+way to do just this. Rather than considering every possible subtree,
+we consider a sequence of trees indexed by a nonnegative tuning
+parameter \( \alpha \).
+</p>
+
+<p>Read more at the following <a href="https://scikit-learn.org/stable/auto_examples/tree/plot_cost_complexity_pruning.html#sphx-glr-auto-examples-tree-plot-cost-complexity-pruning-py" target="_blank">Scikit-Learn link on pruning</a>.</p>
+</section>
+
+<section>
+<h2 id="cost-complexity-pruning">Cost complexity pruning </h2>
+
+<p>For each value of \( \alpha \)  there corresponds a subtree \( T \in T_0 \) such that</p>
+<p>&nbsp;<br>
+$$
+\sum_{m=1}^{\overline{T}}\sum_{i:x_i\in R_m}(y_i-\overline{y}_{R_m})^2+\alpha\overline{T},
+$$
+<p>&nbsp;<br>
+
+<p>is as small as possible. Here \( \overline{T} \) is 
+the number of terminal nodes of the tree \( T \) , \( R_m \) is the
+rectangle (i.e. the subset of predictor space)  corresponding to the \( m \)-th terminal node.
+</p>
+
+<p>The tuning parameter \( \alpha \) controls a trade-off between the subtree&#8217;s
+complexity and its fit to the training data. When \( \alpha = 0 \), then the
+subtree \( T \) will simply equal \( T_0 \), 
+because then the above equation just measures the
+training error. 
+However, as \( \alpha \) increases, there is a price to pay for
+having a tree with many terminal nodes. The above equation will
+tend to be minimized for a smaller subtree. 
+</p>
+
+<p>It turns out that as we increase \( \alpha \) from zero
+branches get pruned from the tree in a nested and predictable fashion,
+so obtaining the whole sequence of subtrees as a function of \( \alpha \) is
+easy. We can select a value of \( \alpha \) using a validation set or using
+cross-validation. We then return to the full data set and obtain the
+subtree corresponding to \( \alpha \). 
+</p>
+</section>
+
+<section>
+<h2 id="schematic-regression-procedure">Schematic Regression Procedure </h2>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Building a Regression Tree</b>
+<p>
+
+<ol>
+<p><li> Use recursive binary splitting to grow a large tree on the training data, stopping only when each terminal node has fewer than some minimum number of observations.</li>
+<p><li> Apply cost complexity pruning to the large tree in order to obtain a sequence of best subtrees, as a function of \( \alpha \).</li>
+<p><li> Use for example \( K \)-fold cross-validation to choose \( \alpha \). Divide the training observations into \( K \) folds. For each \( k=1,2,\dots,K \) we:</li> 
+<ul>
+
+<p><li> repeat steps 1 and 2 on all but the \( k \)-th fold of the training data.</li>
+
+<p><li> Then we valuate the mean squared prediction error on the data in the left-out \( k \)-th fold, as a function of \( \alpha \).</li>
+
+<p><li> Finally  we average the results for each value of \( \alpha \), and pick \( \alpha \) to minimize the average error.</li>
+</ul>
+<p>
+<p><li> Return the subtree from Step 2 that corresponds to the chosen value of \( \alpha \).</li> 
+</ol>
+</div>
+</section>
+
+<section>
+<h2 id="a-classification-tree">A Classification Tree </h2>
+
+<p>A classification tree is very similar to a regression tree, except
+that it is used to predict a qualitative response rather than a
+quantitative one. Recall that for a regression tree, the predicted
+response for an observation is given by the mean response of the
+training observations that belong to the same terminal node. In
+contrast, for a classification tree, we predict that each observation
+belongs to the most commonly occurring class of training observations
+in the region to which it belongs. In interpreting the results of a
+classification tree, we are often interested not only in the class
+prediction corresponding to a particular terminal node region, but
+also in the class proportions among the training observations that
+fall into that region.  
+</p>
+</section>
+
+<section>
+<h2 id="growing-a-classification-tree">Growing a classification tree </h2>
+
+<p>The task of growing a
+classification tree is quite similar to the task of growing a
+regression tree. Just as in the regression setting, we use recursive
+binary splitting to grow a classification tree. However, in the
+classification setting, the MSE cannot be used as a criterion for making
+the binary splits.  A natural alternative to MSE is the <b>classification
+error rate</b>. Since we plan to assign an observation in a given region
+to the most commonly occurring error rate class of training
+observations in that region, the classification error rate is simply
+the fraction of the training observations in that region that do not
+belong to the most common class. 
+</p>
+
+<p>When building a classification tree, either the Gini index or the
+entropy are typically used to evaluate the quality of a particular
+split, since these two approaches are more sensitive to node purity
+than is the classification error rate. 
+</p>
+</section>
+
+<section>
+<h2 id="classification-tree-how-to-split-nodes">Classification tree, how to split nodes </h2>
+
+<p>If our targets are the outcome of a classification process that takes
+for example \( k=1,2,\dots,K \) values, the only thing we need to think of
+is to set up the splitting criteria for each node.
+</p>
+
+<p>We define a PDF \( p_{mk} \) that represents the number of observations of
+a class \( k \) in a region \( R_m \) with \( N_m \) observations. We represent
+this likelihood function in terms of the proportion \( I(y_i=k) \) of
+observations of this class in the region \( R_m \) as
+</p>
+
+<p>&nbsp;<br>
+$$
+p_{mk} = \frac{1}{N_m}\sum_{x_i\in R_m}I(y_i=k).
+$$
+<p>&nbsp;<br>
+
+<p>We let \( p_{mk} \) represent the majority class of observations in region
+\( m \). The three most common ways of splitting a node are given by
+</p>
+
+<ul>
+<p><li> Misclassification error</li> 
+</ul>
+<p>
+<p>&nbsp;<br>
+$$
+p_{mk} = \frac{1}{N_m}\sum_{x_i\in R_m}I(y_i\ne k) = 1-p_{mk}.
+$$
+<p>&nbsp;<br>
+
+<ul>
+<p><li> Gini index \( g \)</li>
+</ul>
+<p>
+<p>&nbsp;<br>
+$$
+g = \sum_{k=1}^K p_{mk}(1-p_{mk}).
+$$
+<p>&nbsp;<br>
+
+<ul>
+<p><li> Information entropy or just entropy \( s \)</li>
+</ul>
+<p>
+<p>&nbsp;<br>
+$$
+s = -\sum_{k=1}^K p_{mk}\log{p_{mk}}.
+$$
+<p>&nbsp;<br>
+</section>
+
+<section>
+<h2 id="visualizing-the-tree-classification">Visualizing the Tree, Classification </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> load_breast_cancer
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> confusion_matrix
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_graphviz
+
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> Image 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+
+
+cancer = load_breast_cancer()
+X = pd.DataFrame(cancer.data, columns=cancer.feature_names)
+<span style="color: #658b00">print</span>(X)
+y = pd.Categorical.from_codes(cancer.target, cancer.target_names)
+y = pd.get_dummies(y)
+<span style="color: #658b00">print</span>(y)
+X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=<span style="color: #B452CD">1</span>)
+tree_clf = DecisionTreeClassifier(max_depth=<span style="color: #B452CD">5</span>)
+tree_clf.fit(X_train, y_train)
+
+export_graphviz(
+    tree_clf,
+    out_file=<span style="color: #CD5555">&quot;DataFiles/cancer.dot&quot;</span>,
+    feature_names=cancer.feature_names,
+    class_names=cancer.target_names,
+    rounded=<span style="color: #8B008B; font-weight: bold">True</span>,
+    filled=<span style="color: #8B008B; font-weight: bold">True</span>
+)
+cmd = <span style="color: #CD5555">&#39;dot -Tpng DataFiles/cancer.dot -o DataFiles/cancer.png&#39;</span>
+os.system(cmd)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
+
+<section>
+<h2 id="visualizing-the-tree-the-moons">Visualizing the Tree, The Moons  </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #228B22"># Common imports</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span>  train_test_split 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> make_moons
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_graphviz
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
 
 np.random.seed(<span style="color: #B452CD">42</span>)
+X, y = make_moons(n_samples=<span style="color: #B452CD">100</span>, noise=<span style="color: #B452CD">0.25</span>, random_state=<span style="color: #B452CD">53</span>)
+X_train, X_test, y_train, y_test = train_test_split(X,y,random_state=<span style="color: #B452CD">0</span>)
+tree_clf = DecisionTreeClassifier(max_depth=<span style="color: #B452CD">5</span>)
+tree_clf.fit(X_train, y_train)
 
-<span style="color: #228B22"># To plot pretty figures</span>
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib</span>
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
-plt.rcParams[<span style="color: #CD5555">&#39;axes.labelsize&#39;</span>] = <span style="color: #B452CD">14</span>
-plt.rcParams[<span style="color: #CD5555">&#39;xtick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
-plt.rcParams[<span style="color: #CD5555">&#39;ytick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
-
-
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets
-
-
-
-X1D = np.linspace(-<span style="color: #B452CD">4</span>, <span style="color: #B452CD">4</span>, <span style="color: #B452CD">9</span>).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
-X2D = np.c_[X1D, X1D**<span style="color: #B452CD">2</span>]
-y = np.array([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>])
-
-plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
-
-plt.subplot(<span style="color: #B452CD">121</span>)
-plt.grid(<span style="color: #8B008B; font-weight: bold">True</span>, which=<span style="color: #CD5555">&#39;both&#39;</span>)
-plt.axhline(y=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.plot(X1D[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">0</span>], np.zeros(<span style="color: #B452CD">4</span>), <span style="color: #CD5555">&quot;bs&quot;</span>)
-plt.plot(X1D[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">1</span>], np.zeros(<span style="color: #B452CD">5</span>), <span style="color: #CD5555">&quot;g^&quot;</span>)
-plt.gca().get_yaxis().set_ticks([])
-plt.xlabel(<span style="color: #CD5555">r&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.axis([-<span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">4.5</span>, -<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">0.2</span>])
-
-plt.subplot(<span style="color: #B452CD">122</span>)
-plt.grid(<span style="color: #8B008B; font-weight: bold">True</span>, which=<span style="color: #CD5555">&#39;both&#39;</span>)
-plt.axhline(y=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.axvline(x=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.plot(X2D[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">0</span>], X2D[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">0</span>], <span style="color: #CD5555">&quot;bs&quot;</span>)
-plt.plot(X2D[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">1</span>], X2D[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;g^&quot;</span>)
-plt.xlabel(<span style="color: #CD5555">r&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.ylabel(<span style="color: #CD5555">r&quot;$x_2$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>, rotation=<span style="color: #B452CD">0</span>)
-plt.gca().get_yaxis().set_ticks([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">4</span>, <span style="color: #B452CD">8</span>, <span style="color: #B452CD">12</span>, <span style="color: #B452CD">16</span>])
-plt.plot([-<span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">4.5</span>], [<span style="color: #B452CD">6.5</span>, <span style="color: #B452CD">6.5</span>], <span style="color: #CD5555">&quot;r--&quot;</span>, linewidth=<span style="color: #B452CD">3</span>)
-plt.axis([-<span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">4.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">17</span>])
-plt.subplots_adjust(right=<span style="color: #B452CD">1</span>)
-plt.show()
+export_graphviz(
+    tree_clf,
+    out_file=<span style="color: #CD5555">&quot;DataFiles/moons.dot&quot;</span>,
+    rounded=<span style="color: #8B008B; font-weight: bold">True</span>,
+    filled=<span style="color: #8B008B; font-weight: bold">True</span>
+)
+cmd = <span style="color: #CD5555">&#39;dot -Tpng DataFiles/moons.dot -o DataFiles/moons.png&#39;</span>
+os.system(cmd)
 </pre>
 </div>
       </div>
@@ -1081,139 +837,248 @@ plt.show()
 </section>
 
 <section>
-<h2 id="the-equations">The equations </h2>
+<h2 id="other-ways-of-visualizing-the-trees">Other ways of visualizing the trees </h2>
 
-<p>Suppose we define a polynomial transformation of degree two only (we continue to live in a plane with \( x_i \) and \( y_i \) as variables)</p>
-<p>&nbsp;<br>
-$$
-z = \phi(x_i) =\left(x_i^2, y_i^2, \sqrt{2}x_iy_i\right).
-$$
-<p>&nbsp;<br>
+<p><b>Scikit-Learn</b> has also another way to visualize the trees which is very useful, here with the Iris data.</p>
 
-<p>With our new basis, the equations we solved earlier are basically the same, that is we have now (without the slack option for simplicity)</p>
-<p>&nbsp;<br>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{z}_i^T\boldsymbol{z}_j,
-$$
-<p>&nbsp;<br>
 
-<p>subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \), and for the support vectors</p>
-<p>&nbsp;<br>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{z}_i+b)= 1 \hspace{0.1cm}\forall i,
-$$
-<p>&nbsp;<br>
-
-<p>from which we also find \( b \).
-To compute \( \boldsymbol{z}_i^T\boldsymbol{z}_j \) we define the kernel \( K(\boldsymbol{x}_i,\boldsymbol{x}_j) \) as
-</p>
-<p>&nbsp;<br>
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=\boldsymbol{z}_i^T\boldsymbol{z}_j= \phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j).
-$$
-<p>&nbsp;<br>
-
-<p>For the above example, the kernel reads</p>
-<p>&nbsp;<br>
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=[x_i^2, y_i^2, \sqrt{2}x_iy_i]^T\begin{bmatrix} x_j^2 \\ y_j^2 \\ \sqrt{2}x_jy_j \end{bmatrix}=x_i^2x_j^2+2x_ix_jy_iy_j+y_i^2y_j^2.
-$$
-<p>&nbsp;<br>
-
-<p>We note that this is nothing but the dot product of the two original
-vectors \( (\boldsymbol{x}_i^T\boldsymbol{x}_j)^2 \). Instead of thus computing the
-product in the Lagrangian of \( \boldsymbol{z}_i^T\boldsymbol{z}_j \) we simply compute
-the dot product \( (\boldsymbol{x}_i^T\boldsymbol{x}_j)^2 \).
-</p>
-
-<p>This leads to the so-called
-kernel trick and the result leads to the same as if we went through
-the trouble of performing the transformation
-\( \phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j) \) during the SVM calculations.
-</p>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> load_iris
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> tree
+X, y = load_iris(return_X_y=<span style="color: #8B008B; font-weight: bold">True</span>)
+tree_clf = tree.DecisionTreeClassifier()
+tree_clf = tree_clf.fit(X, y)
+<span style="color: #228B22"># and then plot the tree</span>
+tree.plot_tree(tree_clf) 
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
 </section>
 
 <section>
-<h2 id="the-problem-to-solve">The problem to solve </h2>
-<p>Using our definition of the kernel We can rewrite again the Lagrangian</p>
-<p>&nbsp;<br>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{z}_j,
-$$
-<p>&nbsp;<br>
+<h2 id="printing-out-as-text">Printing out as text </h2>
 
-<p>subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \) in terms of a convex optimization problem</p>
-<p>&nbsp;<br>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1K(\boldsymbol{x}_1,\boldsymbol{x}_1) & y_1y_2K(\boldsymbol{x}_1,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_1,\boldsymbol{x}_n) \\
-y_2y_1K(\boldsymbol{x}_2,\boldsymbol{x}_1) & y_2y_2(\boldsymbol{x}_2,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_2,\boldsymbol{x}_n) \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1K(\boldsymbol{x}_n,\boldsymbol{x}_1) & y_ny_2K(\boldsymbol{x}_n\boldsymbol{x}_2) & \dots & \dots & y_ny_nK(\boldsymbol{x}_n,\boldsymbol{x}_n) \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{1}\boldsymbol{\lambda}, 
-$$
-<p>&nbsp;<br>
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
-If we add the slack constants this leads to the additional constraint \( 0\leq \lambda_i \leq C \).
+<p>Alternatively, the tree can also be exported in textual format with the function exporttext.
+This method doesn&#8217;t require the installation of external libraries and is more compact:
 </p>
 
-<p>We can rewrite this (see the solutions below) in terms of a convex optimization problem of the type</p>
-<p>&nbsp;<br>
-$$
-\begin{align*}
-    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\boldsymbol{\lambda}^T\boldsymbol{P}\boldsymbol{\lambda}+\boldsymbol{q}^T\boldsymbol{\lambda},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{\lambda} \preceq \boldsymbol{h} \hspace{0.2cm} \wedge \boldsymbol{A}\boldsymbol{\lambda}=f.
-\end{align*}
-$$
-<p>&nbsp;<br>
 
-<p>Below we discuss how to solve these equations. Here we note that the matrix \( \boldsymbol{P} \) has matrix elements \( p_{ij}=y_iy_jK(\boldsymbol{x}_i,\boldsymbol{x}_j) \).
-Given a kernel \( K \) and the targets \( y_i \) this matrix is easy to set up. The constraint \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \) leads to \( f=0 \) and \( \boldsymbol{A}=\boldsymbol{y} \). How to set up the matrix \( \boldsymbol{G} \) is discussed later. Here note that the inequalities \( 0\leq \lambda_i \leq C \) can be split up into
-\( 0\leq \lambda_i \) and \( \lambda_i \leq C \). These two inequalities define then the matrix \( \boldsymbol{G} \) and the vector \( \boldsymbol{h} \).
-</p>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> load_iris
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_text
+iris = load_iris()
+decision_tree = DecisionTreeClassifier(random_state=<span style="color: #B452CD">0</span>, max_depth=<span style="color: #B452CD">2</span>)
+decision_tree = decision_tree.fit(iris.data, iris.target)
+r = export_text(decision_tree, feature_names=iris[<span style="color: #CD5555">&#39;feature_names&#39;</span>])
+<span style="color: #658b00">print</span>(r)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
 </section>
 
 <section>
-<h2 id="different-kernels-and-mercer-s-theorem">Different kernels and Mercer's theorem </h2>
+<h2 id="algorithms-for-setting-up-decision-trees">Algorithms for Setting up Decision Trees </h2>
 
-<p>There are several popular kernels being used. These are</p>
+<p>Two algorithms stand out in the set up of decision trees:</p>
 <ol>
-<p><li> Linear: \( K(\boldsymbol{x},\boldsymbol{y})=\boldsymbol{x}^T\boldsymbol{y} \),</li>
-<p><li> Polynomial: \( K(\boldsymbol{x},\boldsymbol{y})=(\boldsymbol{x}^T\boldsymbol{y}+\gamma)^d \),</li>
-<p><li> Gaussian Radial Basis Function: \( K(\boldsymbol{x},\boldsymbol{y})=\exp{\left(-\gamma\vert\vert\boldsymbol{x}-\boldsymbol{y}\vert\vert^2\right)} \),</li>
-<p><li> Tanh: \( K(\boldsymbol{x},\boldsymbol{y})=\tanh{(\boldsymbol{x}^T\boldsymbol{y}+\gamma)} \),</li>
+<p><li> The CART (Classification And Regression Tree) algorithm for both classification and regression</li>
+<p><li> The ID3 algorithm based on the computation of the information gain for classification</li>
 </ol>
 <p>
-<p>and many other ones.</p>
-
-<p>An important theorem for us is <a href="https://en.wikipedia.org/wiki/Mercer%27s_theorem" target="_blank">Mercer's
-theorem</a>.  The
-theorem states that if a kernel function \( K \) is symmetric, continuous
-and leads to a positive semi-definite matrix \( \boldsymbol{P} \) then there
-exists a function \( \phi \) that maps \( \boldsymbol{x}_i \) and \( \boldsymbol{x}_j \) into
-another space (possibly with much higher dimensions) such that
-</p>
-
-<p>&nbsp;<br>
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=\phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j).
-$$
-<p>&nbsp;<br>
-
-<p>So you can use \( K \) as a kernel since you know \( \phi \) exists, even if
-you don&#8217;t know what \( \phi \) is. 
-</p>
-
-<p>Note that some frequently used kernels (such as the Sigmoid kernel)
-don&#8217;t respect all of Mercer&#8217;s conditions, yet they generally work well
-in practice.
+<p>We discuss both algorithms with applications here. The popular library
+<b>Scikit-Learn</b> uses the CART algorithm. For classification problems
+you can use either the <b>gini</b> index or the <b>entropy</b> to split a tree
+in two branches.
 </p>
 </section>
 
 <section>
-<h2 id="the-moons-example">The moons example </h2>
+<h2 id="the-cart-algorithm-for-classification">The CART algorithm for Classification </h2>
+
+<p>For classification, the CART algorithm splits the data set in two subsets using a single feature \( k \) and a threshold \( t_k \).
+This could be for example a threshold set by a number below a certain circumference of a malign tumor.
+</p>
+
+<p>How do we find these two quantities?
+We search for the pair \( (k,t_k) \) that produces the purest subset using for example the <b>gini</b> factor \( G \).
+The cost function it tries to minimize is then
+</p>
+<p>&nbsp;<br>
+$$
+C(k,t_k) = \frac{m_{\mathrm{left}}}{m}G_{\mathrm{left}}+ \frac{m_{\mathrm{right}}}{m}G_{\mathrm{right}},
+$$
+<p>&nbsp;<br>
+
+<p>where \( G_{\mathrm{left/right}} \) measures the impurity of the left/right subset  and \( m_{\mathrm{left/right}} \)
+ is the number of instances in the left/right subset
+</p>
+
+<p>Once it has successfully split the training set in two, it splits the subsets using the same logic, then the subsubsets
+and so on, recursively. It stops recursing once it reaches the maximum depth (defined by the
+\( max\_depth \) hyperparameter), or if it cannot find a split that will reduce impurity. A few other
+hyperparameters control additional stopping conditions such as the \( min\_samples\_split \),
+\( min\_samples\_leaf \), \( min\_weight\_fraction\_leaf \), and \( max\_leaf\_nodes \).
+</p>
+</section>
+
+<section>
+<h2 id="the-cart-algorithm-for-regression">The CART algorithm for Regression </h2>
+
+<p>The CART algorithm for regression works is similar to the one for classification except that instead of trying to split the
+training set in a way that minimizes say the <b>gini</b> or <b>entropy</b> impurity, it now tries to split the training set in a way that minimizes our well-known mean-squared error (MSE). The cost function is now
+</p>
+<p>&nbsp;<br>
+$$
+C(k,t_k) = \frac{m_{\mathrm{left}}}{m}\mathrm{MSE}_{\mathrm{left}}+ \frac{m_{\mathrm{right}}}{m}\mathrm{MSE}_{\mathrm{right}}.
+$$
+<p>&nbsp;<br>
+
+<p>Here the MSE for a specific node is defined as</p>
+<p>&nbsp;<br>
+$$
+\mathrm{MSE}_{\mathrm{node}}=\frac{1}{m_\mathrm{node}}\sum_{i\in \mathrm{node}}(\overline{y}_{\mathrm{node}}-y_i)^2,
+$$
+<p>&nbsp;<br>
+
+<p>with</p>
+<p>&nbsp;<br>
+$$
+\overline{y}_{\mathrm{node}}=\frac{1}{m_\mathrm{node}}\sum_{i\in \mathrm{node}}y_i,
+$$
+<p>&nbsp;<br>
+
+<p>the mean value of all observations in a specific node.</p>
+
+<p>Without any regularization, the regression task for decision trees, 
+just like for classification tasks, is  prone to overfitting.
+</p>
+</section>
+
+<section>
+<h2 id="why-binary-splits">Why binary splits? </h2>
+
+<p>It is custom to split to a tree uising binary splits. The reason is
+that multiway splits fragment the data too quickly, leaving
+insufficient data at the next level down.  Multiway splits can be
+achieved by a series of binary split and this is normally preferred.
+</p>
+</section>
+
+<section>
+<h2 id="computing-a-tree-using-the-gini-index">Computing a Tree using the Gini Index </h2>
+
+<p>Consider the following example with attributes/features and two
+possible outcomes (classes)  for each attribute.  Assume we wish to find some
+correlations between the average grade of a student as function of the
+number of hours studied and hours slept. We want also to correlate the
+grade in a given course with the general trend, whether the students
+recently has gotten grades below average or above.
+</p>
+
+<p>We have three features/attributes</p>
+<ol>
+<p><li> Trend of average grades before present course, classified as either below  or above  the average grade of the whole class</li>
+
+<p><li> The number of hours studies, classified again as either higher (more than 3 hours per day) or lower . Here we have used a standard for one \( ECTS \) which is scaled to 25-30 hours of work for a semester which lasts 18 weeks, with 15 weeks of lectures and 3 weeks for exams, assuming a total of 30 ECTS per semester.</li> 
+<p><li> The number of hours slept as high for more than \( 8 \) hours and below  for less than 8 hours of sleep, classified again as either high or low</li>
+<p><li> The final grade whether it is above or below average</li>
+</ol>
+</section>
+
+<section>
+<h2 id="the-table">The Table </h2>
+
+<table class="dotable" border="1">
+<thead>
+<tr><th align="center">Grade Trend</th> <th align="center">Hours slept</th> <th align="center">Hours Studied</th> <th align="center">Grade</th> </tr>
+</thead>
+<tbody>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   High           </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+</tbody>
+</table>
+</section>
+
+<section>
+<h2 id="computing-the-various-gini-indices">Computing the various Gini Indices </h2>
+
+<p>In computations we will translate all classes into numbers. Being
+these binary classes, they can easily be split into ones and zeros.
+</p>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Gini index for Average trend</b>
+<p>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+</div>
+</section>
+
+<section>
+<h2 id="computing-the-various-gini-indices-hours-slept">Computing the various Gini Indices, Hours slept </h2>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Gini index for hour slept</b>
+<p>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+</div>
+</section>
+
+<section>
+<h2 id="computing-the-various-gini-indices-hours-studied">Computing the various Gini Indices, Hours studied </h2>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Gini index for hour studied</b>
+<p>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+</div>
+
+<p>For final tree, see the above handwritten notes</p>
+</section>
+
+<section>
+<h2 id="a-possible-code-using-scikit-learn">A possible code using Scikit-Learn </h2>
+
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
 <div class="cell border-box-sizing code_cell rendered">
@@ -1221,193 +1086,1006 @@ in practice.
     <div class="inner_cell">
       <div class="input_area">
         <div class="highlight" style="background: #eeeedd">
-  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">__future__</span> <span style="color: #8B008B; font-weight: bold">import</span> division, print_function, unicode_literals
-
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #228B22"># Common imports</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
-np.random.seed(<span style="color: #B452CD">42</span>)
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_graphviz
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler, OneHotEncoder
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.compose</span> <span style="color: #8B008B; font-weight: bold">import</span> ColumnTransformer
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> Image 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
 
+<span style="color: #228B22"># Where to save the figures and data files</span>
+PROJECT_ROOT_DIR = <span style="color: #CD5555">&quot;Results&quot;</span>
+FIGURE_ID = <span style="color: #CD5555">&quot;Results/FigureFiles&quot;</span>
+DATA_ID = <span style="color: #CD5555">&quot;DataFiles/&quot;</span>
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(PROJECT_ROOT_DIR):
+    os.mkdir(PROJECT_ROOT_DIR)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(FIGURE_ID):
+    os.makedirs(FIGURE_ID)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(DATA_ID):
+    os.makedirs(DATA_ID)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">image_path</span>(fig_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(FIGURE_ID, fig_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">data_path</span>(dat_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(DATA_ID, dat_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">save_fig</span>(fig_id):
+    plt.savefig(image_path(fig_id) + <span style="color: #CD5555">&quot;.png&quot;</span>, <span style="color: #658b00">format</span>=<span style="color: #CD5555">&#39;png&#39;</span>)
+
+infile = <span style="color: #658b00">open</span>(data_path(<span style="color: #CD5555">&quot;grades.csv&quot;</span>),<span style="color: #CD5555">&#39;r&#39;</span>)
+
+<span style="color: #228B22"># Read the experimental data with Pandas</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> display
+grades = pd.read_csv(infile)
+grades = pd.DataFrame(grades)
+display(grades)
+<span style="color: #228B22"># Features and targets</span>
+X = grades.loc[:, grades.columns != <span style="color: #CD5555">&#39;Grade&#39;</span>].values
+y = grades.loc[:, grades.columns == <span style="color: #CD5555">&#39;Grade&#39;</span>].values
+<span style="color: #658b00">print</span>(X)
+<span style="color: #228B22"># Then do a Classification tree</span>
+tree_clf = DecisionTreeClassifier(max_depth=<span style="color: #B452CD">2</span>)
+tree_clf.fit(X, y)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Train set accuracy with Decision Tree: {:.2f}&quot;</span>.format(tree_clf.score(X,y)))
+<span style="color: #228B22">#transfer to a decision tree graph</span>
+export_graphviz(
+    tree_clf,
+    out_file=<span style="color: #CD5555">&quot;DataFiles/grade.dot&quot;</span>,
+    rounded=<span style="color: #8B008B; font-weight: bold">True</span>,
+    filled=<span style="color: #8B008B; font-weight: bold">True</span>
+)
+cmd = <span style="color: #CD5555">&#39;dot -Tpng DataFiles/grade.dot -o DataFiles/grades.png&#39;</span>
+os.system(cmd)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
+
+<section>
+<h2 id="further-example-computing-the-gini-index">Further example: Computing the Gini index </h2>
+
+<p>The next example we will look at is a classical one in many Machine
+Learning applications. Based on various meteorological features, we
+have several so-called attributes which decide whether we at the end
+will do some outdoor activity like skiing, going for a bike ride etc
+etc.  The table here contains the feautures <b>outlook</b>, <b>temperature</b>,
+<b>humidity</b> and <b>wind</b>.  The target or output is whether we ride
+(True=1) or whether we do something else that day (False=0). The
+attributes for each feature are then sunny, overcast and rain for the
+outlook, hot, cold and mild for temperature, high and normal for
+humidity and weak and strong for wind.
+</p>
+
+<p>The table here summarizes the various attributes and</p>
+<table class="dotable" border="1">
+<thead>
+<tr><th align="center">Day</th> <th align="center">Outlook </th> <th align="center">Temperature</th> <th align="center">Humidity</th> <th align="center"> Wind </th> <th align="center">Ride</th> </tr>
+</thead>
+<tbody>
+<tr><td align="center">   1      </td> <td align="center">   Sunny       </td> <td align="center">   Hot            </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   0       </td> </tr>
+<tr><td align="center">   2      </td> <td align="center">   Sunny       </td> <td align="center">   Hot            </td> <td align="center">   High        </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   3      </td> <td align="center">   Overcast    </td> <td align="center">   Hot            </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   4      </td> <td align="center">   Rain        </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   5      </td> <td align="center">   Rain        </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   6      </td> <td align="center">   Rain        </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Strong    </td> <td align="center">   0       </td> </tr>
+<tr><td align="center">   7      </td> <td align="center">   Overcast    </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   8      </td> <td align="center">   Sunny       </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   0       </td> </tr>
+<tr><td align="center">   9      </td> <td align="center">   Sunny       </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   10     </td> <td align="center">   Rain        </td> <td align="center">   Mild           </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   11     </td> <td align="center">   Sunny       </td> <td align="center">   Mild           </td> <td align="center">   Normal      </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   12     </td> <td align="center">   Overcast    </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   13     </td> <td align="center">   Overcast    </td> <td align="center">   Hot            </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   14     </td> <td align="center">   Rain        </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Strong    </td> <td align="center">   0       </td> </tr>
+</tbody>
+</table>
+</section>
+
+<section>
+<h2 id="simple-python-code-to-read-in-data-and-perform-classification">Simple Python Code to read in Data and perform Classification </h2>
+
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #228B22"># Common imports</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_graphviz
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler, OneHotEncoder
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.compose</span> <span style="color: #8B008B; font-weight: bold">import</span> ColumnTransformer
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> Image 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
+
+<span style="color: #228B22"># Where to save the figures and data files</span>
+PROJECT_ROOT_DIR = <span style="color: #CD5555">&quot;Results&quot;</span>
+FIGURE_ID = <span style="color: #CD5555">&quot;Results/FigureFiles&quot;</span>
+DATA_ID = <span style="color: #CD5555">&quot;DataFiles/&quot;</span>
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(PROJECT_ROOT_DIR):
+    os.mkdir(PROJECT_ROOT_DIR)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(FIGURE_ID):
+    os.makedirs(FIGURE_ID)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(DATA_ID):
+    os.makedirs(DATA_ID)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">image_path</span>(fig_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(FIGURE_ID, fig_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">data_path</span>(dat_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(DATA_ID, dat_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">save_fig</span>(fig_id):
+    plt.savefig(image_path(fig_id) + <span style="color: #CD5555">&quot;.png&quot;</span>, <span style="color: #658b00">format</span>=<span style="color: #CD5555">&#39;png&#39;</span>)
+
+infile = <span style="color: #658b00">open</span>(data_path(<span style="color: #CD5555">&quot;rideclass.csv&quot;</span>),<span style="color: #CD5555">&#39;r&#39;</span>)
+
+<span style="color: #228B22"># Read the experimental data with Pandas</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> display
+ridedata = pd.read_csv(infile,names = (<span style="color: #CD5555">&#39;Outlook&#39;</span>,<span style="color: #CD5555">&#39;Temperature&#39;</span>,<span style="color: #CD5555">&#39;Humidity&#39;</span>,<span style="color: #CD5555">&#39;Wind&#39;</span>,<span style="color: #CD5555">&#39;Ride&#39;</span>))
+ridedata = pd.DataFrame(ridedata)
+
+<span style="color: #228B22"># Features and targets</span>
+X = ridedata.loc[:, ridedata.columns != <span style="color: #CD5555">&#39;Ride&#39;</span>].values
+y = ridedata.loc[:, ridedata.columns == <span style="color: #CD5555">&#39;Ride&#39;</span>].values
+
+<span style="color: #228B22"># Create the encoder.</span>
+encoder = OneHotEncoder(handle_unknown=<span style="color: #CD5555">&quot;ignore&quot;</span>)
+<span style="color: #228B22"># Assume for simplicity all features are categorical.</span>
+encoder.fit(X)    
+<span style="color: #228B22"># Apply the encoder.</span>
+X = encoder.transform(X)
+<span style="color: #658b00">print</span>(X)
+<span style="color: #228B22"># Then do a Classification tree</span>
+tree_clf = DecisionTreeClassifier(max_depth=<span style="color: #B452CD">2</span>)
+tree_clf.fit(X, y)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Train set accuracy with Decision Tree: {:.2f}&quot;</span>.format(tree_clf.score(X,y)))
+<span style="color: #228B22">#transfer to a decision tree graph</span>
+export_graphviz(
+    tree_clf,
+    out_file=<span style="color: #CD5555">&quot;DataFiles/ride.dot&quot;</span>,
+    rounded=<span style="color: #8B008B; font-weight: bold">True</span>,
+    filled=<span style="color: #8B008B; font-weight: bold">True</span>
+)
+cmd = <span style="color: #CD5555">&#39;dot -Tpng DataFiles/cancer.dot -o DataFiles/cancer.png&#39;</span>
+os.system(cmd)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
+
+<section>
+<h2 id="computing-the-gini-factor">Computing the Gini Factor  </h2>
+
+<p>The above functions (gini, entropy and misclassification error) are
+important components of the so-called CART algorithm. We will discuss
+this algorithm below after we have discussed the information gain
+algorithm ID3.
+</p>
+
+<p>In the example here we have converted all our attributes into numerical values \( 0,1,2 \) etc.</p>
+
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #228B22"># Split a dataset based on an attribute and an attribute value</span>
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">test_split</span>(index, value, dataset):
+	left, right = <span style="color: #658b00">list</span>(), <span style="color: #658b00">list</span>()
+	<span style="color: #8B008B; font-weight: bold">for</span> row <span style="color: #8B008B">in</span> dataset:
+		<span style="color: #8B008B; font-weight: bold">if</span> row[index] &lt; value:
+			left.append(row)
+		<span style="color: #8B008B; font-weight: bold">else</span>:
+			right.append(row)
+	<span style="color: #8B008B; font-weight: bold">return</span> left, right
+ 
+<span style="color: #228B22"># Calculate the Gini index for a split dataset</span>
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">gini_index</span>(groups, classes):
+	<span style="color: #228B22"># count all samples at split point</span>
+	n_instances = <span style="color: #658b00">float</span>(<span style="color: #658b00">sum</span>([<span style="color: #658b00">len</span>(group) <span style="color: #8B008B; font-weight: bold">for</span> group <span style="color: #8B008B">in</span> groups]))
+	<span style="color: #228B22"># sum weighted Gini index for each group</span>
+	gini = <span style="color: #B452CD">0.0</span>
+	<span style="color: #8B008B; font-weight: bold">for</span> group <span style="color: #8B008B">in</span> groups:
+		size = <span style="color: #658b00">float</span>(<span style="color: #658b00">len</span>(group))
+		<span style="color: #228B22"># avoid divide by zero</span>
+		<span style="color: #8B008B; font-weight: bold">if</span> size == <span style="color: #B452CD">0</span>:
+			<span style="color: #8B008B; font-weight: bold">continue</span>
+		score = <span style="color: #B452CD">0.0</span>
+		<span style="color: #228B22"># score the group based on the score for each class</span>
+		<span style="color: #8B008B; font-weight: bold">for</span> class_val <span style="color: #8B008B">in</span> classes:
+			p = [row[-<span style="color: #B452CD">1</span>] <span style="color: #8B008B; font-weight: bold">for</span> row <span style="color: #8B008B">in</span> group].count(class_val) / size
+			score += p * p
+		<span style="color: #228B22"># weight the group score by its relative size</span>
+		gini += (<span style="color: #B452CD">1.0</span> - score) * (size / n_instances)
+	<span style="color: #8B008B; font-weight: bold">return</span> gini
+
+<span style="color: #228B22"># Select the best split point for a dataset</span>
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">get_split</span>(dataset):
+	class_values = <span style="color: #658b00">list</span>(<span style="color: #658b00">set</span>(row[-<span style="color: #B452CD">1</span>] <span style="color: #8B008B; font-weight: bold">for</span> row <span style="color: #8B008B">in</span> dataset))
+	b_index, b_value, b_score, b_groups = <span style="color: #B452CD">999</span>, <span style="color: #B452CD">999</span>, <span style="color: #B452CD">999</span>, <span style="color: #8B008B; font-weight: bold">None</span>
+	<span style="color: #8B008B; font-weight: bold">for</span> index <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(<span style="color: #658b00">len</span>(dataset[<span style="color: #B452CD">0</span>])-<span style="color: #B452CD">1</span>):
+		<span style="color: #8B008B; font-weight: bold">for</span> row <span style="color: #8B008B">in</span> dataset:
+			groups = test_split(index, row[index], dataset)
+			gini = gini_index(groups, class_values)
+			<span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;X%d &lt; %.3f Gini=%.3f&#39;</span> % ((index+<span style="color: #B452CD">1</span>), row[index], gini))
+			<span style="color: #8B008B; font-weight: bold">if</span> gini &lt; b_score:
+				b_index, b_value, b_score, b_groups = index, row[index], gini, groups
+	<span style="color: #8B008B; font-weight: bold">return</span> {<span style="color: #CD5555">&#39;index&#39;</span>:b_index, <span style="color: #CD5555">&#39;value&#39;</span>:b_value, <span style="color: #CD5555">&#39;groups&#39;</span>:b_groups}
+ 
+dataset = [[<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>],
+            [<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">2</span>,<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">2</span>,<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>],
+            [<span style="color: #B452CD">1</span>,<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>],
+            [<span style="color: #B452CD">0</span>,<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>]]
+
+split = get_split(dataset)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;Split: [X%d &lt; %.3f]&#39;</span> % ((split[<span style="color: #CD5555">&#39;index&#39;</span>]+<span style="color: #B452CD">1</span>), split[<span style="color: #CD5555">&#39;value&#39;</span>]))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
+
+<section>
+<h2 id="regression-trees">Regression trees </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #228B22"># Quadratic training set + noise</span>
+np.random.seed(<span style="color: #B452CD">42</span>)
+m = <span style="color: #B452CD">200</span>
+X = np.random.rand(m, <span style="color: #B452CD">1</span>)
+y = <span style="color: #B452CD">4</span> * (X - <span style="color: #B452CD">0.5</span>) ** <span style="color: #B452CD">2</span>
+y = y + np.random.randn(m, <span style="color: #B452CD">1</span>) / <span style="color: #B452CD">10</span>
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeRegressor
+
+tree_reg = DecisionTreeRegressor(max_depth=<span style="color: #B452CD">2</span>, random_state=<span style="color: #B452CD">42</span>)
+tree_reg.fit(X, y)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
+
+<section>
+<h2 id="final-regressor-code">Final regressor code </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeRegressor
+
+tree_reg1 = DecisionTreeRegressor(random_state=<span style="color: #B452CD">42</span>, max_depth=<span style="color: #B452CD">2</span>)
+tree_reg2 = DecisionTreeRegressor(random_state=<span style="color: #B452CD">42</span>, max_depth=<span style="color: #B452CD">3</span>)
+tree_reg1.fit(X, y)
+tree_reg2.fit(X, y)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">plot_regression_predictions</span>(tree_reg, X, y, axes=[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, -<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1</span>], ylabel=<span style="color: #CD5555">&quot;$y$&quot;</span>):
+    x1 = np.linspace(axes[<span style="color: #B452CD">0</span>], axes[<span style="color: #B452CD">1</span>], <span style="color: #B452CD">500</span>).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
+    y_pred = tree_reg.predict(x1)
+    plt.axis(axes)
+    plt.xlabel(<span style="color: #CD5555">&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
+    <span style="color: #8B008B; font-weight: bold">if</span> ylabel:
+        plt.ylabel(ylabel, fontsize=<span style="color: #B452CD">18</span>, rotation=<span style="color: #B452CD">0</span>)
+    plt.plot(X, y, <span style="color: #CD5555">&quot;b.&quot;</span>)
+    plt.plot(x1, y_pred, <span style="color: #CD5555">&quot;r.-&quot;</span>, linewidth=<span style="color: #B452CD">2</span>, label=<span style="color: #CD5555">r&quot;$\hat{y}$&quot;</span>)
+
+plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
+plt.subplot(<span style="color: #B452CD">121</span>)
+plot_regression_predictions(tree_reg1, X, y)
+<span style="color: #8B008B; font-weight: bold">for</span> split, style <span style="color: #8B008B">in</span> ((<span style="color: #B452CD">0.1973</span>, <span style="color: #CD5555">&quot;k-&quot;</span>), (<span style="color: #B452CD">0.0917</span>, <span style="color: #CD5555">&quot;k--&quot;</span>), (<span style="color: #B452CD">0.7718</span>, <span style="color: #CD5555">&quot;k--&quot;</span>)):
+    plt.plot([split, split], [-<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1</span>], style, linewidth=<span style="color: #B452CD">2</span>)
+plt.text(<span style="color: #B452CD">0.21</span>, <span style="color: #B452CD">0.65</span>, <span style="color: #CD5555">&quot;Depth=0&quot;</span>, fontsize=<span style="color: #B452CD">15</span>)
+plt.text(<span style="color: #B452CD">0.01</span>, <span style="color: #B452CD">0.2</span>, <span style="color: #CD5555">&quot;Depth=1&quot;</span>, fontsize=<span style="color: #B452CD">13</span>)
+plt.text(<span style="color: #B452CD">0.65</span>, <span style="color: #B452CD">0.8</span>, <span style="color: #CD5555">&quot;Depth=1&quot;</span>, fontsize=<span style="color: #B452CD">13</span>)
+plt.legend(loc=<span style="color: #CD5555">&quot;upper center&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
+plt.title(<span style="color: #CD5555">&quot;max_depth=2&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
+
+plt.subplot(<span style="color: #B452CD">122</span>)
+plot_regression_predictions(tree_reg2, X, y, ylabel=<span style="color: #8B008B; font-weight: bold">None</span>)
+<span style="color: #8B008B; font-weight: bold">for</span> split, style <span style="color: #8B008B">in</span> ((<span style="color: #B452CD">0.1973</span>, <span style="color: #CD5555">&quot;k-&quot;</span>), (<span style="color: #B452CD">0.0917</span>, <span style="color: #CD5555">&quot;k--&quot;</span>), (<span style="color: #B452CD">0.7718</span>, <span style="color: #CD5555">&quot;k--&quot;</span>)):
+    plt.plot([split, split], [-<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1</span>], style, linewidth=<span style="color: #B452CD">2</span>)
+<span style="color: #8B008B; font-weight: bold">for</span> split <span style="color: #8B008B">in</span> (<span style="color: #B452CD">0.0458</span>, <span style="color: #B452CD">0.1298</span>, <span style="color: #B452CD">0.2873</span>, <span style="color: #B452CD">0.9040</span>):
+    plt.plot([split, split], [-<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;k:&quot;</span>, linewidth=<span style="color: #B452CD">1</span>)
+plt.text(<span style="color: #B452CD">0.3</span>, <span style="color: #B452CD">0.5</span>, <span style="color: #CD5555">&quot;Depth=2&quot;</span>, fontsize=<span style="color: #B452CD">13</span>)
+plt.title(<span style="color: #CD5555">&quot;max_depth=3&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
+
+plt.show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;">tree_reg1 = DecisionTreeRegressor(random_state=<span style="color: #B452CD">42</span>)
+tree_reg2 = DecisionTreeRegressor(random_state=<span style="color: #B452CD">42</span>, min_samples_leaf=<span style="color: #B452CD">10</span>)
+tree_reg1.fit(X, y)
+tree_reg2.fit(X, y)
+
+x1 = np.linspace(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">500</span>).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
+y_pred1 = tree_reg1.predict(x1)
+y_pred2 = tree_reg2.predict(x1)
+
+plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
+
+plt.subplot(<span style="color: #B452CD">121</span>)
+plt.plot(X, y, <span style="color: #CD5555">&quot;b.&quot;</span>)
+plt.plot(x1, y_pred1, <span style="color: #CD5555">&quot;r.-&quot;</span>, linewidth=<span style="color: #B452CD">2</span>, label=<span style="color: #CD5555">r&quot;$\hat{y}$&quot;</span>)
+plt.axis([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, -<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1.1</span>])
+plt.xlabel(<span style="color: #CD5555">&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
+plt.ylabel(<span style="color: #CD5555">&quot;$y$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>, rotation=<span style="color: #B452CD">0</span>)
+plt.legend(loc=<span style="color: #CD5555">&quot;upper center&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
+plt.title(<span style="color: #CD5555">&quot;No restrictions&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
+
+plt.subplot(<span style="color: #B452CD">122</span>)
+plt.plot(X, y, <span style="color: #CD5555">&quot;b.&quot;</span>)
+plt.plot(x1, y_pred2, <span style="color: #CD5555">&quot;r.-&quot;</span>, linewidth=<span style="color: #B452CD">2</span>, label=<span style="color: #CD5555">r&quot;$\hat{y}$&quot;</span>)
+plt.axis([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, -<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1.1</span>])
+plt.xlabel(<span style="color: #CD5555">&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
+plt.title(<span style="color: #CD5555">&quot;min_samples_leaf={}&quot;</span>.format(tree_reg2.min_samples_leaf), fontsize=<span style="color: #B452CD">14</span>)
+
+plt.show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
+
+<section>
+<h2 id="pros-and-cons-of-trees-pros">Pros and cons of trees, pros </h2>
+
+<ul>
+<p><li> White box, easy to interpret model. Some people believe that decision trees more closely mirror human decision-making than do the regression and classification approaches discussed earlier (think of support vector machines)</li>
+<p><li> Trees are very easy to explain to people. In fact, they are even easier to explain than linear regression!</li>
+<p><li> No feature normalization needed</li>
+<p><li> Tree models can handle both continuous and categorical data (Classification and Regression Trees)</li>
+<p><li> Can model nonlinear relationships</li>
+<p><li> Can model interactions between the different descriptive features</li>
+<p><li> Trees can be displayed graphically, and are easily interpreted even by a non-expert (especially if they are small)</li>
+</ul>
+</section>
+
+<section>
+<h2 id="disadvantages">Disadvantages </h2>
+
+<ul>
+<p><li> Unfortunately, trees generally do not have the same level of predictive accuracy as some of the other regression and classification approaches</li>
+<p><li> If continuous features are used the tree may become quite large and hence less interpretable</li>
+<p><li> Decision trees are prone to overfit the training data and hence do not well generalize the data if no stopping criteria or improvements like pruning, boosting or bagging are implemented</li>
+<p><li> Small changes in the data may lead to a completely different tree. This issue can be addressed by using ensemble methods like bagging, boosting or random forests</li>
+<p><li> Unbalanced datasets where some target feature values occur much more frequently than others may lead to biased trees since the frequently occurring feature values are preferred over the less frequently occurring ones.</li> 
+<p><li> If the number of features is relatively large (high dimensional) and the number of instances is relatively low, the tree might overfit the data</li>
+<p><li> Features with many levels may be preferred over features with less levels since for them it is <em>more easy</em> to split the dataset such that the sub datasets only contain pure target feature values. This issue can be addressed by preferring for instance the information gain ratio as splitting criteria over information gain</li>
+</ul>
+<p>
+<p>However, by aggregating many decision trees, using methods like
+bagging, random forests, and boosting, the predictive performance of
+trees can be substantially improved.
+</p>
+</section>
+
+<section>
+<h2 id="ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods </h2>
+
+<p>As stated above and seen in many of the examples discussed here about
+a single decision tree, we often end up overfitting our training
+data. This normally means that we have a high variance. Can we reduce
+the variance of a statistical learning method?
+</p>
+
+<p>This leads us to a set of different methods that can combine different
+machine learning algorithms or just use one of them to construct
+forests and jungles of trees, homogeneous ones or heterogenous
+ones. These methods are recognized by different names which we will
+try to explain here. These are
+</p>
+
+<ol>
+<p><li> Voting classifiers</li>
+<p><li> Bagging and Pasting</li>
+<p><li> Random forests</li>
+<p><li> Boosting methods, from adaptive to Extreme Gradient Boosting (XGBoost)</li>
+</ol>
+<p>
+<p>We discuss these methods here.</p>
+</section>
+
+<section>
+<h2 id="an-overview-of-ensemble-methods">An Overview of Ensemble Methods  </h2>
+
+<br/><br/>
+<center>
+<p><img src="DataFiles/ensembleoverview.png" width="600" align="bottom"></p>
+</center>
+<br/><br/>
+</section>
+
+<section>
+<h2 id="why-voting">Why Voting? </h2>
+
+<p>The idea behind boosting, and voting as well can be phrased as follows:
+<b>Can a group of people somehow arrive at highly
+reasoned decisions, despite the weak judgement of the individual
+members?</b>
+</p>
+
+<p>The aim is to create a good classifier by combining several weak classifiers.
+<b>A weak classifier is a classifier which is able to produce results that are only slightly better than guessing at random.</b>
+</p>
+
+<p>The basic approach is to apply repeatedly (in boosting this is done in an iterative way) a weak classifier to modifications of the data.
+In voting we simply apply the law of large numbers while in boosting we give more weight to misclassified data in
+each iteration.  
+</p>
+
+<p>Decision trees play an important role as our weak classifier. They serve as the basic method. </p>
+</section>
+
+<section>
+<h2 id="tossing-coins">Tossing coins </h2>
+
+<p>The simplest case is a so-called voting ensemble. To illustrate this,
+think of yourself tossing coins with a biased outcome of 51 per cent
+for heads and 49% for tails. With only few tosses,
+you may not clearly see this distribution for heads and tails. However, after some
+thousands of tosses, there will be a clear majority of heads.  With 2000 tosses
+you should see approximately 1020 heads and 980 tails.
+</p>
+
+<p>We can then state that the outcome is a clear majority of heads. If
+you do this ten thousand times, it is easy to see that there is a 97%
+likelihood of a majority of heads.
+</p>
+
+<p>Another example would be to collect all polls before an
+election. Different polls may show different likelihoods for a
+candidate winning with say a majority  of the popular vote. The majority vote
+would then consist in many polls indicating that this candidate will
+actually win.
+</p>
+
+<p>The example here shows how we can implement the coin tossing case,
+clealry demostrating that after some tosses we see the <a href="https://en.wikipedia.org/wiki/Law_of_large_numbers" target="_blank">law of large</a>
+numbers kicking in.
+</p>
+</section>
+
+<section>
+<h2 id="standard-imports-first">Standard imports first </h2>
+
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #228B22"># Common imports</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> Image 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_graphviz
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler, OneHotEncoder
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.compose</span> <span style="color: #8B008B; font-weight: bold">import</span> ColumnTransformer
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> Image 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
+
+<span style="color: #228B22"># Where to save the figures and data files</span>
+PROJECT_ROOT_DIR = <span style="color: #CD5555">&quot;Results&quot;</span>
+FIGURE_ID = <span style="color: #CD5555">&quot;Results/FigureFiles&quot;</span>
+DATA_ID = <span style="color: #CD5555">&quot;DataFiles/&quot;</span>
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(PROJECT_ROOT_DIR):
+    os.mkdir(PROJECT_ROOT_DIR)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(FIGURE_ID):
+    os.makedirs(FIGURE_ID)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(DATA_ID):
+    os.makedirs(DATA_ID)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">image_path</span>(fig_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(FIGURE_ID, fig_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">data_path</span>(dat_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(DATA_ID, dat_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">save_fig</span>(fig_id):
+    plt.savefig(image_path(fig_id) + <span style="color: #CD5555">&quot;.png&quot;</span>, <span style="color: #658b00">format</span>=<span style="color: #CD5555">&#39;png&#39;</span>)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
+
+<section>
+<h2 id="simple-voting-example-head-or-tail">Simple Voting Example, head or tail </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #228B22"># Common imports</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">matplotlib.colors</span> <span style="color: #8B008B; font-weight: bold">import</span> ListedColormap
 plt.rcParams[<span style="color: #CD5555">&#39;axes.labelsize&#39;</span>] = <span style="color: #B452CD">14</span>
 plt.rcParams[<span style="color: #CD5555">&#39;xtick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
 plt.rcParams[<span style="color: #CD5555">&#39;ytick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
 
+heads_proba = <span style="color: #B452CD">0.51</span>
+coin_tosses = (np.random.rand(<span style="color: #B452CD">10000</span>, <span style="color: #B452CD">10</span>) &lt; heads_proba).astype(np.int32)
+cumulative_heads_ratio = np.cumsum(coin_tosses, axis=<span style="color: #B452CD">0</span>) / np.arange(<span style="color: #B452CD">1</span>, <span style="color: #B452CD">10001</span>).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
+plt.figure(figsize=(<span style="color: #B452CD">8</span>,<span style="color: #B452CD">3.5</span>))
+plt.plot(cumulative_heads_ratio)
+plt.plot([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">10000</span>], [<span style="color: #B452CD">0.51</span>, <span style="color: #B452CD">0.51</span>], <span style="color: #CD5555">&quot;k--&quot;</span>, linewidth=<span style="color: #B452CD">2</span>, label=<span style="color: #CD5555">&quot;51%&quot;</span>)
+plt.plot([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">10000</span>], [<span style="color: #B452CD">0.5</span>, <span style="color: #B452CD">0.5</span>], <span style="color: #CD5555">&quot;k-&quot;</span>, label=<span style="color: #CD5555">&quot;50%&quot;</span>)
+plt.xlabel(<span style="color: #CD5555">&quot;Number of coin tosses&quot;</span>)
+plt.ylabel(<span style="color: #CD5555">&quot;Heads ratio&quot;</span>)
+plt.legend(loc=<span style="color: #CD5555">&quot;lower right&quot;</span>)
+plt.axis([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">10000</span>, <span style="color: #B452CD">0.42</span>, <span style="color: #B452CD">0.58</span>])
+save_fig(<span style="color: #CD5555">&quot;votingsimple&quot;</span>)
+plt.show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
 
+<section>
+<h2 id="using-the-voting-classifier">Using the Voting Classifier </h2>
+
+<p>We can use the voting classifier on other data sets, here the exciting binary case of two distinct objects using the make moons functionality of <b>Scikit-Learn</b>.</p>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> make_moons
+
+X, y = make_moons(n_samples=<span style="color: #B452CD">500</span>, noise=<span style="color: #B452CD">0.30</span>, random_state=<span style="color: #B452CD">42</span>)
+X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=<span style="color: #B452CD">42</span>)
+
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> RandomForestClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> VotingClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> LogisticRegression
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets
+
+log_clf = LogisticRegression(solver=<span style="color: #CD5555">&quot;liblinear&quot;</span>, random_state=<span style="color: #B452CD">42</span>)
+rnd_clf = RandomForestClassifier(n_estimators=<span style="color: #B452CD">10</span>, random_state=<span style="color: #B452CD">42</span>)
+svm_clf = SVC(gamma=<span style="color: #CD5555">&quot;auto&quot;</span>, random_state=<span style="color: #B452CD">42</span>)
+
+voting_clf = VotingClassifier(
+    estimators=[(<span style="color: #CD5555">&#39;lr&#39;</span>, log_clf), (<span style="color: #CD5555">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #CD5555">&#39;svc&#39;</span>, svm_clf)],
+    voting=<span style="color: #CD5555">&#39;hard&#39;</span>)
+
+voting_clf.fit(X_train, y_train)
+
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #8B008B; font-weight: bold">for</span> clf <span style="color: #8B008B">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    <span style="color: #658b00">print</span>(clf.<span style="color: #00688B">__class__</span>.<span style="color: #00688B">__name__</span>, accuracy_score(y_test, y_pred))
+
+log_clf = LogisticRegression(solver=<span style="color: #CD5555">&quot;liblinear&quot;</span>, random_state=<span style="color: #B452CD">42</span>)
+rnd_clf = RandomForestClassifier(n_estimators=<span style="color: #B452CD">10</span>, random_state=<span style="color: #B452CD">42</span>)
+svm_clf = SVC(gamma=<span style="color: #CD5555">&quot;auto&quot;</span>, probability=<span style="color: #8B008B; font-weight: bold">True</span>, random_state=<span style="color: #B452CD">42</span>)
+voting_clf = VotingClassifier(
+    estimators=[(<span style="color: #CD5555">&#39;lr&#39;</span>, log_clf), (<span style="color: #CD5555">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #CD5555">&#39;svc&#39;</span>, svm_clf)],
+    voting=<span style="color: #CD5555">&#39;soft&#39;</span>)
+voting_clf.fit(X_train, y_train)
+
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #8B008B; font-weight: bold">for</span> clf <span style="color: #8B008B">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    <span style="color: #658b00">print</span>(clf.<span style="color: #00688B">__class__</span>.<span style="color: #00688B">__name__</span>, accuracy_score(y_test, y_pred))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
+
+<section>
+<h2 id="voting-and-bagging">Voting and Bagging </h2>
 
 
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> make_moons
+
+X, y = make_moons(n_samples=<span style="color: #B452CD">500</span>, noise=<span style="color: #B452CD">0.30</span>, random_state=<span style="color: #B452CD">42</span>)
+X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=<span style="color: #B452CD">42</span>)
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> RandomForestClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> VotingClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> LogisticRegression
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
+
+log_clf = LogisticRegression(random_state=<span style="color: #B452CD">42</span>)
+rnd_clf = RandomForestClassifier(random_state=<span style="color: #B452CD">42</span>)
+svm_clf = SVC(random_state=<span style="color: #B452CD">42</span>)
+
+voting_clf = VotingClassifier(
+    estimators=[(<span style="color: #CD5555">&#39;lr&#39;</span>, log_clf), (<span style="color: #CD5555">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #CD5555">&#39;svc&#39;</span>, svm_clf)],
+    voting=<span style="color: #CD5555">&#39;hard&#39;</span>)
+voting_clf.fit(X_train, y_train)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #8B008B; font-weight: bold">for</span> clf <span style="color: #8B008B">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    <span style="color: #658b00">print</span>(clf.<span style="color: #00688B">__class__</span>.<span style="color: #00688B">__name__</span>, accuracy_score(y_test, y_pred))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;">log_clf = LogisticRegression(random_state=<span style="color: #B452CD">42</span>)
+rnd_clf = RandomForestClassifier(random_state=<span style="color: #B452CD">42</span>)
+svm_clf = SVC(probability=<span style="color: #8B008B; font-weight: bold">True</span>, random_state=<span style="color: #B452CD">42</span>)
+
+voting_clf = VotingClassifier(
+    estimators=[(<span style="color: #CD5555">&#39;lr&#39;</span>, log_clf), (<span style="color: #CD5555">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #CD5555">&#39;svc&#39;</span>, svm_clf)],
+    voting=<span style="color: #CD5555">&#39;soft&#39;</span>)
+voting_clf.fit(X_train, y_train)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #8B008B; font-weight: bold">for</span> clf <span style="color: #8B008B">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    <span style="color: #658b00">print</span>(clf.<span style="color: #00688B">__class__</span>.<span style="color: #00688B">__name__</span>, accuracy_score(y_test, y_pred))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
+
+<section>
+<h2 id="bagging">Bagging </h2>
+
+<p>The <b>plain</b> decision trees suffer from high
+variance. This means that if we split the training data into two parts
+at random, and fit a decision tree to both halves, the results that we
+get could be quite different. In contrast, a procedure with low
+variance will yield similar results if applied repeatedly to distinct
+data sets; linear regression tends to have low variance, if the ratio
+of \( n \) to \( p \) is moderately large. 
+</p>
+
+<p><b>Bootstrap aggregation</b>, or just <b>bagging</b>, is a
+general-purpose procedure for reducing the variance of a statistical
+learning method. 
+</p>
+</section>
+
+<section>
+<h2 id="more-bagging">More bagging </h2>
+
+<p>Bagging typically results in improved accuracy
+over prediction using a single tree. Unfortunately, however, it can be
+difficult to interpret the resulting model. Recall that one of the
+advantages of decision trees is the attractive and easily interpreted
+diagram that results.
+</p>
+
+<p>However, when we bag a large number of trees, it is no longer
+possible to represent the resulting statistical learning procedure
+using a single tree, and it is no longer clear which variables are
+most important to the procedure. Thus, bagging improves prediction
+accuracy at the expense of interpretability.  Although the collection
+of bagged trees is much more difficult to interpret than a single
+tree, one can obtain an overall summary of the importance of each
+predictor using the MSE (for bagging regression trees) or the Gini
+index (for bagging classification trees). In the case of bagging
+regression trees, we can record the total amount that the MSE is
+decreased due to splits over a given predictor, averaged over all \( B \) possible
+trees. A large value indicates an important predictor. Similarly, in
+the context of bagging classification trees, we can add up the total
+amount that the Gini index  is decreased by splits over a given
+predictor, averaged over all \( B \) trees.
+</p>
+</section>
+
+<section>
+<h2 id="making-your-own-bootstrap-changing-the-level-of-the-decision-tree">Making your own Bootstrap: Changing the Level of the Decision Tree </h2>
+
+<p>Let us bring up our good old boostrap example from the linear regression lectures. We change the linerar regression algorithm with
+a decision tree wth different depths and perform a bootstrap aggregate (in this case we perform as many bootstraps as data points \( n \)).
+</p>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.pipeline</span> <span style="color: #8B008B; font-weight: bold">import</span> make_pipeline
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.utils</span> <span style="color: #8B008B; font-weight: bold">import</span> resample
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeRegressor
+
+n = <span style="color: #B452CD">100</span>
+n_boostraps = <span style="color: #B452CD">100</span>
+maxdepth = <span style="color: #B452CD">8</span>
+
+<span style="color: #228B22"># Make data set.</span>
+x = np.linspace(-<span style="color: #B452CD">3</span>, <span style="color: #B452CD">3</span>, n).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
+y = np.exp(-x**<span style="color: #B452CD">2</span>) + <span style="color: #B452CD">1.5</span> * np.exp(-(x-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>)+ np.random.normal(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0.1</span>, x.shape)
+error = np.zeros(maxdepth)
+bias = np.zeros(maxdepth)
+variance = np.zeros(maxdepth)
+polydegree = np.zeros(maxdepth)
+X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=<span style="color: #B452CD">0.2</span>)
 
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.pipeline</span> <span style="color: #8B008B; font-weight: bold">import</span> Pipeline
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> LinearSVC
+scaler = StandardScaler()
+scaler.fit(X_train)
+X_train_scaled = scaler.transform(X_train)
+X_test_scaled = scaler.transform(X_test)
 
+<span style="color: #228B22"># we produce a simple tree first as benchmark</span>
+simpletree = DecisionTreeRegressor(max_depth=<span style="color: #B452CD">3</span>) 
+simpletree.fit(X_train_scaled, y_train)
+simpleprediction = simpletree.predict(X_test_scaled)
+<span style="color: #8B008B; font-weight: bold">for</span> degree <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(<span style="color: #B452CD">1</span>,maxdepth):
+    model = DecisionTreeRegressor(max_depth=degree) 
+    y_pred = np.empty((y_test.shape[<span style="color: #B452CD">0</span>], n_boostraps))
+    <span style="color: #8B008B; font-weight: bold">for</span> i <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(n_boostraps):
+        x_, y_ = resample(X_train_scaled, y_train)
+        model.fit(x_, y_)
+        y_pred[:, i] = model.predict(X_test_scaled)<span style="color: #228B22">#.ravel()</span>
 
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> make_moons
-X, y = make_moons(n_samples=<span style="color: #B452CD">100</span>, noise=<span style="color: #B452CD">0.15</span>, random_state=<span style="color: #B452CD">42</span>)
-
-<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">plot_dataset</span>(X, y, axes):
-    plt.plot(X[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">0</span>], X[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">0</span>], <span style="color: #CD5555">&quot;bs&quot;</span>)
-    plt.plot(X[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">1</span>], X[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;g^&quot;</span>)
-    plt.axis(axes)
-    plt.grid(<span style="color: #8B008B; font-weight: bold">True</span>, which=<span style="color: #CD5555">&#39;both&#39;</span>)
-    plt.xlabel(<span style="color: #CD5555">r&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-    plt.ylabel(<span style="color: #CD5555">r&quot;$x_2$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>, rotation=<span style="color: #B452CD">0</span>)
-
-plot_dataset(X, y, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plt.show()
-
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> make_moons
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.pipeline</span> <span style="color: #8B008B; font-weight: bold">import</span> Pipeline
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> PolynomialFeatures
-
-polynomial_svm_clf = Pipeline([
-        (<span style="color: #CD5555">&quot;poly_features&quot;</span>, PolynomialFeatures(degree=<span style="color: #B452CD">3</span>)),
-        (<span style="color: #CD5555">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #CD5555">&quot;svm_clf&quot;</span>, LinearSVC(C=<span style="color: #B452CD">10</span>, loss=<span style="color: #CD5555">&quot;hinge&quot;</span>, random_state=<span style="color: #B452CD">42</span>))
-    ])
-
-polynomial_svm_clf.fit(X, y)
-
-<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">plot_predictions</span>(clf, axes):
-    x0s = np.linspace(axes[<span style="color: #B452CD">0</span>], axes[<span style="color: #B452CD">1</span>], <span style="color: #B452CD">100</span>)
-    x1s = np.linspace(axes[<span style="color: #B452CD">2</span>], axes[<span style="color: #B452CD">3</span>], <span style="color: #B452CD">100</span>)
-    x0, x1 = np.meshgrid(x0s, x1s)
-    X = np.c_[x0.ravel(), x1.ravel()]
-    y_pred = clf.predict(X).reshape(x0.shape)
-    y_decision = clf.decision_function(X).reshape(x0.shape)
-    plt.contourf(x0, x1, y_pred, cmap=plt.cm.brg, alpha=<span style="color: #B452CD">0.2</span>)
-    plt.contourf(x0, x1, y_decision, cmap=plt.cm.brg, alpha=<span style="color: #B452CD">0.1</span>)
-
-plot_predictions(polynomial_svm_clf, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plot_dataset(X, y, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-
-plt.show()
-
-
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
-
-poly_kernel_svm_clf = Pipeline([
-        (<span style="color: #CD5555">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #CD5555">&quot;svm_clf&quot;</span>, SVC(kernel=<span style="color: #CD5555">&quot;poly&quot;</span>, degree=<span style="color: #B452CD">3</span>, coef0=<span style="color: #B452CD">1</span>, C=<span style="color: #B452CD">5</span>))
-    ])
-poly_kernel_svm_clf.fit(X, y)
-
-poly100_kernel_svm_clf = Pipeline([
-        (<span style="color: #CD5555">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #CD5555">&quot;svm_clf&quot;</span>, SVC(kernel=<span style="color: #CD5555">&quot;poly&quot;</span>, degree=<span style="color: #B452CD">10</span>, coef0=<span style="color: #B452CD">100</span>, C=<span style="color: #B452CD">5</span>))
-    ])
-poly100_kernel_svm_clf.fit(X, y)
-
-plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
-
-plt.subplot(<span style="color: #B452CD">121</span>)
-plot_predictions(poly_kernel_svm_clf, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plot_dataset(X, y, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plt.title(<span style="color: #CD5555">r&quot;$d=3, r=1, C=5$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
-
-plt.subplot(<span style="color: #B452CD">122</span>)
-plot_predictions(poly100_kernel_svm_clf, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plot_dataset(X, y, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plt.title(<span style="color: #CD5555">r&quot;$d=10, r=100, C=5$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
-
-plt.show()
-
-<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">gaussian_rbf</span>(x, landmark, gamma):
-    <span style="color: #8B008B; font-weight: bold">return</span> np.exp(-gamma * np.linalg.norm(x - landmark, axis=<span style="color: #B452CD">1</span>)**<span style="color: #B452CD">2</span>)
-
-gamma = <span style="color: #B452CD">0.3</span>
-
-x1s = np.linspace(-<span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">200</span>).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
-x2s = gaussian_rbf(x1s, -<span style="color: #B452CD">2</span>, gamma)
-x3s = gaussian_rbf(x1s, <span style="color: #B452CD">1</span>, gamma)
-
-XK = np.c_[gaussian_rbf(X1D, -<span style="color: #B452CD">2</span>, gamma), gaussian_rbf(X1D, <span style="color: #B452CD">1</span>, gamma)]
-yk = np.array([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>])
-
-plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
-
-plt.subplot(<span style="color: #B452CD">121</span>)
-plt.grid(<span style="color: #8B008B; font-weight: bold">True</span>, which=<span style="color: #CD5555">&#39;both&#39;</span>)
-plt.axhline(y=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.scatter(x=[-<span style="color: #B452CD">2</span>, <span style="color: #B452CD">1</span>], y=[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>], s=<span style="color: #B452CD">150</span>, alpha=<span style="color: #B452CD">0.5</span>, c=<span style="color: #CD5555">&quot;red&quot;</span>)
-plt.plot(X1D[:, <span style="color: #B452CD">0</span>][yk==<span style="color: #B452CD">0</span>], np.zeros(<span style="color: #B452CD">4</span>), <span style="color: #CD5555">&quot;bs&quot;</span>)
-plt.plot(X1D[:, <span style="color: #B452CD">0</span>][yk==<span style="color: #B452CD">1</span>], np.zeros(<span style="color: #B452CD">5</span>), <span style="color: #CD5555">&quot;g^&quot;</span>)
-plt.plot(x1s, x2s, <span style="color: #CD5555">&quot;g--&quot;</span>)
-plt.plot(x1s, x3s, <span style="color: #CD5555">&quot;b:&quot;</span>)
-plt.gca().get_yaxis().set_ticks([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0.25</span>, <span style="color: #B452CD">0.5</span>, <span style="color: #B452CD">0.75</span>, <span style="color: #B452CD">1</span>])
-plt.xlabel(<span style="color: #CD5555">r&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.ylabel(<span style="color: #CD5555">r&quot;Similarity&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
-plt.annotate(<span style="color: #CD5555">r&#39;$\mathbf{x}$&#39;</span>,
-             xy=(X1D[<span style="color: #B452CD">3</span>, <span style="color: #B452CD">0</span>], <span style="color: #B452CD">0</span>),
-             xytext=(-<span style="color: #B452CD">0.5</span>, <span style="color: #B452CD">0.20</span>),
-             ha=<span style="color: #CD5555">&quot;center&quot;</span>,
-             arrowprops=<span style="color: #658b00">dict</span>(facecolor=<span style="color: #CD5555">&#39;black&#39;</span>, shrink=<span style="color: #B452CD">0.1</span>),
-             fontsize=<span style="color: #B452CD">18</span>,
-            )
-plt.text(-<span style="color: #B452CD">2</span>, <span style="color: #B452CD">0.9</span>, <span style="color: #CD5555">&quot;$x_2$&quot;</span>, ha=<span style="color: #CD5555">&quot;center&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.text(<span style="color: #B452CD">1</span>, <span style="color: #B452CD">0.9</span>, <span style="color: #CD5555">&quot;$x_3$&quot;</span>, ha=<span style="color: #CD5555">&quot;center&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.axis([-<span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">4.5</span>, -<span style="color: #B452CD">0.1</span>, <span style="color: #B452CD">1.1</span>])
-
-plt.subplot(<span style="color: #B452CD">122</span>)
-plt.grid(<span style="color: #8B008B; font-weight: bold">True</span>, which=<span style="color: #CD5555">&#39;both&#39;</span>)
-plt.axhline(y=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.axvline(x=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.plot(XK[:, <span style="color: #B452CD">0</span>][yk==<span style="color: #B452CD">0</span>], XK[:, <span style="color: #B452CD">1</span>][yk==<span style="color: #B452CD">0</span>], <span style="color: #CD5555">&quot;bs&quot;</span>)
-plt.plot(XK[:, <span style="color: #B452CD">0</span>][yk==<span style="color: #B452CD">1</span>], XK[:, <span style="color: #B452CD">1</span>][yk==<span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;g^&quot;</span>)
-plt.xlabel(<span style="color: #CD5555">r&quot;$x_2$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.ylabel(<span style="color: #CD5555">r&quot;$x_3$  &quot;</span>, fontsize=<span style="color: #B452CD">20</span>, rotation=<span style="color: #B452CD">0</span>)
-plt.annotate(<span style="color: #CD5555">r&#39;$\phi\left(\mathbf{x}\right)$&#39;</span>,
-             xy=(XK[<span style="color: #B452CD">3</span>, <span style="color: #B452CD">0</span>], XK[<span style="color: #B452CD">3</span>, <span style="color: #B452CD">1</span>]),
-             xytext=(<span style="color: #B452CD">0.65</span>, <span style="color: #B452CD">0.50</span>),
-             ha=<span style="color: #CD5555">&quot;center&quot;</span>,
-             arrowprops=<span style="color: #658b00">dict</span>(facecolor=<span style="color: #CD5555">&#39;black&#39;</span>, shrink=<span style="color: #B452CD">0.1</span>),
-             fontsize=<span style="color: #B452CD">18</span>,
-            )
-plt.plot([-<span style="color: #B452CD">0.1</span>, <span style="color: #B452CD">1.1</span>], [<span style="color: #B452CD">0.57</span>, -<span style="color: #B452CD">0.1</span>], <span style="color: #CD5555">&quot;r--&quot;</span>, linewidth=<span style="color: #B452CD">3</span>)
-plt.axis([-<span style="color: #B452CD">0.1</span>, <span style="color: #B452CD">1.1</span>, -<span style="color: #B452CD">0.1</span>, <span style="color: #B452CD">1.1</span>])
-    
-plt.subplots_adjust(right=<span style="color: #B452CD">1</span>)
-
-plt.show()
-
-
-x1_example = X1D[<span style="color: #B452CD">3</span>, <span style="color: #B452CD">0</span>]
-<span style="color: #8B008B; font-weight: bold">for</span> landmark <span style="color: #8B008B">in</span> (-<span style="color: #B452CD">2</span>, <span style="color: #B452CD">1</span>):
-    k = gaussian_rbf(np.array([[x1_example]]), np.array([[landmark]]), gamma)
-    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Phi({}, {}) = {}&quot;</span>.format(x1_example, landmark, k))
-
-rbf_kernel_svm_clf = Pipeline([
-        (<span style="color: #CD5555">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #CD5555">&quot;svm_clf&quot;</span>, SVC(kernel=<span style="color: #CD5555">&quot;rbf&quot;</span>, gamma=<span style="color: #B452CD">5</span>, C=<span style="color: #B452CD">0.001</span>))
-    ])
-rbf_kernel_svm_clf.fit(X, y)
-
-
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
-
-gamma1, gamma2 = <span style="color: #B452CD">0.1</span>, <span style="color: #B452CD">5</span>
-C1, C2 = <span style="color: #B452CD">0.001</span>, <span style="color: #B452CD">1000</span>
-hyperparams = (gamma1, C1), (gamma1, C2), (gamma2, C1), (gamma2, C2)
-
-svm_clfs = []
-<span style="color: #8B008B; font-weight: bold">for</span> gamma, C <span style="color: #8B008B">in</span> hyperparams:
-    rbf_kernel_svm_clf = Pipeline([
-            (<span style="color: #CD5555">&quot;scaler&quot;</span>, StandardScaler()),
-            (<span style="color: #CD5555">&quot;svm_clf&quot;</span>, SVC(kernel=<span style="color: #CD5555">&quot;rbf&quot;</span>, gamma=gamma, C=C))
-        ])
-    rbf_kernel_svm_clf.fit(X, y)
-    svm_clfs.append(rbf_kernel_svm_clf)
-
-plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">7</span>))
-
-<span style="color: #8B008B; font-weight: bold">for</span> i, svm_clf <span style="color: #8B008B">in</span> <span style="color: #658b00">enumerate</span>(svm_clfs):
-    plt.subplot(<span style="color: #B452CD">221</span> + i)
-    plot_predictions(svm_clf, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-    plot_dataset(X, y, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-    gamma, C = hyperparams[i]
-    plt.title(<span style="color: #CD5555">r&quot;$\gamma = {}, C = {}$&quot;</span>.format(gamma, C), fontsize=<span style="color: #B452CD">16</span>)
-
+    polydegree[degree] = degree
+    error[degree] = np.mean( np.mean((y_test - y_pred)**<span style="color: #B452CD">2</span>, axis=<span style="color: #B452CD">1</span>, keepdims=<span style="color: #8B008B; font-weight: bold">True</span>) )
+    bias[degree] = np.mean( (y_test - np.mean(y_pred, axis=<span style="color: #B452CD">1</span>, keepdims=<span style="color: #8B008B; font-weight: bold">True</span>))**<span style="color: #B452CD">2</span> )
+    variance[degree] = np.mean( np.var(y_pred, axis=<span style="color: #B452CD">1</span>, keepdims=<span style="color: #8B008B; font-weight: bold">True</span>) )
+    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;Polynomial degree:&#39;</span>, degree)
+    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;Error:&#39;</span>, error[degree])
+    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;Bias^2:&#39;</span>, bias[degree])
+    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;Var:&#39;</span>, variance[degree])
+    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;{} &gt;= {} + {} = {}&#39;</span>.format(error[degree], bias[degree], variance[degree], bias[degree]+variance[degree]))
+ 
+mse_simpletree= np.mean( np.mean((y_test - simpleprediction)**<span style="color: #B452CD">2</span>))
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Simple tree:&quot;</span>,mse_simpletree)
+plt.xlim(<span style="color: #B452CD">1</span>,maxdepth)
+plt.plot(polydegree, error, label=<span style="color: #CD5555">&#39;MSE&#39;</span>)
+plt.plot(polydegree, bias, label=<span style="color: #CD5555">&#39;bias&#39;</span>)
+plt.plot(polydegree, variance, label=<span style="color: #CD5555">&#39;Variance&#39;</span>)
+plt.legend()
+save_fig(<span style="color: #CD5555">&quot;baggingboot&quot;</span>)
 plt.show()
 </pre>
 </div>
@@ -1426,43 +2104,567 @@ plt.show()
 </section>
 
 <section>
-<h2 id="mathematical-optimization-of-convex-functions">Mathematical optimization of convex functions </h2>
+<h2 id="random-forests">Random forests </h2>
+
+<p>Random forests provide an improvement over bagged trees by way of a
+small tweak that decorrelates the trees. 
+</p>
+
+<p>As in bagging, we build a
+number of decision trees on bootstrapped training samples. But when
+building these decision trees, each time a split in a tree is
+considered, a random sample of \( m \) predictors is chosen as split
+candidates from the full set of \( p \) predictors. The split is allowed to
+use only one of those \( m \) predictors. 
+</p>
+
+<p>A fresh sample of \( m \) predictors is
+taken at each split, and typically we choose 
+</p>
 
-<p>A mathematical (quadratic) optimization problem, or just optimization problem, has the form</p>
 <p>&nbsp;<br>
 $$
-\begin{align*}
-    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\boldsymbol{\lambda}^T\boldsymbol{P}\boldsymbol{\lambda}+\boldsymbol{q}^T\boldsymbol{\lambda},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{\lambda} \preceq \boldsymbol{h} \wedge  \boldsymbol{A}\boldsymbol{\lambda}=f.
-\end{align*}
+m\approx \sqrt{p}.
 $$
 <p>&nbsp;<br>
 
-<p>subject to some constraints for say a selected set \( i=1,2,\dots, n \).
-In our case we are optimizing with respect to the Lagrangian multipliers \( \lambda_i \), and the
-vector \( \boldsymbol{\lambda}=[\lambda_1, \lambda_2,\dots, \lambda_n] \) is the optimization variable we are dealing with.
+<p>In building a random forest, at
+each split in the tree, the algorithm is not even allowed to consider
+a majority of the available predictors. 
 </p>
 
-<p>In our case we are particularly interested in a class of optimization problems called convex optmization problems. 
-In our discussion on gradient descent methods we discussed at length the definition of a convex function. 
+<p>The reason for this is rather clever. Suppose that there is one very
+strong predictor in the data set, along with a number of other
+moderately strong predictors. Then in the collection of bagged
+variable importance random forest trees, most or all of the trees will
+use this strong predictor in the top split. Consequently, all of the
+bagged trees will look quite similar to each other. Hence the
+predictions from the bagged trees will be highly correlated.
+Unfortunately, averaging many highly correlated quantities does not
+lead to as large of a reduction in variance as averaging many
+uncorrelated quantities. In particular, this means that bagging will
+not lead to a substantial reduction in variance over a single tree in
+this setting.
 </p>
-
-<p>Convex optimization problems play a central role in applied mathematics and we recommend strongly <a href="http://web.stanford.edu/~boyd/cvxbook/" target="_blank">Boyd and Vandenberghe's text on the topics</a>.</p>
 </section>
 
 <section>
-<h2 id="how-do-we-solve-these-problems">How do we solve these problems? </h2>
+<h2 id="random-forest-algorithm">Random Forest Algorithm </h2>
+<p>The algorithm described here can be applied to both classification and regression problems.</p>
 
-<p>If we use Python as programming language and wish to venture beyond
-<b>scikit-learn</b>, <b>tensorflow</b> and similar software which makes our
-lives so much easier, we need to dive into the wonderful world of
-quadratic programming. We can, if we wish, solve the minimization
-problem using say standard gradient methods or conjugate gradient
-methods. However, these methods tend to exhibit a rather slow
-converge. So, welcome to the promised land of quadratic programming.
+<p>We will grow of forest of say \( B \) trees.</p>
+<ol>
+<p><li> For \( b=1:B \)</li>
+<ul>
+
+<p><li> Draw a bootstrap sample from the training data organized in our \( \boldsymbol{X} \) matrix.</li>
+
+<p><li> We grow then a random forest tree \( T_b \) based on the bootstrapped data by repeating the steps outlined till we reach the maximum node size is reached</li>
+<ol>
+
+<p><li> we select \( m \le p \) variables at random from the \( p \) predictors/features</li>
+
+<p><li> pick the best split point among the \( m \) features using for example the CART algorithm and create a new node</li>
+
+<p><li> split the node into daughter nodes</li>
+</ol>
+<p>
+</ul>
+<p>
+<p><li> Output then the ensemble of trees \( \{T_b\}_1^{B} \) and make predictions for either a regression type of problem or a classification type of problem.</li> 
+</ol>
+</section>
+
+<section>
+<h2 id="random-forests-compared-with-other-methods-on-the-cancer-data">Random Forests Compared with other Methods on the Cancer Data </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span>  train_test_split 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> load_breast_cancer
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> LogisticRegression
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> BaggingClassifier
+
+<span style="color: #228B22"># Load the data</span>
+cancer = load_breast_cancer()
+
+X_train, X_test, y_train, y_test = train_test_split(cancer.data,cancer.target,random_state=<span style="color: #B452CD">0</span>)
+<span style="color: #658b00">print</span>(X_train.shape)
+<span style="color: #658b00">print</span>(X_test.shape)
+<span style="color: #228B22">#define methods</span>
+<span style="color: #228B22"># Logistic Regression</span>
+logreg = LogisticRegression(solver=<span style="color: #CD5555">&#39;lbfgs&#39;</span>)
+<span style="color: #228B22"># Support vector machine</span>
+svm = SVC(gamma=<span style="color: #CD5555">&#39;auto&#39;</span>, C=<span style="color: #B452CD">100</span>)
+<span style="color: #228B22"># Decision Trees</span>
+deep_tree_clf = DecisionTreeClassifier(max_depth=<span style="color: #8B008B; font-weight: bold">None</span>)
+<span style="color: #228B22">#Scale the data</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler
+scaler = StandardScaler()
+scaler.fit(X_train)
+X_train_scaled = scaler.transform(X_train)
+X_test_scaled = scaler.transform(X_test)
+<span style="color: #228B22"># Logistic Regression</span>
+logreg.fit(X_train_scaled, y_train)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Test set accuracy Logistic Regression with scaled data: {:.2f}&quot;</span>.format(logreg.score(X_test_scaled,y_test)))
+<span style="color: #228B22"># Support Vector Machine</span>
+svm.fit(X_train_scaled, y_train)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Test set accuracy SVM with scaled data: {:.2f}&quot;</span>.format(logreg.score(X_test_scaled,y_test)))
+<span style="color: #228B22"># Decision Trees</span>
+deep_tree_clf.fit(X_train_scaled, y_train)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Test set accuracy with Decision Trees and scaled data: {:.2f}&quot;</span>.format(deep_tree_clf.score(X_test_scaled,y_test)))
+
+
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> RandomForestClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> LabelEncoder
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> cross_validate
+<span style="color: #228B22"># Data set not specificied</span>
+<span style="color: #228B22">#Instantiate the model with 500 trees and entropy as splitting criteria</span>
+Random_Forest_model = RandomForestClassifier(n_estimators=<span style="color: #B452CD">500</span>,criterion=<span style="color: #CD5555">&quot;entropy&quot;</span>)
+Random_Forest_model.fit(X_train_scaled, y_train)
+<span style="color: #228B22">#Cross validation</span>
+accuracy = cross_validate(Random_Forest_model,X_test_scaled,y_test,cv=<span style="color: #B452CD">10</span>)[<span style="color: #CD5555">&#39;test_score&#39;</span>]
+<span style="color: #658b00">print</span>(accuracy)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Test set accuracy with Random Forests and scaled data: {:.2f}&quot;</span>.format(Random_Forest_model.score(X_test_scaled,y_test)))
+
+
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">scikitplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">skplt</span>
+y_pred = Random_Forest_model.predict(X_test_scaled)
+skplt.metrics.plot_confusion_matrix(y_test, y_pred, normalize=<span style="color: #8B008B; font-weight: bold">True</span>)
+plt.show()
+y_probas = Random_Forest_model.predict_proba(X_test_scaled)
+skplt.metrics.plot_roc(y_test, y_probas)
+plt.show()
+skplt.metrics.plot_cumulative_gain(y_test, y_probas)
+plt.show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+<p>Recall that the cumulative gains curve shows the percentage of the
+overall number of cases in a given category <em>gained</em> by targeting a
+percentage of the total number of cases.
 </p>
 
-<p>The functions we need are contained in the <a href="https://cvxopt.org/userguide/coneprog.html" target="_blank">quadratic programming library CVXOPT</a> and we need to import it together with <b>numpy</b> as</p>
+<p>Similarly, the receiver operating characteristic curve, or ROC curve,
+displays the diagnostic ability of a binary classifier system as its
+discrimination threshold is varied. It plots the true positive rate against the false positive rate.
+</p>
+</section>
+
+<section>
+<h2 id="compare-bagging-on-trees-with-random-forests">Compare  Bagging on Trees with Random Forests  </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;">bag_clf = BaggingClassifier(
+    DecisionTreeClassifier(splitter=<span style="color: #CD5555">&quot;random&quot;</span>, max_leaf_nodes=<span style="color: #B452CD">16</span>, random_state=<span style="color: #B452CD">42</span>),
+    n_estimators=<span style="color: #B452CD">500</span>, max_samples=<span style="color: #B452CD">1.0</span>, bootstrap=<span style="color: #8B008B; font-weight: bold">True</span>, n_jobs=-<span style="color: #B452CD">1</span>, random_state=<span style="color: #B452CD">42</span>)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="font-size: 80%; line-height: 125%;">bag_clf.fit(X_train, y_train)
+y_pred = bag_clf.predict(X_test)
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> RandomForestClassifier
+rnd_clf = RandomForestClassifier(n_estimators=<span style="color: #B452CD">500</span>, max_leaf_nodes=<span style="color: #B452CD">16</span>, n_jobs=-<span style="color: #B452CD">1</span>, random_state=<span style="color: #B452CD">42</span>)
+rnd_clf.fit(X_train, y_train)
+y_pred_rf = rnd_clf.predict(X_test)
+np.sum(y_pred == y_pred_rf) / <span style="color: #658b00">len</span>(y_pred) 
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+</section>
+
+<section>
+<h2 id="boosting-a-bird-s-eye-view">Boosting, a Bird's Eye View </h2>
+
+<p>The basic idea is to combine weak classifiers in order to create a good
+classifier. With a weak classifier we often intend a classifier which
+produces results which are only slightly better than we would get by
+random guesses.
+</p>
+
+<p>This is done by applying in an iterative way a weak (or a standard
+classifier like decision trees) to modify the data. In each iteration
+we emphasize those observations which are misclassified by weighting
+them with a factor.
+</p>
+</section>
+
+<section>
+<h2 id="what-is-boosting-additive-modelling-iterative-fitting">What is boosting? Additive Modelling/Iterative Fitting </h2>
+
+<p>Boosting is a way of fitting an additive expansion in a set of
+elementary basis functions like for example some simple polynomials.
+Assume for example that we have a function
+</p>
+<p>&nbsp;<br>
+$$
+f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m),
+$$
+<p>&nbsp;<br>
+
+<p>where \( \beta_m \) are the expansion parameters to be determined in a
+minimization process and \( b(x;\gamma_m) \) are some simple functions of
+the multivariable parameter \( x \) which is characterized by the
+parameters \( \gamma_m \).
+</p>
+
+<p>As an example, consider the Sigmoid function we used in logistic
+regression. In that case, we can translate the function
+\( b(x;\gamma_m) \) into the Sigmoid function
+</p>
+
+<p>&nbsp;<br>
+$$
+\sigma(t) = \frac{1}{1+\exp{(-t)}},
+$$
+<p>&nbsp;<br>
+
+<p>where \( t=\gamma_0+\gamma_1 x \) and the parameters \( \gamma_0 \) and
+\( \gamma_1 \) were determined by the Logistic Regression fitting
+algorithm.
+</p>
+
+<p>As another example, consider the cost function we defined for linear regression</p>
+<p>&nbsp;<br>
+$$
+C(\boldsymbol{y},\boldsymbol{f}) = \frac{1}{n} \sum_{i=0}^{n-1}(y_i-f(x_i))^2.
+$$
+<p>&nbsp;<br>
+
+<p>In this case the function \( f(x) \) was replaced by the design matrix
+\( \boldsymbol{X} \) and the unknown linear regression parameters \( \boldsymbol{\beta} \),
+that is \( \boldsymbol{f}=\boldsymbol{X}\boldsymbol{\beta} \). In linear regression we can 
+simply invert a matrix and obtain the parameters \( \beta \) by
+</p>
+
+<p>&nbsp;<br>
+$$
+\boldsymbol{\beta}=\left(\boldsymbol{X}^T\boldsymbol{X}\right)^{-1}\boldsymbol{X}^T\boldsymbol{y}.
+$$
+<p>&nbsp;<br>
+
+<p>In iterative fitting or additive modeling, we minimize the cost function with respect to the parameters \( \beta_m \) and \( \gamma_m \).</p>
+</section>
+
+<section>
+<h2 id="iterative-fitting-regression-and-squared-error-cost-function">Iterative Fitting, Regression and Squared-error Cost Function </h2>
+
+<p>The way we proceed is as follows (here we specialize to the squared-error cost function)</p>
+
+<ol>
+<p><li> Establish a cost function, here \( C(\boldsymbol{y},\boldsymbol{f}) = \frac{1}{n} \sum_{i=0}^{n-1}(y_i-f_M(x_i))^2 \) with \( f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m) \).</li>
+<p><li> Initialize with a guess \( f_0(x) \). It could be one or even zero or some random numbers.</li>
+<p><li> For \( m=1:M \)
+<ol type="a"></li>
+ <p><li> minimize \( \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2 \) wrt \( \gamma \) and \( \beta \)</li>
+ <p><li> This gives the optimal values \( \beta_m \) and \( \gamma_m \)</li>
+ <p><li> Determine then the new values \( f_m(x)=f_{m-1}(x) +\beta_m b(x;\gamma_m) \)</li>
+</ol>
+<p>
+</ol>
+<p>
+<p>We could use any of the algorithms we have discussed till now. If we
+use trees, \( \gamma \) parameterizes the split variables and split points
+at the internal nodes, and the predictions at the terminal nodes.
+</p>
+</section>
+
+<section>
+<h2 id="squared-error-example-and-iterative-fitting">Squared-Error Example and Iterative Fitting </h2>
+
+<p>To better understand what happens, let us develop the steps for the iterative fitting using the above squared error function.</p>
+
+<p>For simplicity we assume also that our functions \( b(x;\gamma)=1+\gamma x \). </p>
+
+<p>This means that for every iteration \( m \), we need to optimize</p>
+
+<p>&nbsp;<br>
+$$
+(\beta_m,\gamma_m) = \mathrm{argmin}_{\beta,\lambda}\hspace{0.1cm} \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2=\sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta(1+\gamma x_i))^2.
+$$
+<p>&nbsp;<br>
+
+<p>We start our iteration by simply setting \( f_0(x)=0 \). 
+Taking the derivatives  with respect to \( \beta \) and \( \gamma \) we obtain
+</p>
+<p>&nbsp;<br>
+$$
+\frac{\partial {\cal C}}{\partial \beta} = -2\sum_{i}(1+\gamma x_i)(y_i-\beta(1+\gamma x_i))=0,
+$$
+<p>&nbsp;<br>
+
+<p>and</p>
+<p>&nbsp;<br>
+$$
+\frac{\partial {\cal C}}{\partial \gamma} =-2\sum_{i}\beta x_i(y_i-\beta(1+\gamma x_i))=0.
+$$
+<p>&nbsp;<br>
+
+<p>We can then rewrite these equations as (defining \( \boldsymbol{w}=\boldsymbol{e}+\gamma \boldsymbol{x}) \) with \( \boldsymbol{e} \) being the unit vector)</p>
+<p>&nbsp;<br>
+$$
+\gamma \boldsymbol{w}^T(\boldsymbol{y}-\beta\gamma \boldsymbol{w})=0,
+$$
+<p>&nbsp;<br>
+
+<p>which gives us \( \beta = \boldsymbol{w}^T\boldsymbol{y}/(\boldsymbol{w}^T\boldsymbol{w}) \). Similarly we have </p>
+<p>&nbsp;<br>
+$$
+\beta\gamma \boldsymbol{x}^T(\boldsymbol{y}-\beta(1+\gamma \boldsymbol{x}))=0,
+$$
+<p>&nbsp;<br>
+
+<p>which leads to \( \gamma =(\boldsymbol{x}^T\boldsymbol{y}-\beta\boldsymbol{x}^T\boldsymbol{e})/(\beta\boldsymbol{x}^T\boldsymbol{x}) \).  Inserting
+for \( \beta \) gives us an equation for \( \gamma \). This is a non-linear equation in the unknown \( \gamma \) and has to be solved numerically. 
+</p>
+
+<p>The solution to these two equations gives us in turn \( \beta_1 \) and \( \gamma_1 \) leading to the new expression for \( f_1(x) \) as
+\( f_1(x) = \beta_1(1+\gamma_1x) \). Doing this \( M \) times results in our final estimate for the function \( f \). 
+</p>
+</section>
+
+<section>
+<h2 id="iterative-fitting-classification-and-adaboost">Iterative Fitting, Classification and AdaBoost </h2>
+
+<p>Let us consider a binary classification problem with two outcomes \( y_i \in \{-1,1\} \) and \( i=0,1,2,\dots,n-1 \) as our set of
+observations. We define a classification function \( G(x) \) which produces a prediction taking one or the other of the two values 
+\( \{-1,1\} \).
+</p>
+
+<p>The error rate of the training sample is then</p>
+
+<p>&nbsp;<br>
+$$
+\mathrm{\overline{err}}=\frac{1}{n} \sum_{i=0}^{n-1} I(y_i\ne G(x_i)). 
+$$
+<p>&nbsp;<br>
+
+<p>The iterative procedure starts with defining a weak classifier whose
+error rate is barely better than random guessing.  The iterative
+procedure in boosting is to sequentially apply a  weak
+classification algorithm to repeatedly modified versions of the data
+producing a sequence of weak classifiers \( G_m(x) \).
+</p>
+
+<p>Here we will express our  function \( f(x) \) in terms of \( G(x) \). That is</p>
+<p>&nbsp;<br>
+$$
+f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m),
+$$
+<p>&nbsp;<br>
+
+<p>will be a function of </p>
+<p>&nbsp;<br>
+$$
+G_M(x) = \mathrm{sign} \sum_{i=1}^M \alpha_m G_m(x).
+$$
+<p>&nbsp;<br>
+</section>
+
+<section>
+<h2 id="adaptive-boosting-adaboost">Adaptive Boosting, AdaBoost </h2>
+
+<p>In our iterative procedure we define thus</p>
+<p>&nbsp;<br>
+$$
+f_m(x) = f_{m-1}(x)+\beta_mG_m(x).
+$$
+<p>&nbsp;<br>
+
+<p>The simplest possible cost function which leads (also simple from a computational point of view) to the AdaBoost algorithm is the
+exponential cost/loss function defined as
+</p>
+<p>&nbsp;<br>
+$$
+C(\boldsymbol{y},\boldsymbol{f}) = \sum_{i=0}^{n-1}\exp{(-y_i(f_{m-1}(x_i)+\beta G(x_i))}.
+$$
+<p>&nbsp;<br>
+
+<p>We optimize \( \beta \) and \( G \) for each value of \( m=1:M \) as we did in the regression case.
+This is normally done in two steps. Let us however first rewrite the cost function as 
+</p>
+
+<p>&nbsp;<br>
+$$
+C(\boldsymbol{y},\boldsymbol{f}) = \sum_{i=0}^{n-1}w_i^{m}\exp{(-y_i\beta G(x_i))},
+$$
+<p>&nbsp;<br>
+
+<p>where we have defined \( w_i^m= \exp{(-y_if_{m-1}(x_i))} \).</p>
+</section>
+
+<section>
+<h2 id="building-up-adaboost">Building up AdaBoost </h2>
+
+<p>First, for any \( \beta > 0 \), we optimize \( G \) by setting</p>
+<p>&nbsp;<br>
+$$
+G_m(x) = \mathrm{sign} \sum_{i=0}^{n-1} w_i^m I(y_i \ne G_(x_i)),
+$$
+<p>&nbsp;<br>
+
+<p>which is the classifier that minimizes the weighted error rate in predicting \( y \).</p>
+
+<p>We can do this by rewriting</p>
+<p>&nbsp;<br>
+$$
+\exp{-(\beta)}\sum_{y_i=G(x_i)}w_i^m+\exp{(\beta)}\sum_{y_i\ne G(x_i)}w_i^m,
+$$
+<p>&nbsp;<br>
+
+<p>which can be rewritten as</p>
+<p>&nbsp;<br>
+$$
+(\exp{(\beta)}-\exp{-(\beta)})\sum_{i=0}^{n-1}w_i^mI(y_i\ne G(x_i))+\exp{(-\beta)}\sum_{i=0}^{n-1}w_i^m=0,
+$$
+<p>&nbsp;<br>
+
+<p>which leads to</p>
+<p>&nbsp;<br>
+$$
+\beta_m = \frac{1}{2}\log{\frac{1-\mathrm{\overline{err}}}{\mathrm{\overline{err}}}},
+$$
+<p>&nbsp;<br>
+
+<p>where we have redefined the error as </p>
+<p>&nbsp;<br>
+$$
+\mathrm{\overline{err}}_m=\frac{1}{n}\frac{\sum_{i=0}^{n-1}w_i^mI(y_i\ne G(x_i)}{\sum_{i=0}^{n-1}w_i^m},
+$$
+<p>&nbsp;<br>
+
+<p>which leads to an update of</p>
+<p>&nbsp;<br>
+$$
+f_m(x) = f_{m-1}(x) +\beta_m G_m(x).
+$$
+<p>&nbsp;<br>
+
+<p>This leads to the new weights</p>
+<p>&nbsp;<br>
+$$
+w_i^{m+1} = w_i^m \exp{(-y_i\beta_m G_m(x_i))}
+$$
+<p>&nbsp;<br>
+</section>
+
+<section>
+<h2 id="adaptive-boosting-adaboost-basic-algorithm">Adaptive boosting: AdaBoost, Basic Algorithm </h2>
+
+<p>The algorithm here is rather straightforward. Assume that our weak
+classifier is a decision tree and we consider a binary set of outputs
+with \( y_i \in \{-1,1\} \) and \( i=0,1,2,\dots,n-1 \) as our set of
+observations. Our design matrix is given in terms of the
+feature/predictor vectors
+\( \boldsymbol{X}=[\boldsymbol{x}_0\boldsymbol{x}_1\dots\boldsymbol{x}_{p-1}] \). Finally, we define also a
+classifier determined by our data via a function \( G(x) \). This function tells us how well we are able to classify our outputs/targets \( \boldsymbol{y} \). 
+</p>
+
+<p>We have already defined the misclassification error \( \mathrm{err} \) as</p>
+<p>&nbsp;<br>
+$$
+\mathrm{err}=\frac{1}{n}\sum_{i=0}^{n-1}I(y_i\ne G(x_i)),
+$$
+<p>&nbsp;<br>
+
+<p>where the function \( I() \) is one if we misclassify and zero if we classify correctly. </p>
+</section>
+
+<section>
+<h2 id="basic-steps-of-adaboost">Basic Steps of AdaBoost </h2>
+
+<p>With the above definitions we are now ready to set up the algorithm for AdaBoost.
+The basic idea is to set up weights which will be used to scale the correctly classified and the misclassified cases.
+</p>
+<ol>
+<p><li> We start by initializing all weights to \( w_i = 1/n \), with \( i=0,1,2,\dots n-1 \). It is easy to see that we must have \( \sum_{i=0}^{n-1}w_i = 1 \).</li>
+<p><li> We rewrite the misclassification error as</li> 
+</ol>
+<p>
+<p>&nbsp;<br>
+$$
+\mathrm{\overline{err}}_m=\frac{\sum_{i=0}^{n-1}w_i^m I(y_i\ne G(x_i))}{\sum_{i=0}^{n-1}w_i},
+$$
+<p>&nbsp;<br>
+
+<ol>
+<p><li> Then we start looping over all attempts at classifying, namely we start an iterative process for \( m=1:M \), where \( M \) is the final number of classifications. Our given classifier could for example be a plain decision tree.
+<ol type="a"></li>
+ <p><li> Fit then a given classifier to the training set using the weights \( w_i \).</li>
+ <p><li> Compute then \( \mathrm{err} \) and figure out which events are classified properly and which are classified wrongly.</li>
+ <p><li> Define a quantity \( \alpha_{m} = \log{(1-\mathrm{\overline{err}}_m)/\mathrm{\overline{err}}_m} \)</li>
+ <p><li> Set the new weights to \( w_i = w_i\times \exp{(\alpha_m I(y_i\ne G(x_i)} \).</li>
+</ol>
+<p>
+<p><li> Compute the new classifier \( G(x)= \sum_{i=0}^{n-1}\alpha_m I(y_i\ne G(x_i) \).</li>
+</ol>
+<p>
+<p>For the iterations with \( m \le 2 \) the weights are modified
+individually at each steps. The observations which were misclassified
+at iteration \( m-1 \) have a weight which is larger than those which were
+classified properly. As this proceeds, the observations which were
+difficult to classifiy correctly are given a larger influence. Each
+new classification step \( m \) is then forced to concentrate on those
+observations that are missed in the previous iterations.
+</p>
+</section>
+
+<section>
+<h2 id="adaboost-examples">AdaBoost Examples </h2>
+
+<p>Using <b>Scikit-Learn</b> it is easy to apply the adaptive boosting algorithm, as done here.</p>
 
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
@@ -1471,8 +2673,27 @@ converge. So, welcome to the promised land of quadratic programming.
     <div class="inner_cell">
       <div class="input_area">
         <div class="highlight" style="background: #eeeedd">
-  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span>
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">cvxopt</span>
+  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> AdaBoostClassifier
+
+ada_clf = AdaBoostClassifier(
+    DecisionTreeClassifier(max_depth=<span style="color: #B452CD">1</span>), n_estimators=<span style="color: #B452CD">200</span>,
+    algorithm=<span style="color: #CD5555">&quot;SAMME.R&quot;</span>, learning_rate=<span style="color: #B452CD">0.5</span>, random_state=<span style="color: #B452CD">42</span>)
+ada_clf.fit(X_train, y_train)
+
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> AdaBoostClassifier
+
+ada_clf = AdaBoostClassifier(
+    DecisionTreeClassifier(max_depth=<span style="color: #B452CD">1</span>), n_estimators=<span style="color: #B452CD">200</span>,
+    algorithm=<span style="color: #CD5555">&quot;SAMME.R&quot;</span>, learning_rate=<span style="color: #B452CD">0.5</span>, random_state=<span style="color: #B452CD">42</span>)
+ada_clf.fit(X_train_scaled, y_train)
+y_pred = ada_clf.predict(X_test_scaled)
+skplt.metrics.plot_confusion_matrix(y_test, y_pred, normalize=<span style="color: #8B008B; font-weight: bold">True</span>)
+plt.show()
+y_probas = ada_clf.predict_proba(X_test_scaled)
+skplt.metrics.plot_roc(y_test, y_probas)
+plt.show()
+skplt.metrics.plot_cumulative_gain(y_test, y_probas)
+plt.show()
 </pre>
 </div>
       </div>
@@ -1487,134 +2708,6 @@ converge. So, welcome to the promised land of quadratic programming.
     </div>
   </div>
 </div>
-
-<p>This will make our life much easier. You don't need t0 write your own optimizer.</p>
-</section>
-
-<section>
-<h2 id="a-simple-example">A simple example </h2>
-
-<p>We remind ourselves about the general problem we want to solve</p>
-<p>&nbsp;<br>
-$$
-\begin{align*}
-    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}\boldsymbol{x}^T\boldsymbol{P}\boldsymbol{x}+\boldsymbol{q}^T\boldsymbol{x},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm} to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{x} \preceq \boldsymbol{h} \wedge  \boldsymbol{A}\boldsymbol{x}=f.
-\end{align*}
-$$
-<p>&nbsp;<br>
-
-<p>Let us show how to perform the optmization using a simple case. Assume we want to optimize the following problem</p>
-<p>&nbsp;<br>
-$$
-\begin{align*}
-    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}x^2+5x+3y \\ \nonumber
-    &\mathrm{subject}\hspace{0.5cm} \mathrm{to} \\ \nonumber
-    &x, y \geq 0 \\ \nonumber
-    &x+3y  \geq 15 \\ \nonumber
-    &2x+5y  \leq  100 \\ \nonumber
-    &3x+4y  \leq  80.  \\ \nonumber
-\end{align*}
-$$
-<p>&nbsp;<br>
-
-<p>The minimization problem can be rewritten in terms of vectors and matrices as (with \( x \) and \( y \) being the unknowns)</p>
-<p>&nbsp;<br>
-$$
-\frac{1}{2}\begin{bmatrix} x\\ y \end{bmatrix}^T   \begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}  \begin{bmatrix} x \\ y \end{bmatrix}  + \begin{bmatrix}3\\ 4  \end{bmatrix}^T \begin{bmatrix}x \\ y  \end{bmatrix}.  
-$$
-<p>&nbsp;<br>
-
-<p>Similarly, we can now set up the inequalities (we need to change \( \geq \) to \( \leq \) by multiplying with \( -1 \) on bot sides) as the following matrix-vector equation</p>
-<p>&nbsp;<br>
-$$
-\begin{bmatrix} -1 & 0 \\ 0 & -1 \\ -1 & -3 \\ 2 & 5 \\ 3 & 4\end{bmatrix}\begin{bmatrix} x \\ y\end{bmatrix} \preceq \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
-$$
-<p>&nbsp;<br>
-
-<p>We have collapsed all the inequalities into a single matrix \( \boldsymbol{G} \). We see also that our matrix </p>
-<p>&nbsp;<br>
-$$
-\boldsymbol{P} =\begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}
-$$
-<p>&nbsp;<br>
-
-<p>is clearly positive semi-definite (all eigenvalues larger or equal zero). 
-Finally, the vector \( \boldsymbol{h} \) is defined as 
-</p>
-<p>&nbsp;<br>
-$$
-\boldsymbol{h} = \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
-$$
-<p>&nbsp;<br>
-
-<p>Since we don't have any equalities the matrix \( \boldsymbol{A} \) is set to zero
-The following code solves the equations for us
-</p>
-
-<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #eeeedd">
-  <pre style="font-size: 80%; line-height: 125%;"><span style="color: #228B22"># Import the necessary packages</span>
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span>
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">cvxopt</span> <span style="color: #8B008B; font-weight: bold">import</span> matrix
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">cvxopt</span> <span style="color: #8B008B; font-weight: bold">import</span> solvers
-P = matrix(numpy.diag([<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>]), tc=<span style="color: #CD5555">&#39;d&#39;</span>)
-q = matrix(numpy.array([<span style="color: #B452CD">3</span>,<span style="color: #B452CD">4</span>]), tc=<span style="color: #CD5555">&#39;d&#39;</span>)
-G = matrix(numpy.array([[-<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>],[<span style="color: #B452CD">0</span>,-<span style="color: #B452CD">1</span>],[-<span style="color: #B452CD">1</span>,-<span style="color: #B452CD">3</span>],[<span style="color: #B452CD">2</span>,<span style="color: #B452CD">5</span>],[<span style="color: #B452CD">3</span>,<span style="color: #B452CD">4</span>]]), tc=<span style="color: #CD5555">&#39;d&#39;</span>)
-h = matrix(numpy.array([<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,-<span style="color: #B452CD">15</span>,<span style="color: #B452CD">100</span>,<span style="color: #B452CD">80</span>]), tc=<span style="color: #CD5555">&#39;d&#39;</span>)
-<span style="color: #228B22"># Construct the QP, invoke solver</span>
-sol = solvers.qp(P,q,G,h)
-<span style="color: #228B22"># Extract optimal value and solution</span>
-sol[<span style="color: #CD5555">&#39;x&#39;</span>] 
-sol[<span style="color: #CD5555">&#39;primal objective&#39;</span>]
-<span style="color: #658b00">print</span>(sol[<span style="color: #CD5555">&#39;x&#39;</span>] )
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
-</section>
-
-<section>
-<h2 id="back-to-the-more-realistic-cases">Back to the more realistic cases </h2>
-
-<p>We are now ready to return to our setup of the optmization problem for a more realistic case. Introducing the <b>slack</b> parameter \( C \) we have</p>
-<p>&nbsp;<br>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1K(\boldsymbol{x}_1,\boldsymbol{x}_1) & y_1y_2K(\boldsymbol{x}_1,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_1,\boldsymbol{x}_n) \\
-y_2y_1K(\boldsymbol{x}_2,\boldsymbol{x}_1) & y_2y_2K(\boldsymbol{x}_2,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_2,\boldsymbol{x}_n) \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1K(\boldsymbol{x}_n,\boldsymbol{x}_1) & y_ny_2K(\boldsymbol{x}_n\boldsymbol{x}_2) & \dots & \dots & y_ny_nK(\boldsymbol{x}_n,\boldsymbol{x}_n) \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{I}\boldsymbol{\lambda}, 
-$$
-<p>&nbsp;<br>
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
-With  the slack constants this leads to the additional constraint \( 0\leq \lambda_i \leq C \).
-</p>
-
-<p>Using the <b>CVXOPT</b> library, the matrix \( P \) would then be defined by the above matrix while the KKT conditions would all be collected by the matrix \( G \).</p>
-</section>
-
-<section>
-<h2 id="support-vector-machines-for-regression">Support vector machines for regression </h2>
-
-<p>Material will be added here.</p>
 </section>
 
 
diff --git a/doc/pub/week46/html/week46-solarized.html b/doc/pub/week46/html/week46-solarized.html
index 2184bd858..ee7edc6df 100644
--- a/doc/pub/week46/html/week46-solarized.html
+++ b/doc/pub/week46/html/week46-solarized.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --pygments_html_style=perldoc --html_style=sol
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <link href="https://cdn.rawgit.com/doconce/doconce/master/bundled/html_styles/style_solarized_box/css/solarized_light_code.css" rel="stylesheet" type="text/css" title="light"/>
 <script src="https://cdn.rawgit.com/doconce/doconce/master/bundled/html_styles/style_solarized_box/js/highlight.pack.js"></script>
 <script>hljs.initHighlightingOnLoad();</script>
@@ -63,68 +63,190 @@ div.toc p,a {
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -146,7 +268,7 @@ MathJax.Hub.Config({
 
 <!-- ------------------- main content ---------------------- -->
 <center>
-<h1>Week 46: Support Vector Machines and Project 3. </h1>
+<h1>Week 46: Decision Trees, Ensemble methods  and Random Forests</h1>
 </center>  <!-- document title -->
 
 <!-- author(s): Morten Hjorth-Jensen -->
@@ -162,704 +284,138 @@ MathJax.Hub.Config({
 </center>
 <br>
 <center>
-<h4>Nov 19, 2022</h4>
+<h4>Week 46, November 13-17</h4>
 </center> <!-- date -->
 <br>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="overview-of-week-46">Overview of week 46 </h2>
-
-<ul>
-<li> <b>Thursday</b>: Support Vector Machines, discussion of project 3</li>
-<ul>
-  <li> <a href="https://youtu.be/F3CkH-opbdY" target="_blank">Video of lecture</a></li>
-</ul>
-<li> <b>Friday</b>: Support vector machines and start Principal Component Analysis</li>
-<ul>
-  <li> <a href="https://youtu.be/BbupEEvMXtg" target="_blank">Video of lecture</a></li>
-</ul>
-</ul>
-<div class="alert alert-block alert-block alert-text-normal">
-<b>Reading</b>
-<p>
-<p>Reading recommendations:</p>
-<ol>
- <li> See lecture notes for week 46 at <a href="https://compphysics.github.io/MachineLearning/doc/web/course.html." target="_blank"><tt>https://compphysics.github.io/MachineLearning/doc/web/course.html.</tt></a></li>
- <li> Hastie et al chapter 12</li>
- <li> Bishop chapter 7.1 and 7.2</li>
-</ol>
-</div>
-
+<h2 id="plan-for-week-46">Plan for week 46 </h2>
 
 <div class="alert alert-block alert-block alert-text-normal">
-<b>Videos</b>
+<b>Active learning sessions on Tuesday and Wednesday</b>
 <p>
-<ol>
-<li> <a href="https://www.youtube.com/watch?v=efR1C6CvhmE&ab_channel=StatQuestwithJoshStarmer" target="_blank">Overview video on Support Vector Machines</a></li>
-<li> See also <a href="https://www.youtube.com/watch?v=N1vOgolbjSc&ab_channel=AliceZhao" target="_blank">this video</a>.</li>
-</ol>
+<ul>
+  <li> Work and Discussion of project 2</li>
+  <li> Discussion of project 3 as well</li>
+</ul>
+</div>
+  
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Material for the lecture on Thursday November 16, 2023</b>
+<p>
+<ul>
+  <li> Thursday: Basics of decision trees, classification and regression algorithms and ensemble models</li> 
+  <li> Readings and Videos:</li>
+<ul>
+    <li> These lecture notes</li>
+    <li> <a href="https://youtu.be/" target="_blank">Video of lecture to be added</a></li>
+    <li> <a href="https://www.youtube.com/watch?v=RmajweUFKvM&ab_channel=Simplilearn" target="_blank">Video on Decision trees</a></li>
+    <li> Decision Trees: Geron's chapter 6 covers decision trees while ensemble models, voting and bagging are discussed in chapter 7. See also lecture from <a href="https://www.uio.no/studier/emner/matnat/math/STK-IN4300/h20/slides/lecture_7.pdf" target="_blank">STK-IN4300, lecture 7</a>. Chapter 9.2 of Hastie et al contains also a good discussion.</li>
+</ul>
+</ul>
 </div>
 
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="support-vector-machines-overarching-aims">Support Vector Machines, overarching aims  </h2>
+<h2 id="decision-trees-overarching-aims">Decision trees, overarching aims  </h2>
 
-<p>A Support Vector Machine (SVM) is a very powerful and versatile
-Machine Learning method, capable of performing linear or nonlinear
-classification, regression, and even outlier detection. It is one of
-the most popular models in Machine Learning, and anyone interested in
-Machine Learning should have it in their toolbox. SVMs are
-particularly well suited for classification of complex but small-sized or
-medium-sized datasets.  
+<p>We start here with the most basic algorithm, the so-called decision
+tree. With this basic algorithm we can in turn build more complex
+networks, spanning from homogeneous and heterogenous forests (bagging,
+random forests and more) to one of the most popular supervised
+algorithms nowadays, the extreme gradient boosting, or just
+XGBoost. But let us start with the simplest possible ingredient.
 </p>
 
-<p>The case with two well-separated classes only can be understood in an
-intuitive way in terms of lines in a two-dimensional space separating
-the two classes (see figure below).
+<p>Decision trees are supervised learning algorithms used for both,
+classification and regression tasks.
 </p>
 
-<p>The basic mathematics behind the SVM is however less familiar to most of us. 
-It relies on the definition of hyperplanes and the
-definition of a <b>margin</b> which separates classes (in case of
-classification problems) of variables. It is also used for regression
-problems.
+<p>The main idea of decision trees
+is to find those descriptive features which contain the most
+<b>information</b> regarding the target feature and then split the dataset
+along the values of these features such that the target feature values
+for the resulting underlying datasets are as pure as possible.
 </p>
 
-<p>With SVMs we distinguish between hard margin and soft margins. The
-latter introduces a so-called softening parameter to be discussed
-below.  We distinguish also between linear and non-linear
-approaches. The latter are the most frequent ones since it is rather
-unlikely that we can separate classes easily by say straight lines.
+<p>The descriptive features which reproduce best the target/output features are normally  said
+to be the most informative ones. The process of finding the <b>most
+informative</b> feature is done until we accomplish a stopping criteria
+where we then finally end up in so called <b>leaf nodes</b>. 
 </p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="hyperplanes-and-all-that">Hyperplanes and all that </h2>
+<h2 id="basics-of-a-tree">Basics of a tree </h2>
 
-<p>The theory behind support vector machines (SVM hereafter) is based on
-the mathematical description of so-called hyperplanes. Let us start
-with a two-dimensional case. This will also allow us to introduce our
-first SVM examples. These will be tailored to the case of two specific
-classes, as displayed in the figure here based on the usage of the petal data.
+<p>A decision tree is typically divided into a <b>root node</b>, the <b>interior nodes</b>,
+and the final <b>leaf nodes</b> or just <b>leaves</b>. These entities are then connected by so-called <b>branches</b>.
 </p>
 
-<p>We assume here that our data set can be well separated into two
-domains, where a straight line does the job in the separating the two
-classes. Here the two classes are represented by either squares or
-circles.
-</p>
-
-<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #eeeedd">
-  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC, LinearSVC
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> SGDClassifier
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib</span>
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
-plt.rcParams[<span style="color: #CD5555">&#39;axes.labelsize&#39;</span>] = <span style="color: #B452CD">14</span>
-plt.rcParams[<span style="color: #CD5555">&#39;xtick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
-plt.rcParams[<span style="color: #CD5555">&#39;ytick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
-
-
-iris = datasets.load_iris()
-X = iris[<span style="color: #CD5555">&quot;data&quot;</span>][:, (<span style="color: #B452CD">2</span>, <span style="color: #B452CD">3</span>)]  <span style="color: #228B22"># petal length, petal width</span>
-y = iris[<span style="color: #CD5555">&quot;target&quot;</span>]
-
-setosa_or_versicolor = (y == <span style="color: #B452CD">0</span>) | (y == <span style="color: #B452CD">1</span>)
-X = X[setosa_or_versicolor]
-y = y[setosa_or_versicolor]
-
-
-
-C = <span style="color: #B452CD">5</span>
-alpha = <span style="color: #B452CD">1</span> / (C * <span style="color: #658b00">len</span>(X))
-
-lin_clf = LinearSVC(loss=<span style="color: #CD5555">&quot;hinge&quot;</span>, C=C, random_state=<span style="color: #B452CD">42</span>)
-svm_clf = SVC(kernel=<span style="color: #CD5555">&quot;linear&quot;</span>, C=C)
-sgd_clf = SGDClassifier(loss=<span style="color: #CD5555">&quot;hinge&quot;</span>, learning_rate=<span style="color: #CD5555">&quot;constant&quot;</span>, eta0=<span style="color: #B452CD">0.001</span>, alpha=alpha,
-                        max_iter=<span style="color: #B452CD">100000</span>, random_state=<span style="color: #B452CD">42</span>)
-
-scaler = StandardScaler()
-X_scaled = scaler.fit_transform(X)
-
-lin_clf.fit(X_scaled, y)
-svm_clf.fit(X_scaled, y)
-sgd_clf.fit(X_scaled, y)
-
-<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;LinearSVC:                   &quot;</span>, lin_clf.intercept_, lin_clf.coef_)
-<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;SVC:                         &quot;</span>, svm_clf.intercept_, svm_clf.coef_)
-<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;SGDClassifier(alpha={:.5f}):&quot;</span>.format(sgd_clf.alpha), sgd_clf.intercept_, sgd_clf.coef_)
-
-<span style="color: #228B22"># Compute the slope and bias of each decision boundary</span>
-w1 = -lin_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>]/lin_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-b1 = -lin_clf.intercept_[<span style="color: #B452CD">0</span>]/lin_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-w2 = -svm_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>]/svm_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-b2 = -svm_clf.intercept_[<span style="color: #B452CD">0</span>]/svm_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-w3 = -sgd_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>]/sgd_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-b3 = -sgd_clf.intercept_[<span style="color: #B452CD">0</span>]/sgd_clf.coef_[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>]
-
-<span style="color: #228B22"># Transform the decision boundary lines back to the original scale</span>
-line1 = scaler.inverse_transform([[-<span style="color: #B452CD">10</span>, -<span style="color: #B452CD">10</span> * w1 + b1], [<span style="color: #B452CD">10</span>, <span style="color: #B452CD">10</span> * w1 + b1]])
-line2 = scaler.inverse_transform([[-<span style="color: #B452CD">10</span>, -<span style="color: #B452CD">10</span> * w2 + b2], [<span style="color: #B452CD">10</span>, <span style="color: #B452CD">10</span> * w2 + b2]])
-line3 = scaler.inverse_transform([[-<span style="color: #B452CD">10</span>, -<span style="color: #B452CD">10</span> * w3 + b3], [<span style="color: #B452CD">10</span>, <span style="color: #B452CD">10</span> * w3 + b3]])
-
-<span style="color: #228B22"># Plot all three decision boundaries</span>
-plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
-plt.plot(line1[:, <span style="color: #B452CD">0</span>], line1[:, <span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;k:&quot;</span>, label=<span style="color: #CD5555">&quot;LinearSVC&quot;</span>)
-plt.plot(line2[:, <span style="color: #B452CD">0</span>], line2[:, <span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;b--&quot;</span>, linewidth=<span style="color: #B452CD">2</span>, label=<span style="color: #CD5555">&quot;SVC&quot;</span>)
-plt.plot(line3[:, <span style="color: #B452CD">0</span>], line3[:, <span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;r-&quot;</span>, label=<span style="color: #CD5555">&quot;SGDClassifier&quot;</span>)
-plt.plot(X[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">1</span>], X[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;bs&quot;</span>) <span style="color: #228B22"># label=&quot;Iris-Versicolor&quot;</span>
-plt.plot(X[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">0</span>], X[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">0</span>], <span style="color: #CD5555">&quot;yo&quot;</span>) <span style="color: #228B22"># label=&quot;Iris-Setosa&quot;</span>
-plt.xlabel(<span style="color: #CD5555">&quot;Petal length&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
-plt.ylabel(<span style="color: #CD5555">&quot;Petal width&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
-plt.legend(loc=<span style="color: #CD5555">&quot;upper center&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
-plt.axis([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">5.5</span>, <span style="color: #B452CD">0</span>, <span style="color: #B452CD">2</span>])
-
-plt.show()
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
-
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="what-is-a-hyperplane">What is a hyperplane? </h2>
-
-<p>The aim of the SVM algorithm is to find a hyperplane in a
-\( p \)-dimensional space, where \( p \) is the number of features that
-distinctly classifies the data points.
-</p>
-
-<p>In a \( p \)-dimensional space, a hyperplane is what we call an affine subspace of dimension of \( p-1 \).
-As an example, in two dimension, a hyperplane is simply as straight line while in three dimensions it is 
-a two-dimensional subspace, or stated simply, a plane. 
-</p>
-
-<p>In two dimensions, with the variables \( x_1 \) and \( x_2 \), the hyperplane is defined as</p>
-$$
-b+w_1x_1+w_2x_2=0,
-$$
-
-<p>where \( b \) is the intercept and \( w_1 \) and \( w_2 \) define the elements of a vector orthogonal to the line 
-\( b+w_1x_1+w_2x_2=0 \). 
-In two dimensions we define the vectors \( \boldsymbol{x} =[x1,x2] \) and \( \boldsymbol{w}=[w1,w2] \). 
-We can then rewrite the above equation as 
-</p>
-
-$$
-\boldsymbol{x}^T\boldsymbol{w}+b=0.
-$$
-
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="a-p-dimensional-space-of-features">A \( p \)-dimensional space of features </h2>
-
-<p>We limit ourselves to two classes of outputs \( y_i \) and assign these classes the values \( y_i = \pm 1 \). 
-In a \( p \)-dimensional space of say \( p \) features we have a hyperplane defines as 
-</p>
-$$
-b+wx_1+w_2x_2+\dots +w_px_p=0.
-$$
-
-<p>If we define a 
-matrix \( \boldsymbol{X}=\left[\boldsymbol{x}_1,\boldsymbol{x}_2,\dots, \boldsymbol{x}_p\right] \)
-of dimension \( n\times p \), where \( n \) represents the observations for each feature and each vector \( x_i \) is a column vector of the matrix \( \boldsymbol{X} \), 
-</p>
-$$
-\boldsymbol{x}_i = \begin{bmatrix} x_{i1} \\ x_{i2} \\ \dots \\ \dots \\ x_{ip} \end{bmatrix}.
-$$
-
-<p>If the above condition is not met for a given vector \( \boldsymbol{x}_i \) we have </p>
-$$
-b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} >0,
-$$
-
-<p>if our output \( y_i=1 \).
-In this case we say that \( \boldsymbol{x}_i \) lies on one of the sides of the hyperplane and if 
-</p>
-$$
-b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} < 0,
-$$
-
-<p>for the class of observations \( y_i=-1 \), 
-then \( \boldsymbol{x}_i \) lies on the other side. 
-</p>
-
-<p>Equivalently, for the two classes of observations we have </p>
-$$
-y_i\left(b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip}\right) > 0. 
-$$
-
-<p>When we try to separate hyperplanes, if it exists, we can use it to construct a natural classifier: a test observation is assigned a given class depending on which side of the hyperplane it is located.</p>
-
-<!-- !split  -->
-<h2 id="the-two-dimensional-case">The two-dimensional case </h2>
-
-<p>Let us try to develop our intuition about SVMs by limiting ourselves to a two-dimensional
-plane.  To separate the two classes of data points, there are many
-possible lines (hyperplanes if you prefer a more strict naming)  
-that could be chosen. Our objective is to find a
-plane that has the maximum margin, i.e the maximum distance between
-data points of both classes. Maximizing the margin distance provides
-some reinforcement so that future data points can be classified with
-more confidence.
-</p>
-
-<p>What a linear classifier attempts to accomplish is to split the
-feature space into two half spaces by placing a hyperplane between the
-data points.  This hyperplane will be our decision boundary.  All
-points on one side of the plane will belong to class one and all points
-on the other side of the plane will belong to the second class two.
-</p>
-
-<p>Unfortunately there are many ways in which we can place a hyperplane
-to divide the data.  Below is an example of two candidate hyperplanes
-for our data sample.
+<p>The leaf nodes
+contain the predictions we will make for new query instances presented
+to our trained model. This is possible since the model has 
+learned the underlying structure of the training data and hence can,
+given some assumptions, make predictions about the target feature value
+(class) of unseen query instances.
 </p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="getting-into-the-details">Getting into the details </h2>
+<h2 id="a-sketch-of-a-tree-regression-problem">A Sketch of a Tree, Regression problem  </h2>
 
-<p>Let us define the function</p>
-$$
-f(x) = \boldsymbol{w}^T\boldsymbol{x}+b = 0,
-$$
-
-<p>as the function that determines the line \( L \) that separates two classes (our two features), see the figure here. </p>
-
-<p>Any point defined by \( \boldsymbol{x}_i \) and \( \boldsymbol{x}_2 \) on the line \( L \) will satisfy \( \boldsymbol{w}^T(\boldsymbol{x}_1-\boldsymbol{x}_2)=0 \). </p>
-
-<p>The signed distance \( \delta \) from any point defined by a vector \( \boldsymbol{x} \) and a point \( \boldsymbol{x}_0 \) on the line \( L \) is then</p>
-$$
-\delta = \frac{1}{\vert\vert \boldsymbol{w}\vert\vert}(\boldsymbol{w}^T\boldsymbol{x}+b).
-$$
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
 
+<!-- FIGURE: [DataFiles/Regsimpletree.png, width=600 frac=0.8] -->
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="first-attempt-at-a-minimization-approach">First attempt at a minimization approach </h2>
-
-<p>How do we find the parameter \( b \) and the vector \( \boldsymbol{w} \)? What we could
-do is to define a cost function which now contains the set of all
-misclassified points \( M \) and attempt to minimize this function
-</p>
-
-$$
-C(\boldsymbol{w},b) = -\sum_{i\in M} y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b).
-$$
-
-<p>We could now for example define all values \( y_i =1 \) as misclassified in case we have \( \boldsymbol{w}^T\boldsymbol{x}_i+b < 0 \) and the opposite if we have \( y_i=-1 \). Taking the derivatives gives us</p>
-$$
-\frac{\partial C}{\partial b} = -\sum_{i\in M} y_i,
-$$
-
-<p>and </p>
-$$
-\frac{\partial C}{\partial \boldsymbol{w}} = -\sum_{i\in M} y_ix_i.
-$$
+<h2 id="a-sketch-of-a-tree-classification-problem">A Sketch of a Tree, Classification  problem  </h2>
 
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+<!-- FIGURE: [DataFiles/Classimpletree.png, width=600 frac=0.8] -->
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="solving-the-equations">Solving the equations </h2>
+<h2 id="a-typical-decision-tree-with-its-pertinent-jargon-classification-problem">A typical Decision Tree with its pertinent Jargon, Classification Problem </h2>
 
-<p>We can now use the Newton-Raphson method or different variants of the gradient descent family (from plain gradient descent to various stochastic gradient descent approaches) to solve the equations</p>
-$$
-b \leftarrow b +\eta \frac{\partial C}{\partial b},
-$$
+<br/><br/>
+<center>
+<p><img src="DataFiles/cancer.png" width="600" align="bottom"></p>
+</center>
+<br/><br/>
 
-<p>and</p>
-$$
-\boldsymbol{w} \leftarrow \boldsymbol{w} +\eta \frac{\partial C}{\partial \boldsymbol{w}},
-$$
-
-<p>where \( \eta \) is our by now well-known learning rate. </p>
+<p>This tree was produced using the Wisconsin cancer data (discussed here as well, see code examples below) using <b>Scikit-Learn</b>'s decision tree classifier. Here we have used the so-called <b>gini</b> index (see below) to split the various branches.</p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="code-example">Code Example </h2>
+<h2 id="general-features">General Features </h2>
 
-<p>The equations we discussed above can be coded rather easily (the
-framework is similar to what we developed for logistic
-regression). We are going to set up a simple case with two classes only and we want to find a line which separates them the best possible way.
-</p>
+<p>The overarching approach to decision trees is a top-down approach.</p>
 
-<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #eeeedd">
-  <pre style="line-height: 125%;">
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
-
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="problems-with-the-simpler-approach">Problems with the Simpler Approach </h2>
-
-<p>There are however problems with this approach, although it looks
-pretty straightforward to implement. When running the above code, we see that we can easily end up with many diffeent lines which separate the two classes.
-</p>
-
-<p>For small
-gaps between the entries, we may also end up needing many iterations
-before the solutions converge and if the data cannot be separated
-properly into two distinct classes, we may not experience a converge
-at all.
+<ul>
+<li> A leaf provides the classification of a given instance.</li>
+<li> A node specifies a test of some attribute of the instance.</li>
+<li> A branch corresponds to a possible values of an attribute.</li>
+<li> An instance is classified by starting at the root node of the tree, testing the attribute specified by this node, then moving down the tree branch corresponding to the value of the attribute in the given example.</li>
+</ul>
+<p>This process is then repeated for the subtree rooted at the new
+node.
 </p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="a-better-approach">A better approach </h2>
+<h2 id="how-do-we-set-it-up">How do we set it up? </h2>
 
-<p>A better approach is rather to try to define a large margin between
-the two classes (if they are well separated from the beginning).
+<p>In simplified terms, the process of training a decision tree and
+predicting the target features of query instances is as follows:
 </p>
 
-<p>Thus, we wish to find a margin \( M \) with \( \boldsymbol{w} \) normalized to
-\( \vert\vert \boldsymbol{w}\vert\vert =1 \) subject to the condition
-</p>
-
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, p. 
-$$
-
-<p>All points are thus at a signed distance from the decision boundary defined by the line \( L \). The parameters \( b \) and \( w_1 \) and \( w_2 \) define this line. </p>
-
-<p>We seek thus the largest value \( M \) defined by</p>
-$$
-\frac{1}{\vert \vert \boldsymbol{w}\vert\vert}y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, n, 
-$$
-
-<p>or just </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M\vert \vert \boldsymbol{w}\vert\vert \hspace{0.1cm}\forall i. 
-$$
-
-<p>If we scale the equation so that \( \vert \vert \boldsymbol{w}\vert\vert = 1/M \), we have to find the minimum of 
-\( \boldsymbol{w}^T\boldsymbol{w}=\vert \vert \boldsymbol{w}\vert\vert \) (the norm) subject to the condition
-</p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq 1 \hspace{0.1cm}\forall i.
-$$
-
-<p>We have thus defined our margin as the invers of the norm of
-\( \boldsymbol{w} \). We want to minimize the norm in order to have a as large as
-possible margin \( M \). Before we proceed, we need to remind ourselves
-about Lagrangian multipliers.
-</p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="a-quick-reminder-on-lagrangian-multipliers">A quick Reminder on Lagrangian Multipliers </h2>
-
-<p>Consider a function of three independent variables \( f(x,y,z) \) . For the function \( f \) to be an
-extreme we have
-</p>
-$$
-df=0.
-$$
-
-<p>A necessary and sufficient condition is</p>
-$$
-\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
-$$
-
-<p>due to</p>
-$$
-df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz.
-$$
-
-<p>In many problems the variables \( x,y,z \) are often subject to constraints (such as those above for the margin)
-so that they are no longer all independent. It is possible at least in principle to use each 
-constraint to eliminate one variable
-and to proceed with a new and smaller set of independent varables.
-</p>
-
-<p>The use of so-called Lagrangian  multipliers is an alternative technique  when the elimination
-of variables is incovenient or undesirable.  Assume that we have an equation of constraint on 
-the variables \( x,y,z \)
-</p>
-$$
-\phi(x,y,z) = 0,
-$$
-
-<p> resulting in</p>
-$$
-d\phi = \frac{\partial \phi}{\partial x}dx+\frac{\partial \phi}{\partial y}dy+\frac{\partial \phi}{\partial z}dz =0.
-$$
-
-<p>Now we cannot set anymore</p>
-$$
-\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
-$$
-
-<p>if \( df=0 \) is wanted
-because there are now only two independent variables!  Assume \( x \) and \( y \) are the independent 
-variables.
-Then \( dz \) is no longer arbitrary.
-</p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="adding-the-multiplier">Adding the Multiplier </h2>
-
-<p>However, we can add to</p>
-$$
-df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz,
-$$
-
-<p>a multiplum of \( d\phi \), viz. \( \lambda d\phi \), resulting  in</p>
-$$
-df+\lambda d\phi = (\frac{\partial f}{\partial z}+\lambda
-\frac{\partial \phi}{\partial x})dx+(\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y})dy+
-(\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z})dz =0.
-$$
-
-<p>Our multiplier is chosen so that</p>
-$$
-\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z} =0.
-$$
-
-<p>We need to remember that we took \( dx \) and \( dy \) to be arbitrary and thus we must have</p>
-$$
-\frac{\partial f}{\partial x}+\lambda\frac{\partial \phi}{\partial x} =0,
-$$
-
-<p>and</p>
-$$
-\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y} =0.
-$$
-
-<p>When all these equations are satisfied, \( df=0 \).  We have four unknowns, \( x,y,z \) and
-\( \lambda \). Actually we want only \( x,y,z \), \( \lambda \) needs not to be determined, 
-it is therefore often called
-Lagrange's undetermined multiplier.
-If we have a set of constraints \( \phi_k \) we have the equations
-</p>
-$$
-\frac{\partial f}{\partial x_i}+\sum_k\lambda_k\frac{\partial \phi_k}{\partial x_i} =0.
-$$
-
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="setting-up-the-problem">Setting up the Problem </h2>
-<p>In order to solve the above problem, we define the following Lagrangian function to be minimized </p>
-$$
-{\cal L}(\lambda,b,\boldsymbol{w})=\frac{1}{2}\boldsymbol{w}^T\boldsymbol{w}-\sum_{i=1}^n\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)-1\right],
-$$
-
-<p>where \( \lambda_i \) is a so-called Lagrange multiplier subject to the condition \( \lambda_i \geq 0 \).</p>
-
-<p>Taking the derivatives  with respect to \( b \) and \( \boldsymbol{w} \) we obtain </p>
-$$
-\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
-$$
-
-<p>and </p>
-$$
-\frac{\partial {\cal L}}{\partial \boldsymbol{w}} = 0 = \boldsymbol{w}-\sum_{i} \lambda_iy_i\boldsymbol{x}_i.
-$$
-
-<p>Inserting these constraints into the equation for \( {\cal L} \) we obtain</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-
-<p>subject to the constraints \( \lambda_i\geq 0 \) and \( \sum_i\lambda_iy_i=0 \). 
-We must in addition satisfy the <a href="https://en.wikipedia.org/wiki/Karush%E2%80%93Kuhn%E2%80%93Tucker_conditions" target="_blank">Karush-Kuhn-Tucker</a> (KKT) condition
-</p>
-$$
-\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -1\right] \hspace{0.1cm}\forall i.
-$$
-
 <ol>
-<li> If \( \lambda_i > 0 \), then \( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1 \) and we say that \( x_i \) is on the boundary.</li>
-<li> If \( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)> 1 \), we say \( x_i \) is not on the boundary and we set \( \lambda_i=0 \).</li> 
+<li> Present a dataset containing of a number of training instances characterized by a number of descriptive features and a target feature</li>
+<li> Train the decision tree model by continuously splitting the target feature along the values of the descriptive features using a measure of information gain during the training process</li>
+<li> Grow the tree until we accomplish a stopping criteria create leaf nodes which represent the <em>predictions</em> we want to make for new query instances</li>
+<li> Show query instances to the tree and run down the tree until we arrive at leaf nodes</li>
 </ol>
-<p>When \( \lambda_i > 0 \), the vectors \( \boldsymbol{x}_i \) are called support vectors. They are the vectors closest to the line (or hyperplane) and define the margin \( M \). </p>
+<p>Then we are essentially done!</p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="the-problem-to-solve">The problem to solve </h2>
-
-<p>We can rewrite </p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-
-<p>and its constraints in terms of a matrix-vector problem where we minimize w.r.t. \( \lambda \) the following problem</p>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1\boldsymbol{x}_1^T\boldsymbol{x}_1 & y_1y_2\boldsymbol{x}_1^T\boldsymbol{x}_2 & \dots & \dots & y_1y_n\boldsymbol{x}_1^T\boldsymbol{x}_n \\
-y_2y_1\boldsymbol{x}_2^T\boldsymbol{x}_1 & y_2y_2\boldsymbol{x}_2^T\boldsymbol{x}_2 & \dots & \dots & y_1y_n\boldsymbol{x}_2^T\boldsymbol{x}_n \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1\boldsymbol{x}_n^T\boldsymbol{x}_1 & y_ny_2\boldsymbol{x}_n^T\boldsymbol{x}_2 & \dots & \dots & y_ny_n\boldsymbol{x}_n^T\boldsymbol{x}_n \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{1}\boldsymbol{\lambda}, 
-$$
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
-</p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="the-last-steps">The last steps </h2>
-
-<p>Solving the above problem, yields the values of \( \lambda_i \).
-To find the coefficients of your hyperplane we need simply to compute 
-</p>
-$$
-\boldsymbol{w}=\sum_{i} \lambda_iy_i\boldsymbol{x}_i.
-$$
-
-<p>With our vector \( \boldsymbol{w} \) we can in turn find the value of the intercept \( b \) (here in two dimensions) via </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1,
-$$
-
-<p>resulting in</p>
-$$
-b = \frac{1}{y_i}-\boldsymbol{w}^T\boldsymbol{x}_i,
-$$
-
-<p>or if we write it out in terms of the support vectors only, with \( N_s \) being their number,  we have</p>
-$$
-b = \frac{1}{N_s}\sum_{j\in N_s}\left(y_j-\sum_{i=1}^n\lambda_iy_i\boldsymbol{x}_i^T\boldsymbol{x}_j\right).
-$$
-
-<p>With our hyperplane coefficients we can use our classifier to assign any observation by simply using </p>
-$$
-y_i = \mathrm{sign}(\boldsymbol{w}^T\boldsymbol{x}_i+b).
-$$
-
-<p>Below we discuss how to find the optimal values of \( \lambda_i \). Before we proceed however, we discuss now the so-called soft classifier. </p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="a-soft-classifier">A soft classifier </h2>
-
-<p>Till now, the margin is strictly defined by the support vectors. This defines what is called a hard classifier, that is the margins are well defined.</p>
-
-<p>Suppose now that classes overlap in feature space, as shown in the
-figure here. One way to deal with this problem before we define the
-so-called <b>kernel approach</b>, is to allow a kind of slack in the sense
-that we allow some points to be on the wrong side of the margin.
-</p>
-
-<p>We introduce thus the so-called <b>slack</b> variables \( \boldsymbol{\xi} =[\xi_1,x_2,\dots,x_n] \) and 
-modify our previous equation
-</p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1,
-$$
-
-<p>to </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1-\xi_i,
-$$
-
-<p>with the requirement \( \xi_i\geq 0 \). The total violation is now \( \sum_i\xi \). 
-The value \( \xi_i \) in the constraint the last constraint corresponds to the  amount by which the prediction
-\( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1 \) is on the wrong side of its margin. Hence by bounding the sum \( \sum_i \xi_i \),
-we bound the total amount by which predictions fall on the wrong side of their margins.
-</p>
-
-<p>Misclassifications occur when \( \xi_i > 1 \). Thus bounding the total sum by some value \( C \) bounds in turn the total number of
-misclassifications.
-</p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="soft-optmization-problem">Soft optmization problem </h2>
-
-<p>This has in turn the consequences that we change our optmization problem to finding the minimum of </p>
-$$
-{\cal L}=\frac{1}{2}\boldsymbol{w}^T\boldsymbol{w}-\sum_{i=1}^n\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)-(1-\xi_)\right]+C\sum_{i=1}^n\xi_i-\sum_{i=1}^n\gamma_i\xi_i,
-$$
-
-<p>subject to </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1-\xi_i \hspace{0.1cm}\forall i,
-$$
-
-<p>with the requirement \( \xi_i\geq 0 \).</p>
-
-<p>Taking the derivatives  with respect to \( b \) and \( \boldsymbol{w} \) we obtain </p>
-$$
-\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
-$$
-
-<p>and </p>
-$$
-\frac{\partial {\cal L}}{\partial \boldsymbol{w}} = 0 = \boldsymbol{w}-\sum_{i} \lambda_iy_i\boldsymbol{x}_i, 
-$$
-
-<p>and</p>
-$$
-\lambda_i = C-\gamma_i \hspace{0.1cm}\forall i.
-$$
-
-<p>Inserting these constraints into the equation for \( {\cal L} \) we obtain the same equation as before</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-
-<p>but now subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \) and \( 0\leq\lambda_i \leq C \). 
-We must in addition satisfy the Karush-Kuhn-Tucker condition which now reads
-</p>
-$$
-\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -(1-\xi_)\right]=0 \hspace{0.1cm}\forall i,
-$$
-
-$$
-\gamma_i\xi_i = 0,
-$$
-
-<p>and </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -(1-\xi_) \geq 0 \hspace{0.1cm}\forall i.
-$$
-
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="kernels-and-non-linearity">Kernels and non-linearity </h2>
-
-<p>The cases we have studied till now, were all characterized by two classes
-with a close to linear separability. The classifiers we have described
-so far find linear boundaries in our input feature space. It is
-possible to make our procedure more flexible by exploring the feature
-space using other basis expansions such as higher-order polynomials,
-wavelets, splines etc.
-</p>
-
-<p>If our feature space is not easy to separate, as shown in the figure
-here, we can achieve a better separation by introducing more complex
-basis functions. The ideal would be, as shown in the next figure, to, via a specific transformation to 
-obtain a separation between the classes which is almost linear. 
-</p>
-
-<p>The change of basis, from \( x\rightarrow z=\phi(x) \) leads to the same type of equations to be solved, except that
-we need to introduce for example a polynomial transformation to a two-dimensional training set.
-</p>
-
+<h2 id="decision-trees-and-regression">Decision trees and Regression  </h2>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
 <div class="cell border-box-sizing code_cell rendered">
@@ -868,51 +424,444 @@ we need to introduce for example a polynomial transformation to a two-dimensiona
       <div class="input_area">
         <div class="highlight" style="background: #eeeedd">
   <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> PolynomialFeatures
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> LinearRegression
+
+steps=<span style="color: #B452CD">250</span>
+
+distance=<span style="color: #B452CD">0</span>
+x=<span style="color: #B452CD">0</span>
+distance_list=[]
+steps_list=[]
+<span style="color: #8B008B; font-weight: bold">while</span> x&lt;steps:
+    distance+=np.random.randint(-<span style="color: #B452CD">1</span>,<span style="color: #B452CD">2</span>)
+    distance_list.append(distance)
+    x+=<span style="color: #B452CD">1</span>
+    steps_list.append(x)
+plt.plot(steps_list,distance_list, color=<span style="color: #CD5555">&#39;green&#39;</span>, label=<span style="color: #CD5555">&quot;Random Walk Data&quot;</span>)
+
+steps_list=np.asarray(steps_list)
+distance_list=np.asarray(distance_list)
+
+X=steps_list[:,np.newaxis]
+
+<span style="color: #228B22">#Polynomial fits</span>
+
+<span style="color: #228B22">#Degree 2</span>
+poly_features=PolynomialFeatures(degree=<span style="color: #B452CD">2</span>, include_bias=<span style="color: #8B008B; font-weight: bold">False</span>)
+X_poly=poly_features.fit_transform(X)
+
+lin_reg=LinearRegression()
+poly_fit=lin_reg.fit(X_poly,distance_list)
+b=lin_reg.coef_
+c=lin_reg.intercept_
+<span style="color: #658b00">print</span> (<span style="color: #CD5555">&quot;2nd degree coefficients:&quot;</span>)
+<span style="color: #658b00">print</span> (<span style="color: #CD5555">&quot;zero power: &quot;</span>,c)
+<span style="color: #658b00">print</span> (<span style="color: #CD5555">&quot;first power: &quot;</span>, b[<span style="color: #B452CD">0</span>])
+<span style="color: #658b00">print</span> (<span style="color: #CD5555">&quot;second power: &quot;</span>,b[<span style="color: #B452CD">1</span>])
+
+z = np.arange(<span style="color: #B452CD">0</span>, steps, <span style="color: #B452CD">.01</span>)
+z_mod=b[<span style="color: #B452CD">1</span>]*z**<span style="color: #B452CD">2</span>+b[<span style="color: #B452CD">0</span>]*z+c
+
+fit_mod=b[<span style="color: #B452CD">1</span>]*X**<span style="color: #B452CD">2</span>+b[<span style="color: #B452CD">0</span>]*X+c
+plt.plot(z, z_mod, color=<span style="color: #CD5555">&#39;r&#39;</span>, label=<span style="color: #CD5555">&quot;2nd Degree Fit&quot;</span>)
+plt.title(<span style="color: #CD5555">&quot;Polynomial Regression&quot;</span>)
+
+plt.xlabel(<span style="color: #CD5555">&quot;Steps&quot;</span>)
+plt.ylabel(<span style="color: #CD5555">&quot;Distance&quot;</span>)
+
+<span style="color: #228B22">#Degree 10</span>
+poly_features10=PolynomialFeatures(degree=<span style="color: #B452CD">10</span>, include_bias=<span style="color: #8B008B; font-weight: bold">False</span>)
+X_poly10=poly_features10.fit_transform(X)
+
+poly_fit10=lin_reg.fit(X_poly10,distance_list)
+
+y_plot=poly_fit10.predict(X_poly10)
+plt.plot(X, y_plot, color=<span style="color: #CD5555">&#39;black&#39;</span>, label=<span style="color: #CD5555">&quot;10th Degree Fit&quot;</span>)
+
+plt.legend()
+plt.show()
+
+
+<span style="color: #228B22">#Decision Tree Regression</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeRegressor
+regr_1=DecisionTreeRegressor(max_depth=<span style="color: #B452CD">2</span>)
+regr_2=DecisionTreeRegressor(max_depth=<span style="color: #B452CD">5</span>)
+regr_3=DecisionTreeRegressor(max_depth=<span style="color: #B452CD">7</span>)
+regr_1.fit(X, distance_list)
+regr_2.fit(X, distance_list)
+regr_3.fit(X, distance_list)
+
+X_test = np.arange(<span style="color: #B452CD">0.0</span>, steps, <span style="color: #B452CD">0.01</span>)[:, np.newaxis]
+y_1 = regr_1.predict(X_test)
+y_2 = regr_2.predict(X_test)
+y_3=regr_3.predict(X_test)
+
+<span style="color: #228B22"># Plot the results</span>
+plt.figure()
+plt.scatter(X, distance_list, s=<span style="color: #B452CD">2.5</span>, c=<span style="color: #CD5555">&quot;black&quot;</span>, label=<span style="color: #CD5555">&quot;data&quot;</span>)
+plt.plot(X_test, y_1, color=<span style="color: #CD5555">&quot;red&quot;</span>,
+         label=<span style="color: #CD5555">&quot;max_depth=2&quot;</span>, linewidth=<span style="color: #B452CD">2</span>)
+plt.plot(X_test, y_2, color=<span style="color: #CD5555">&quot;green&quot;</span>, label=<span style="color: #CD5555">&quot;max_depth=5&quot;</span>, linewidth=<span style="color: #B452CD">2</span>)
+plt.plot(X_test, y_3, color=<span style="color: #CD5555">&quot;m&quot;</span>, label=<span style="color: #CD5555">&quot;max_depth=7&quot;</span>, linewidth=<span style="color: #B452CD">2</span>)
+
+plt.xlabel(<span style="color: #CD5555">&quot;Data&quot;</span>)
+plt.ylabel(<span style="color: #CD5555">&quot;Darget&quot;</span>)
+plt.title(<span style="color: #CD5555">&quot;Decision Tree Regression&quot;</span>)
+plt.legend()
+plt.show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="building-a-tree-regression">Building a tree, regression </h2>
+
+<p>There are mainly two steps</p>
+<ol>
+<li> We split the predictor space (the set of possible values \( x_1,x_2,\dots, x_p \)) into \( J \) distinct and non-non-overlapping regions, \( R_1,R_2,\dots,R_J \).</li>  
+<li> For every observation that falls into the region \( R_j \) , we make the same prediction, which is simply the mean of the response values for the training observations in \( R_j \).</li>
+</ol>
+<p>How do we construct the regions \( R_1,\dots,R_J \)?  In theory, the
+regions could have any shape. However, we choose to divide the
+predictor space into high-dimensional rectangles, or boxes, for
+simplicity and for ease of interpretation of the resulting predictive
+model. The goal is to find boxes \( R_1,\dots,R_J \) that minimize the
+MSE, given by
+</p>
+
+$$
+\sum_{j=1}^J\sum_{i\in R_j}(y_i-\overline{y}_{R_j})^2,
+$$
+
+<p>where \( \overline{y}_{R_j} \)  is the mean response for the training observations 
+within box \( j \). 
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="a-top-down-approach-recursive-binary-splitting">A top-down approach, recursive binary splitting </h2>
+
+<p>Unfortunately, it is computationally infeasible to consider every
+possible partition of the feature space into \( J \) boxes.  The common
+strategy is to take a top-down approach
+</p>
+
+<p>The approach is top-down because it begins at the top of the tree (all
+observations belong to a single region) and then successively splits
+the predictor space; each split is indicated via two new branches
+further down on the tree. It is greedy because at each step of the
+tree-building process, the best split is made at that particular step,
+rather than looking ahead and picking a split that will lead to a
+better tree in some future step.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="making-a-tree">Making a tree </h2>
+
+<p>In order to implement the recursive binary splitting we start by selecting
+the predictor \( x_j \) and a cutpoint \( s \) that splits the predictor space into two regions \( R_1 \) and \( R_2 \)
+</p>
+$$
+\left\{X\vert x_j < s\right\},
+$$
+
+<p>and</p>
+$$
+\left\{X\vert x_j \geq s\right\},
+$$
+
+<p>so that we obtain the lowest MSE, that is</p>
+$$
+\sum_{i:x_i\in R_j}(y_i-\overline{y}_{R_1})^2+\sum_{i:x_i\in R_2}(y_i-\overline{y}_{R_2})^2,
+$$
+
+<p>which we want to minimize by considering all predictors
+\( x_1,x_2,\dots,x_p \).  We consider also all possible values of \( s \) for
+each predictor. These values could be determined by randomly assigned
+numbers or by starting at the midpoint and then proceed till we find
+an optimal value.
+</p>
+
+<p>For any \( j \) and \( s \), we define the pair of half-planes where
+\( \overline{y}_{R_1} \) is the mean response for the training
+observations in \( R_1(j,s) \), and \( \overline{y}_{R_2} \) is the mean
+response for the training observations in \( R_2(j,s) \).
+</p>
+
+<p>Finding the values of \( j \) and \( s \) that minimize the above equation can be
+done quite quickly, especially when the number of features \( p \) is not
+too large.
+</p>
+
+<p>Next, we repeat the process, looking
+for the best predictor and best cutpoint in order to split the data
+further so as to minimize the MSE within each of the resulting
+regions. However, this time, instead of splitting the entire predictor
+space, we split one of the two previously identified regions. We now
+have three regions. Again, we look to split one of these three regions
+further, so as to minimize the MSE. The process continues until a
+stopping criterion is reached; for instance, we may continue until no
+region contains more than five observations.
+</p>
+
+<!-- !split  -->
+<h2 id="pruning-the-tree">Pruning the tree </h2>
+
+<p>The above procedure is rather straightforward, but leads often to
+overfitting and unnecessarily large and complicated trees. The basic
+idea is to grow a large tree \( T_0 \) and then prune it back in order to
+obtain a subtree. A smaller tree with fewer splits (fewer regions) can
+lead to smaller variance and better interpretation at the cost of a
+little more bias.
+</p>
+
+<p>The so-called Cost complexity pruning algorithm gives us a
+way to do just this. Rather than considering every possible subtree,
+we consider a sequence of trees indexed by a nonnegative tuning
+parameter \( \alpha \).
+</p>
+
+<p>Read more at the following <a href="https://scikit-learn.org/stable/auto_examples/tree/plot_cost_complexity_pruning.html#sphx-glr-auto-examples-tree-plot-cost-complexity-pruning-py" target="_blank">Scikit-Learn link on pruning</a>.</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="cost-complexity-pruning">Cost complexity pruning </h2>
+
+<p>For each value of \( \alpha \)  there corresponds a subtree \( T \in T_0 \) such that</p>
+$$
+\sum_{m=1}^{\overline{T}}\sum_{i:x_i\in R_m}(y_i-\overline{y}_{R_m})^2+\alpha\overline{T},
+$$
+
+<p>is as small as possible. Here \( \overline{T} \) is 
+the number of terminal nodes of the tree \( T \) , \( R_m \) is the
+rectangle (i.e. the subset of predictor space)  corresponding to the \( m \)-th terminal node.
+</p>
+
+<p>The tuning parameter \( \alpha \) controls a trade-off between the subtree&#8217;s
+complexity and its fit to the training data. When \( \alpha = 0 \), then the
+subtree \( T \) will simply equal \( T_0 \), 
+because then the above equation just measures the
+training error. 
+However, as \( \alpha \) increases, there is a price to pay for
+having a tree with many terminal nodes. The above equation will
+tend to be minimized for a smaller subtree. 
+</p>
+
+<p>It turns out that as we increase \( \alpha \) from zero
+branches get pruned from the tree in a nested and predictable fashion,
+so obtaining the whole sequence of subtrees as a function of \( \alpha \) is
+easy. We can select a value of \( \alpha \) using a validation set or using
+cross-validation. We then return to the full data set and obtain the
+subtree corresponding to \( \alpha \). 
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="schematic-regression-procedure">Schematic Regression Procedure </h2>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Building a Regression Tree</b>
+<p>
+
+<ol>
+<li> Use recursive binary splitting to grow a large tree on the training data, stopping only when each terminal node has fewer than some minimum number of observations.</li>
+<li> Apply cost complexity pruning to the large tree in order to obtain a sequence of best subtrees, as a function of \( \alpha \).</li>
+<li> Use for example \( K \)-fold cross-validation to choose \( \alpha \). Divide the training observations into \( K \) folds. For each \( k=1,2,\dots,K \) we:</li> 
+<ul>
+  <li> repeat steps 1 and 2 on all but the \( k \)-th fold of the training data.</li> 
+  <li> Then we valuate the mean squared prediction error on the data in the left-out \( k \)-th fold, as a function of \( \alpha \).</li>
+  <li> Finally  we average the results for each value of \( \alpha \), and pick \( \alpha \) to minimize the average error.</li>
+</ul>
+<li> Return the subtree from Step 2 that corresponds to the chosen value of \( \alpha \).</li> 
+</ol>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="a-classification-tree">A Classification Tree </h2>
+
+<p>A classification tree is very similar to a regression tree, except
+that it is used to predict a qualitative response rather than a
+quantitative one. Recall that for a regression tree, the predicted
+response for an observation is given by the mean response of the
+training observations that belong to the same terminal node. In
+contrast, for a classification tree, we predict that each observation
+belongs to the most commonly occurring class of training observations
+in the region to which it belongs. In interpreting the results of a
+classification tree, we are often interested not only in the class
+prediction corresponding to a particular terminal node region, but
+also in the class proportions among the training observations that
+fall into that region.  
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="growing-a-classification-tree">Growing a classification tree </h2>
+
+<p>The task of growing a
+classification tree is quite similar to the task of growing a
+regression tree. Just as in the regression setting, we use recursive
+binary splitting to grow a classification tree. However, in the
+classification setting, the MSE cannot be used as a criterion for making
+the binary splits.  A natural alternative to MSE is the <b>classification
+error rate</b>. Since we plan to assign an observation in a given region
+to the most commonly occurring error rate class of training
+observations in that region, the classification error rate is simply
+the fraction of the training observations in that region that do not
+belong to the most common class. 
+</p>
+
+<p>When building a classification tree, either the Gini index or the
+entropy are typically used to evaluate the quality of a particular
+split, since these two approaches are more sensitive to node purity
+than is the classification error rate. 
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="classification-tree-how-to-split-nodes">Classification tree, how to split nodes </h2>
+
+<p>If our targets are the outcome of a classification process that takes
+for example \( k=1,2,\dots,K \) values, the only thing we need to think of
+is to set up the splitting criteria for each node.
+</p>
+
+<p>We define a PDF \( p_{mk} \) that represents the number of observations of
+a class \( k \) in a region \( R_m \) with \( N_m \) observations. We represent
+this likelihood function in terms of the proportion \( I(y_i=k) \) of
+observations of this class in the region \( R_m \) as
+</p>
+
+$$
+p_{mk} = \frac{1}{N_m}\sum_{x_i\in R_m}I(y_i=k).
+$$
+
+<p>We let \( p_{mk} \) represent the majority class of observations in region
+\( m \). The three most common ways of splitting a node are given by
+</p>
+
+<ul>
+<li> Misclassification error</li> 
+</ul>
+$$
+p_{mk} = \frac{1}{N_m}\sum_{x_i\in R_m}I(y_i\ne k) = 1-p_{mk}.
+$$
+
+<ul>
+<li> Gini index \( g \)</li>
+</ul>
+$$
+g = \sum_{k=1}^K p_{mk}(1-p_{mk}).
+$$
+
+<ul>
+<li> Information entropy or just entropy \( s \)</li>
+</ul>
+$$
+s = -\sum_{k=1}^K p_{mk}\log{p_{mk}}.
+$$
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="visualizing-the-tree-classification">Visualizing the Tree, Classification </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> load_breast_cancer
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> confusion_matrix
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_graphviz
+
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> Image 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+
+
+cancer = load_breast_cancer()
+X = pd.DataFrame(cancer.data, columns=cancer.feature_names)
+<span style="color: #658b00">print</span>(X)
+y = pd.Categorical.from_codes(cancer.target, cancer.target_names)
+y = pd.get_dummies(y)
+<span style="color: #658b00">print</span>(y)
+X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=<span style="color: #B452CD">1</span>)
+tree_clf = DecisionTreeClassifier(max_depth=<span style="color: #B452CD">5</span>)
+tree_clf.fit(X_train, y_train)
+
+export_graphviz(
+    tree_clf,
+    out_file=<span style="color: #CD5555">&quot;DataFiles/cancer.dot&quot;</span>,
+    feature_names=cancer.feature_names,
+    class_names=cancer.target_names,
+    rounded=<span style="color: #8B008B; font-weight: bold">True</span>,
+    filled=<span style="color: #8B008B; font-weight: bold">True</span>
+)
+cmd = <span style="color: #CD5555">&#39;dot -Tpng DataFiles/cancer.dot -o DataFiles/cancer.png&#39;</span>
+os.system(cmd)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="visualizing-the-tree-the-moons">Visualizing the Tree, The Moons  </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #228B22"># Common imports</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span>  train_test_split 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> make_moons
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_graphviz
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
 
 np.random.seed(<span style="color: #B452CD">42</span>)
+X, y = make_moons(n_samples=<span style="color: #B452CD">100</span>, noise=<span style="color: #B452CD">0.25</span>, random_state=<span style="color: #B452CD">53</span>)
+X_train, X_test, y_train, y_test = train_test_split(X,y,random_state=<span style="color: #B452CD">0</span>)
+tree_clf = DecisionTreeClassifier(max_depth=<span style="color: #B452CD">5</span>)
+tree_clf.fit(X_train, y_train)
 
-<span style="color: #228B22"># To plot pretty figures</span>
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib</span>
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
-plt.rcParams[<span style="color: #CD5555">&#39;axes.labelsize&#39;</span>] = <span style="color: #B452CD">14</span>
-plt.rcParams[<span style="color: #CD5555">&#39;xtick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
-plt.rcParams[<span style="color: #CD5555">&#39;ytick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
-
-
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets
-
-
-
-X1D = np.linspace(-<span style="color: #B452CD">4</span>, <span style="color: #B452CD">4</span>, <span style="color: #B452CD">9</span>).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
-X2D = np.c_[X1D, X1D**<span style="color: #B452CD">2</span>]
-y = np.array([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>])
-
-plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
-
-plt.subplot(<span style="color: #B452CD">121</span>)
-plt.grid(<span style="color: #8B008B; font-weight: bold">True</span>, which=<span style="color: #CD5555">&#39;both&#39;</span>)
-plt.axhline(y=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.plot(X1D[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">0</span>], np.zeros(<span style="color: #B452CD">4</span>), <span style="color: #CD5555">&quot;bs&quot;</span>)
-plt.plot(X1D[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">1</span>], np.zeros(<span style="color: #B452CD">5</span>), <span style="color: #CD5555">&quot;g^&quot;</span>)
-plt.gca().get_yaxis().set_ticks([])
-plt.xlabel(<span style="color: #CD5555">r&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.axis([-<span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">4.5</span>, -<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">0.2</span>])
-
-plt.subplot(<span style="color: #B452CD">122</span>)
-plt.grid(<span style="color: #8B008B; font-weight: bold">True</span>, which=<span style="color: #CD5555">&#39;both&#39;</span>)
-plt.axhline(y=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.axvline(x=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.plot(X2D[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">0</span>], X2D[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">0</span>], <span style="color: #CD5555">&quot;bs&quot;</span>)
-plt.plot(X2D[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">1</span>], X2D[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;g^&quot;</span>)
-plt.xlabel(<span style="color: #CD5555">r&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.ylabel(<span style="color: #CD5555">r&quot;$x_2$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>, rotation=<span style="color: #B452CD">0</span>)
-plt.gca().get_yaxis().set_ticks([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">4</span>, <span style="color: #B452CD">8</span>, <span style="color: #B452CD">12</span>, <span style="color: #B452CD">16</span>])
-plt.plot([-<span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">4.5</span>], [<span style="color: #B452CD">6.5</span>, <span style="color: #B452CD">6.5</span>], <span style="color: #CD5555">&quot;r--&quot;</span>, linewidth=<span style="color: #B452CD">3</span>)
-plt.axis([-<span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">4.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">17</span>])
-plt.subplots_adjust(right=<span style="color: #B452CD">1</span>)
-plt.show()
+export_graphviz(
+    tree_clf,
+    out_file=<span style="color: #CD5555">&quot;DataFiles/moons.dot&quot;</span>,
+    rounded=<span style="color: #8B008B; font-weight: bold">True</span>,
+    filled=<span style="color: #8B008B; font-weight: bold">True</span>
+)
+cmd = <span style="color: #CD5555">&#39;dot -Tpng DataFiles/moons.dot -o DataFiles/moons.png&#39;</span>
+os.system(cmd)
 </pre>
 </div>
       </div>
@@ -930,117 +879,231 @@ plt.show()
 
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="the-equations">The equations </h2>
+<h2 id="other-ways-of-visualizing-the-trees">Other ways of visualizing the trees </h2>
 
-<p>Suppose we define a polynomial transformation of degree two only (we continue to live in a plane with \( x_i \) and \( y_i \) as variables)</p>
-$$
-z = \phi(x_i) =\left(x_i^2, y_i^2, \sqrt{2}x_iy_i\right).
-$$
+<p><b>Scikit-Learn</b> has also another way to visualize the trees which is very useful, here with the Iris data.</p>
 
-<p>With our new basis, the equations we solved earlier are basically the same, that is we have now (without the slack option for simplicity)</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{z}_i^T\boldsymbol{z}_j,
-$$
 
-<p>subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \), and for the support vectors</p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{z}_i+b)= 1 \hspace{0.1cm}\forall i,
-$$
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> load_iris
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> tree
+X, y = load_iris(return_X_y=<span style="color: #8B008B; font-weight: bold">True</span>)
+tree_clf = tree.DecisionTreeClassifier()
+tree_clf = tree_clf.fit(X, y)
+<span style="color: #228B22"># and then plot the tree</span>
+tree.plot_tree(tree_clf) 
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
 
-<p>from which we also find \( b \).
-To compute \( \boldsymbol{z}_i^T\boldsymbol{z}_j \) we define the kernel \( K(\boldsymbol{x}_i,\boldsymbol{x}_j) \) as
-</p>
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=\boldsymbol{z}_i^T\boldsymbol{z}_j= \phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j).
-$$
-
-<p>For the above example, the kernel reads</p>
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=[x_i^2, y_i^2, \sqrt{2}x_iy_i]^T\begin{bmatrix} x_j^2 \\ y_j^2 \\ \sqrt{2}x_jy_j \end{bmatrix}=x_i^2x_j^2+2x_ix_jy_iy_j+y_i^2y_j^2.
-$$
-
-<p>We note that this is nothing but the dot product of the two original
-vectors \( (\boldsymbol{x}_i^T\boldsymbol{x}_j)^2 \). Instead of thus computing the
-product in the Lagrangian of \( \boldsymbol{z}_i^T\boldsymbol{z}_j \) we simply compute
-the dot product \( (\boldsymbol{x}_i^T\boldsymbol{x}_j)^2 \).
-</p>
-
-<p>This leads to the so-called
-kernel trick and the result leads to the same as if we went through
-the trouble of performing the transformation
-\( \phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j) \) during the SVM calculations.
-</p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="the-problem-to-solve">The problem to solve </h2>
-<p>Using our definition of the kernel We can rewrite again the Lagrangian</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{z}_j,
-$$
+<h2 id="printing-out-as-text">Printing out as text </h2>
 
-<p>subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \) in terms of a convex optimization problem</p>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1K(\boldsymbol{x}_1,\boldsymbol{x}_1) & y_1y_2K(\boldsymbol{x}_1,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_1,\boldsymbol{x}_n) \\
-y_2y_1K(\boldsymbol{x}_2,\boldsymbol{x}_1) & y_2y_2(\boldsymbol{x}_2,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_2,\boldsymbol{x}_n) \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1K(\boldsymbol{x}_n,\boldsymbol{x}_1) & y_ny_2K(\boldsymbol{x}_n\boldsymbol{x}_2) & \dots & \dots & y_ny_nK(\boldsymbol{x}_n,\boldsymbol{x}_n) \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{1}\boldsymbol{\lambda}, 
-$$
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
-If we add the slack constants this leads to the additional constraint \( 0\leq \lambda_i \leq C \).
+<p>Alternatively, the tree can also be exported in textual format with the function exporttext.
+This method doesn&#8217;t require the installation of external libraries and is more compact:
 </p>
 
-<p>We can rewrite this (see the solutions below) in terms of a convex optimization problem of the type</p>
-$$
-\begin{align*}
-    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\boldsymbol{\lambda}^T\boldsymbol{P}\boldsymbol{\lambda}+\boldsymbol{q}^T\boldsymbol{\lambda},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{\lambda} \preceq \boldsymbol{h} \hspace{0.2cm} \wedge \boldsymbol{A}\boldsymbol{\lambda}=f.
-\end{align*}
-$$
 
-<p>Below we discuss how to solve these equations. Here we note that the matrix \( \boldsymbol{P} \) has matrix elements \( p_{ij}=y_iy_jK(\boldsymbol{x}_i,\boldsymbol{x}_j) \).
-Given a kernel \( K \) and the targets \( y_i \) this matrix is easy to set up. The constraint \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \) leads to \( f=0 \) and \( \boldsymbol{A}=\boldsymbol{y} \). How to set up the matrix \( \boldsymbol{G} \) is discussed later. Here note that the inequalities \( 0\leq \lambda_i \leq C \) can be split up into
-\( 0\leq \lambda_i \) and \( \lambda_i \leq C \). These two inequalities define then the matrix \( \boldsymbol{G} \) and the vector \( \boldsymbol{h} \).
-</p>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> load_iris
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_text
+iris = load_iris()
+decision_tree = DecisionTreeClassifier(random_state=<span style="color: #B452CD">0</span>, max_depth=<span style="color: #B452CD">2</span>)
+decision_tree = decision_tree.fit(iris.data, iris.target)
+r = export_text(decision_tree, feature_names=iris[<span style="color: #CD5555">&#39;feature_names&#39;</span>])
+<span style="color: #658b00">print</span>(r)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="different-kernels-and-mercer-s-theorem">Different kernels and Mercer's theorem </h2>
+<h2 id="algorithms-for-setting-up-decision-trees">Algorithms for Setting up Decision Trees </h2>
 
-<p>There are several popular kernels being used. These are</p>
+<p>Two algorithms stand out in the set up of decision trees:</p>
 <ol>
-<li> Linear: \( K(\boldsymbol{x},\boldsymbol{y})=\boldsymbol{x}^T\boldsymbol{y} \),</li>
-<li> Polynomial: \( K(\boldsymbol{x},\boldsymbol{y})=(\boldsymbol{x}^T\boldsymbol{y}+\gamma)^d \),</li>
-<li> Gaussian Radial Basis Function: \( K(\boldsymbol{x},\boldsymbol{y})=\exp{\left(-\gamma\vert\vert\boldsymbol{x}-\boldsymbol{y}\vert\vert^2\right)} \),</li>
-<li> Tanh: \( K(\boldsymbol{x},\boldsymbol{y})=\tanh{(\boldsymbol{x}^T\boldsymbol{y}+\gamma)} \),</li>
+<li> The CART (Classification And Regression Tree) algorithm for both classification and regression</li>
+<li> The ID3 algorithm based on the computation of the information gain for classification</li>
 </ol>
-<p>and many other ones.</p>
-
-<p>An important theorem for us is <a href="https://en.wikipedia.org/wiki/Mercer%27s_theorem" target="_blank">Mercer's
-theorem</a>.  The
-theorem states that if a kernel function \( K \) is symmetric, continuous
-and leads to a positive semi-definite matrix \( \boldsymbol{P} \) then there
-exists a function \( \phi \) that maps \( \boldsymbol{x}_i \) and \( \boldsymbol{x}_j \) into
-another space (possibly with much higher dimensions) such that
-</p>
-
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=\phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j).
-$$
-
-<p>So you can use \( K \) as a kernel since you know \( \phi \) exists, even if
-you don&#8217;t know what \( \phi \) is. 
-</p>
-
-<p>Note that some frequently used kernels (such as the Sigmoid kernel)
-don&#8217;t respect all of Mercer&#8217;s conditions, yet they generally work well
-in practice.
+<p>We discuss both algorithms with applications here. The popular library
+<b>Scikit-Learn</b> uses the CART algorithm. For classification problems
+you can use either the <b>gini</b> index or the <b>entropy</b> to split a tree
+in two branches.
 </p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="the-moons-example">The moons example </h2>
+<h2 id="the-cart-algorithm-for-classification">The CART algorithm for Classification </h2>
+
+<p>For classification, the CART algorithm splits the data set in two subsets using a single feature \( k \) and a threshold \( t_k \).
+This could be for example a threshold set by a number below a certain circumference of a malign tumor.
+</p>
+
+<p>How do we find these two quantities?
+We search for the pair \( (k,t_k) \) that produces the purest subset using for example the <b>gini</b> factor \( G \).
+The cost function it tries to minimize is then
+</p>
+$$
+C(k,t_k) = \frac{m_{\mathrm{left}}}{m}G_{\mathrm{left}}+ \frac{m_{\mathrm{right}}}{m}G_{\mathrm{right}},
+$$
+
+<p>where \( G_{\mathrm{left/right}} \) measures the impurity of the left/right subset  and \( m_{\mathrm{left/right}} \)
+ is the number of instances in the left/right subset
+</p>
+
+<p>Once it has successfully split the training set in two, it splits the subsets using the same logic, then the subsubsets
+and so on, recursively. It stops recursing once it reaches the maximum depth (defined by the
+\( max\_depth \) hyperparameter), or if it cannot find a split that will reduce impurity. A few other
+hyperparameters control additional stopping conditions such as the \( min\_samples\_split \),
+\( min\_samples\_leaf \), \( min\_weight\_fraction\_leaf \), and \( max\_leaf\_nodes \).
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="the-cart-algorithm-for-regression">The CART algorithm for Regression </h2>
+
+<p>The CART algorithm for regression works is similar to the one for classification except that instead of trying to split the
+training set in a way that minimizes say the <b>gini</b> or <b>entropy</b> impurity, it now tries to split the training set in a way that minimizes our well-known mean-squared error (MSE). The cost function is now
+</p>
+$$
+C(k,t_k) = \frac{m_{\mathrm{left}}}{m}\mathrm{MSE}_{\mathrm{left}}+ \frac{m_{\mathrm{right}}}{m}\mathrm{MSE}_{\mathrm{right}}.
+$$
+
+<p>Here the MSE for a specific node is defined as</p>
+$$
+\mathrm{MSE}_{\mathrm{node}}=\frac{1}{m_\mathrm{node}}\sum_{i\in \mathrm{node}}(\overline{y}_{\mathrm{node}}-y_i)^2,
+$$
+
+<p>with</p>
+$$
+\overline{y}_{\mathrm{node}}=\frac{1}{m_\mathrm{node}}\sum_{i\in \mathrm{node}}y_i,
+$$
+
+<p>the mean value of all observations in a specific node.</p>
+
+<p>Without any regularization, the regression task for decision trees, 
+just like for classification tasks, is  prone to overfitting.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="why-binary-splits">Why binary splits? </h2>
+
+<p>It is custom to split to a tree uising binary splits. The reason is
+that multiway splits fragment the data too quickly, leaving
+insufficient data at the next level down.  Multiway splits can be
+achieved by a series of binary split and this is normally preferred.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="computing-a-tree-using-the-gini-index">Computing a Tree using the Gini Index </h2>
+
+<p>Consider the following example with attributes/features and two
+possible outcomes (classes)  for each attribute.  Assume we wish to find some
+correlations between the average grade of a student as function of the
+number of hours studied and hours slept. We want also to correlate the
+grade in a given course with the general trend, whether the students
+recently has gotten grades below average or above.
+</p>
+
+<p>We have three features/attributes</p>
+<ol>
+<li> Trend of average grades before present course, classified as either below  or above  the average grade of the whole class</li>  
+<li> The number of hours studies, classified again as either higher (more than 3 hours per day) or lower . Here we have used a standard for one \( ECTS \) which is scaled to 25-30 hours of work for a semester which lasts 18 weeks, with 15 weeks of lectures and 3 weeks for exams, assuming a total of 30 ECTS per semester.</li> 
+<li> The number of hours slept as high for more than \( 8 \) hours and below  for less than 8 hours of sleep, classified again as either high or low</li>
+<li> The final grade whether it is above or below average</li>
+</ol>
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="the-table">The Table </h2>
+
+<table class="dotable" border="1">
+<thead>
+<tr><th align="center">Grade Trend</th> <th align="center">Hours slept</th> <th align="center">Hours Studied</th> <th align="center">Grade</th> </tr>
+</thead>
+<tbody>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   High           </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+</tbody>
+</table>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="computing-the-various-gini-indices">Computing the various Gini Indices </h2>
+
+<p>In computations we will translate all classes into numbers. Being
+these binary classes, they can easily be split into ones and zeros.
+</p>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Gini index for Average trend</b>
+<p>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="computing-the-various-gini-indices-hours-slept">Computing the various Gini Indices, Hours slept </h2>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Gini index for hour slept</b>
+<p>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="computing-the-various-gini-indices-hours-studied">Computing the various Gini Indices, Hours studied </h2>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Gini index for hour studied</b>
+<p>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+</div>
+
+
+<p>For final tree, see the above handwritten notes</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="a-possible-code-using-scikit-learn">A possible code using Scikit-Learn </h2>
+
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
 <div class="cell border-box-sizing code_cell rendered">
@@ -1048,193 +1111,994 @@ in practice.
     <div class="inner_cell">
       <div class="input_area">
         <div class="highlight" style="background: #eeeedd">
-  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">__future__</span> <span style="color: #8B008B; font-weight: bold">import</span> division, print_function, unicode_literals
-
+  <pre style="line-height: 125%;"><span style="color: #228B22"># Common imports</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
-np.random.seed(<span style="color: #B452CD">42</span>)
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_graphviz
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler, OneHotEncoder
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.compose</span> <span style="color: #8B008B; font-weight: bold">import</span> ColumnTransformer
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> Image 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
 
+<span style="color: #228B22"># Where to save the figures and data files</span>
+PROJECT_ROOT_DIR = <span style="color: #CD5555">&quot;Results&quot;</span>
+FIGURE_ID = <span style="color: #CD5555">&quot;Results/FigureFiles&quot;</span>
+DATA_ID = <span style="color: #CD5555">&quot;DataFiles/&quot;</span>
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(PROJECT_ROOT_DIR):
+    os.mkdir(PROJECT_ROOT_DIR)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(FIGURE_ID):
+    os.makedirs(FIGURE_ID)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(DATA_ID):
+    os.makedirs(DATA_ID)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">image_path</span>(fig_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(FIGURE_ID, fig_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">data_path</span>(dat_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(DATA_ID, dat_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">save_fig</span>(fig_id):
+    plt.savefig(image_path(fig_id) + <span style="color: #CD5555">&quot;.png&quot;</span>, <span style="color: #658b00">format</span>=<span style="color: #CD5555">&#39;png&#39;</span>)
+
+infile = <span style="color: #658b00">open</span>(data_path(<span style="color: #CD5555">&quot;grades.csv&quot;</span>),<span style="color: #CD5555">&#39;r&#39;</span>)
+
+<span style="color: #228B22"># Read the experimental data with Pandas</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> display
+grades = pd.read_csv(infile)
+grades = pd.DataFrame(grades)
+display(grades)
+<span style="color: #228B22"># Features and targets</span>
+X = grades.loc[:, grades.columns != <span style="color: #CD5555">&#39;Grade&#39;</span>].values
+y = grades.loc[:, grades.columns == <span style="color: #CD5555">&#39;Grade&#39;</span>].values
+<span style="color: #658b00">print</span>(X)
+<span style="color: #228B22"># Then do a Classification tree</span>
+tree_clf = DecisionTreeClassifier(max_depth=<span style="color: #B452CD">2</span>)
+tree_clf.fit(X, y)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Train set accuracy with Decision Tree: {:.2f}&quot;</span>.format(tree_clf.score(X,y)))
+<span style="color: #228B22">#transfer to a decision tree graph</span>
+export_graphviz(
+    tree_clf,
+    out_file=<span style="color: #CD5555">&quot;DataFiles/grade.dot&quot;</span>,
+    rounded=<span style="color: #8B008B; font-weight: bold">True</span>,
+    filled=<span style="color: #8B008B; font-weight: bold">True</span>
+)
+cmd = <span style="color: #CD5555">&#39;dot -Tpng DataFiles/grade.dot -o DataFiles/grades.png&#39;</span>
+os.system(cmd)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="further-example-computing-the-gini-index">Further example: Computing the Gini index </h2>
+
+<p>The next example we will look at is a classical one in many Machine
+Learning applications. Based on various meteorological features, we
+have several so-called attributes which decide whether we at the end
+will do some outdoor activity like skiing, going for a bike ride etc
+etc.  The table here contains the feautures <b>outlook</b>, <b>temperature</b>,
+<b>humidity</b> and <b>wind</b>.  The target or output is whether we ride
+(True=1) or whether we do something else that day (False=0). The
+attributes for each feature are then sunny, overcast and rain for the
+outlook, hot, cold and mild for temperature, high and normal for
+humidity and weak and strong for wind.
+</p>
+
+<p>The table here summarizes the various attributes and</p>
+<table class="dotable" border="1">
+<thead>
+<tr><th align="center">Day</th> <th align="center">Outlook </th> <th align="center">Temperature</th> <th align="center">Humidity</th> <th align="center"> Wind </th> <th align="center">Ride</th> </tr>
+</thead>
+<tbody>
+<tr><td align="center">   1      </td> <td align="center">   Sunny       </td> <td align="center">   Hot            </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   0       </td> </tr>
+<tr><td align="center">   2      </td> <td align="center">   Sunny       </td> <td align="center">   Hot            </td> <td align="center">   High        </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   3      </td> <td align="center">   Overcast    </td> <td align="center">   Hot            </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   4      </td> <td align="center">   Rain        </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   5      </td> <td align="center">   Rain        </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   6      </td> <td align="center">   Rain        </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Strong    </td> <td align="center">   0       </td> </tr>
+<tr><td align="center">   7      </td> <td align="center">   Overcast    </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   8      </td> <td align="center">   Sunny       </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   0       </td> </tr>
+<tr><td align="center">   9      </td> <td align="center">   Sunny       </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   10     </td> <td align="center">   Rain        </td> <td align="center">   Mild           </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   11     </td> <td align="center">   Sunny       </td> <td align="center">   Mild           </td> <td align="center">   Normal      </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   12     </td> <td align="center">   Overcast    </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   13     </td> <td align="center">   Overcast    </td> <td align="center">   Hot            </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   14     </td> <td align="center">   Rain        </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Strong    </td> <td align="center">   0       </td> </tr>
+</tbody>
+</table>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="simple-python-code-to-read-in-data-and-perform-classification">Simple Python Code to read in Data and perform Classification </h2>
+
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #228B22"># Common imports</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_graphviz
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler, OneHotEncoder
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.compose</span> <span style="color: #8B008B; font-weight: bold">import</span> ColumnTransformer
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> Image 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
+
+<span style="color: #228B22"># Where to save the figures and data files</span>
+PROJECT_ROOT_DIR = <span style="color: #CD5555">&quot;Results&quot;</span>
+FIGURE_ID = <span style="color: #CD5555">&quot;Results/FigureFiles&quot;</span>
+DATA_ID = <span style="color: #CD5555">&quot;DataFiles/&quot;</span>
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(PROJECT_ROOT_DIR):
+    os.mkdir(PROJECT_ROOT_DIR)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(FIGURE_ID):
+    os.makedirs(FIGURE_ID)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(DATA_ID):
+    os.makedirs(DATA_ID)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">image_path</span>(fig_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(FIGURE_ID, fig_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">data_path</span>(dat_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(DATA_ID, dat_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">save_fig</span>(fig_id):
+    plt.savefig(image_path(fig_id) + <span style="color: #CD5555">&quot;.png&quot;</span>, <span style="color: #658b00">format</span>=<span style="color: #CD5555">&#39;png&#39;</span>)
+
+infile = <span style="color: #658b00">open</span>(data_path(<span style="color: #CD5555">&quot;rideclass.csv&quot;</span>),<span style="color: #CD5555">&#39;r&#39;</span>)
+
+<span style="color: #228B22"># Read the experimental data with Pandas</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> display
+ridedata = pd.read_csv(infile,names = (<span style="color: #CD5555">&#39;Outlook&#39;</span>,<span style="color: #CD5555">&#39;Temperature&#39;</span>,<span style="color: #CD5555">&#39;Humidity&#39;</span>,<span style="color: #CD5555">&#39;Wind&#39;</span>,<span style="color: #CD5555">&#39;Ride&#39;</span>))
+ridedata = pd.DataFrame(ridedata)
+
+<span style="color: #228B22"># Features and targets</span>
+X = ridedata.loc[:, ridedata.columns != <span style="color: #CD5555">&#39;Ride&#39;</span>].values
+y = ridedata.loc[:, ridedata.columns == <span style="color: #CD5555">&#39;Ride&#39;</span>].values
+
+<span style="color: #228B22"># Create the encoder.</span>
+encoder = OneHotEncoder(handle_unknown=<span style="color: #CD5555">&quot;ignore&quot;</span>)
+<span style="color: #228B22"># Assume for simplicity all features are categorical.</span>
+encoder.fit(X)    
+<span style="color: #228B22"># Apply the encoder.</span>
+X = encoder.transform(X)
+<span style="color: #658b00">print</span>(X)
+<span style="color: #228B22"># Then do a Classification tree</span>
+tree_clf = DecisionTreeClassifier(max_depth=<span style="color: #B452CD">2</span>)
+tree_clf.fit(X, y)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Train set accuracy with Decision Tree: {:.2f}&quot;</span>.format(tree_clf.score(X,y)))
+<span style="color: #228B22">#transfer to a decision tree graph</span>
+export_graphviz(
+    tree_clf,
+    out_file=<span style="color: #CD5555">&quot;DataFiles/ride.dot&quot;</span>,
+    rounded=<span style="color: #8B008B; font-weight: bold">True</span>,
+    filled=<span style="color: #8B008B; font-weight: bold">True</span>
+)
+cmd = <span style="color: #CD5555">&#39;dot -Tpng DataFiles/cancer.dot -o DataFiles/cancer.png&#39;</span>
+os.system(cmd)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="computing-the-gini-factor">Computing the Gini Factor  </h2>
+
+<p>The above functions (gini, entropy and misclassification error) are
+important components of the so-called CART algorithm. We will discuss
+this algorithm below after we have discussed the information gain
+algorithm ID3.
+</p>
+
+<p>In the example here we have converted all our attributes into numerical values \( 0,1,2 \) etc.</p>
+
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #228B22"># Split a dataset based on an attribute and an attribute value</span>
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">test_split</span>(index, value, dataset):
+	left, right = <span style="color: #658b00">list</span>(), <span style="color: #658b00">list</span>()
+	<span style="color: #8B008B; font-weight: bold">for</span> row <span style="color: #8B008B">in</span> dataset:
+		<span style="color: #8B008B; font-weight: bold">if</span> row[index] &lt; value:
+			left.append(row)
+		<span style="color: #8B008B; font-weight: bold">else</span>:
+			right.append(row)
+	<span style="color: #8B008B; font-weight: bold">return</span> left, right
+ 
+<span style="color: #228B22"># Calculate the Gini index for a split dataset</span>
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">gini_index</span>(groups, classes):
+	<span style="color: #228B22"># count all samples at split point</span>
+	n_instances = <span style="color: #658b00">float</span>(<span style="color: #658b00">sum</span>([<span style="color: #658b00">len</span>(group) <span style="color: #8B008B; font-weight: bold">for</span> group <span style="color: #8B008B">in</span> groups]))
+	<span style="color: #228B22"># sum weighted Gini index for each group</span>
+	gini = <span style="color: #B452CD">0.0</span>
+	<span style="color: #8B008B; font-weight: bold">for</span> group <span style="color: #8B008B">in</span> groups:
+		size = <span style="color: #658b00">float</span>(<span style="color: #658b00">len</span>(group))
+		<span style="color: #228B22"># avoid divide by zero</span>
+		<span style="color: #8B008B; font-weight: bold">if</span> size == <span style="color: #B452CD">0</span>:
+			<span style="color: #8B008B; font-weight: bold">continue</span>
+		score = <span style="color: #B452CD">0.0</span>
+		<span style="color: #228B22"># score the group based on the score for each class</span>
+		<span style="color: #8B008B; font-weight: bold">for</span> class_val <span style="color: #8B008B">in</span> classes:
+			p = [row[-<span style="color: #B452CD">1</span>] <span style="color: #8B008B; font-weight: bold">for</span> row <span style="color: #8B008B">in</span> group].count(class_val) / size
+			score += p * p
+		<span style="color: #228B22"># weight the group score by its relative size</span>
+		gini += (<span style="color: #B452CD">1.0</span> - score) * (size / n_instances)
+	<span style="color: #8B008B; font-weight: bold">return</span> gini
+
+<span style="color: #228B22"># Select the best split point for a dataset</span>
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">get_split</span>(dataset):
+	class_values = <span style="color: #658b00">list</span>(<span style="color: #658b00">set</span>(row[-<span style="color: #B452CD">1</span>] <span style="color: #8B008B; font-weight: bold">for</span> row <span style="color: #8B008B">in</span> dataset))
+	b_index, b_value, b_score, b_groups = <span style="color: #B452CD">999</span>, <span style="color: #B452CD">999</span>, <span style="color: #B452CD">999</span>, <span style="color: #8B008B; font-weight: bold">None</span>
+	<span style="color: #8B008B; font-weight: bold">for</span> index <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(<span style="color: #658b00">len</span>(dataset[<span style="color: #B452CD">0</span>])-<span style="color: #B452CD">1</span>):
+		<span style="color: #8B008B; font-weight: bold">for</span> row <span style="color: #8B008B">in</span> dataset:
+			groups = test_split(index, row[index], dataset)
+			gini = gini_index(groups, class_values)
+			<span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;X%d &lt; %.3f Gini=%.3f&#39;</span> % ((index+<span style="color: #B452CD">1</span>), row[index], gini))
+			<span style="color: #8B008B; font-weight: bold">if</span> gini &lt; b_score:
+				b_index, b_value, b_score, b_groups = index, row[index], gini, groups
+	<span style="color: #8B008B; font-weight: bold">return</span> {<span style="color: #CD5555">&#39;index&#39;</span>:b_index, <span style="color: #CD5555">&#39;value&#39;</span>:b_value, <span style="color: #CD5555">&#39;groups&#39;</span>:b_groups}
+ 
+dataset = [[<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>],
+            [<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">2</span>,<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">2</span>,<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>],
+            [<span style="color: #B452CD">1</span>,<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>],
+            [<span style="color: #B452CD">0</span>,<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>],
+            [<span style="color: #B452CD">2</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>]]
+
+split = get_split(dataset)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;Split: [X%d &lt; %.3f]&#39;</span> % ((split[<span style="color: #CD5555">&#39;index&#39;</span>]+<span style="color: #B452CD">1</span>), split[<span style="color: #CD5555">&#39;value&#39;</span>]))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="regression-trees">Regression trees </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #228B22"># Quadratic training set + noise</span>
+np.random.seed(<span style="color: #B452CD">42</span>)
+m = <span style="color: #B452CD">200</span>
+X = np.random.rand(m, <span style="color: #B452CD">1</span>)
+y = <span style="color: #B452CD">4</span> * (X - <span style="color: #B452CD">0.5</span>) ** <span style="color: #B452CD">2</span>
+y = y + np.random.randn(m, <span style="color: #B452CD">1</span>) / <span style="color: #B452CD">10</span>
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeRegressor
+
+tree_reg = DecisionTreeRegressor(max_depth=<span style="color: #B452CD">2</span>, random_state=<span style="color: #B452CD">42</span>)
+tree_reg.fit(X, y)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="final-regressor-code">Final regressor code </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeRegressor
+
+tree_reg1 = DecisionTreeRegressor(random_state=<span style="color: #B452CD">42</span>, max_depth=<span style="color: #B452CD">2</span>)
+tree_reg2 = DecisionTreeRegressor(random_state=<span style="color: #B452CD">42</span>, max_depth=<span style="color: #B452CD">3</span>)
+tree_reg1.fit(X, y)
+tree_reg2.fit(X, y)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">plot_regression_predictions</span>(tree_reg, X, y, axes=[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, -<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1</span>], ylabel=<span style="color: #CD5555">&quot;$y$&quot;</span>):
+    x1 = np.linspace(axes[<span style="color: #B452CD">0</span>], axes[<span style="color: #B452CD">1</span>], <span style="color: #B452CD">500</span>).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
+    y_pred = tree_reg.predict(x1)
+    plt.axis(axes)
+    plt.xlabel(<span style="color: #CD5555">&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
+    <span style="color: #8B008B; font-weight: bold">if</span> ylabel:
+        plt.ylabel(ylabel, fontsize=<span style="color: #B452CD">18</span>, rotation=<span style="color: #B452CD">0</span>)
+    plt.plot(X, y, <span style="color: #CD5555">&quot;b.&quot;</span>)
+    plt.plot(x1, y_pred, <span style="color: #CD5555">&quot;r.-&quot;</span>, linewidth=<span style="color: #B452CD">2</span>, label=<span style="color: #CD5555">r&quot;$\hat{y}$&quot;</span>)
+
+plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
+plt.subplot(<span style="color: #B452CD">121</span>)
+plot_regression_predictions(tree_reg1, X, y)
+<span style="color: #8B008B; font-weight: bold">for</span> split, style <span style="color: #8B008B">in</span> ((<span style="color: #B452CD">0.1973</span>, <span style="color: #CD5555">&quot;k-&quot;</span>), (<span style="color: #B452CD">0.0917</span>, <span style="color: #CD5555">&quot;k--&quot;</span>), (<span style="color: #B452CD">0.7718</span>, <span style="color: #CD5555">&quot;k--&quot;</span>)):
+    plt.plot([split, split], [-<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1</span>], style, linewidth=<span style="color: #B452CD">2</span>)
+plt.text(<span style="color: #B452CD">0.21</span>, <span style="color: #B452CD">0.65</span>, <span style="color: #CD5555">&quot;Depth=0&quot;</span>, fontsize=<span style="color: #B452CD">15</span>)
+plt.text(<span style="color: #B452CD">0.01</span>, <span style="color: #B452CD">0.2</span>, <span style="color: #CD5555">&quot;Depth=1&quot;</span>, fontsize=<span style="color: #B452CD">13</span>)
+plt.text(<span style="color: #B452CD">0.65</span>, <span style="color: #B452CD">0.8</span>, <span style="color: #CD5555">&quot;Depth=1&quot;</span>, fontsize=<span style="color: #B452CD">13</span>)
+plt.legend(loc=<span style="color: #CD5555">&quot;upper center&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
+plt.title(<span style="color: #CD5555">&quot;max_depth=2&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
+
+plt.subplot(<span style="color: #B452CD">122</span>)
+plot_regression_predictions(tree_reg2, X, y, ylabel=<span style="color: #8B008B; font-weight: bold">None</span>)
+<span style="color: #8B008B; font-weight: bold">for</span> split, style <span style="color: #8B008B">in</span> ((<span style="color: #B452CD">0.1973</span>, <span style="color: #CD5555">&quot;k-&quot;</span>), (<span style="color: #B452CD">0.0917</span>, <span style="color: #CD5555">&quot;k--&quot;</span>), (<span style="color: #B452CD">0.7718</span>, <span style="color: #CD5555">&quot;k--&quot;</span>)):
+    plt.plot([split, split], [-<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1</span>], style, linewidth=<span style="color: #B452CD">2</span>)
+<span style="color: #8B008B; font-weight: bold">for</span> split <span style="color: #8B008B">in</span> (<span style="color: #B452CD">0.0458</span>, <span style="color: #B452CD">0.1298</span>, <span style="color: #B452CD">0.2873</span>, <span style="color: #B452CD">0.9040</span>):
+    plt.plot([split, split], [-<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;k:&quot;</span>, linewidth=<span style="color: #B452CD">1</span>)
+plt.text(<span style="color: #B452CD">0.3</span>, <span style="color: #B452CD">0.5</span>, <span style="color: #CD5555">&quot;Depth=2&quot;</span>, fontsize=<span style="color: #B452CD">13</span>)
+plt.title(<span style="color: #CD5555">&quot;max_depth=3&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
+
+plt.show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;">tree_reg1 = DecisionTreeRegressor(random_state=<span style="color: #B452CD">42</span>)
+tree_reg2 = DecisionTreeRegressor(random_state=<span style="color: #B452CD">42</span>, min_samples_leaf=<span style="color: #B452CD">10</span>)
+tree_reg1.fit(X, y)
+tree_reg2.fit(X, y)
+
+x1 = np.linspace(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">500</span>).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
+y_pred1 = tree_reg1.predict(x1)
+y_pred2 = tree_reg2.predict(x1)
+
+plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
+
+plt.subplot(<span style="color: #B452CD">121</span>)
+plt.plot(X, y, <span style="color: #CD5555">&quot;b.&quot;</span>)
+plt.plot(x1, y_pred1, <span style="color: #CD5555">&quot;r.-&quot;</span>, linewidth=<span style="color: #B452CD">2</span>, label=<span style="color: #CD5555">r&quot;$\hat{y}$&quot;</span>)
+plt.axis([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, -<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1.1</span>])
+plt.xlabel(<span style="color: #CD5555">&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
+plt.ylabel(<span style="color: #CD5555">&quot;$y$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>, rotation=<span style="color: #B452CD">0</span>)
+plt.legend(loc=<span style="color: #CD5555">&quot;upper center&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
+plt.title(<span style="color: #CD5555">&quot;No restrictions&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
+
+plt.subplot(<span style="color: #B452CD">122</span>)
+plt.plot(X, y, <span style="color: #CD5555">&quot;b.&quot;</span>)
+plt.plot(x1, y_pred2, <span style="color: #CD5555">&quot;r.-&quot;</span>, linewidth=<span style="color: #B452CD">2</span>, label=<span style="color: #CD5555">r&quot;$\hat{y}$&quot;</span>)
+plt.axis([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, -<span style="color: #B452CD">0.2</span>, <span style="color: #B452CD">1.1</span>])
+plt.xlabel(<span style="color: #CD5555">&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
+plt.title(<span style="color: #CD5555">&quot;min_samples_leaf={}&quot;</span>.format(tree_reg2.min_samples_leaf), fontsize=<span style="color: #B452CD">14</span>)
+
+plt.show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="pros-and-cons-of-trees-pros">Pros and cons of trees, pros </h2>
+
+<ul>
+<li> White box, easy to interpret model. Some people believe that decision trees more closely mirror human decision-making than do the regression and classification approaches discussed earlier (think of support vector machines)</li>
+<li> Trees are very easy to explain to people. In fact, they are even easier to explain than linear regression!</li>
+<li> No feature normalization needed</li>
+<li> Tree models can handle both continuous and categorical data (Classification and Regression Trees)</li>
+<li> Can model nonlinear relationships</li>
+<li> Can model interactions between the different descriptive features</li>
+<li> Trees can be displayed graphically, and are easily interpreted even by a non-expert (especially if they are small)</li>
+</ul>
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="disadvantages">Disadvantages </h2>
+
+<ul>
+<li> Unfortunately, trees generally do not have the same level of predictive accuracy as some of the other regression and classification approaches</li>
+<li> If continuous features are used the tree may become quite large and hence less interpretable</li>
+<li> Decision trees are prone to overfit the training data and hence do not well generalize the data if no stopping criteria or improvements like pruning, boosting or bagging are implemented</li>
+<li> Small changes in the data may lead to a completely different tree. This issue can be addressed by using ensemble methods like bagging, boosting or random forests</li>
+<li> Unbalanced datasets where some target feature values occur much more frequently than others may lead to biased trees since the frequently occurring feature values are preferred over the less frequently occurring ones.</li> 
+<li> If the number of features is relatively large (high dimensional) and the number of instances is relatively low, the tree might overfit the data</li>
+<li> Features with many levels may be preferred over features with less levels since for them it is <em>more easy</em> to split the dataset such that the sub datasets only contain pure target feature values. This issue can be addressed by preferring for instance the information gain ratio as splitting criteria over information gain</li>
+</ul>
+<p>However, by aggregating many decision trees, using methods like
+bagging, random forests, and boosting, the predictive performance of
+trees can be substantially improved.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods </h2>
+
+<p>As stated above and seen in many of the examples discussed here about
+a single decision tree, we often end up overfitting our training
+data. This normally means that we have a high variance. Can we reduce
+the variance of a statistical learning method?
+</p>
+
+<p>This leads us to a set of different methods that can combine different
+machine learning algorithms or just use one of them to construct
+forests and jungles of trees, homogeneous ones or heterogenous
+ones. These methods are recognized by different names which we will
+try to explain here. These are
+</p>
+
+<ol>
+<li> Voting classifiers</li>
+<li> Bagging and Pasting</li>
+<li> Random forests</li>
+<li> Boosting methods, from adaptive to Extreme Gradient Boosting (XGBoost)</li>
+</ol>
+<p>We discuss these methods here.</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="an-overview-of-ensemble-methods">An Overview of Ensemble Methods  </h2>
+
+<br/><br/>
+<center>
+<p><img src="DataFiles/ensembleoverview.png" width="600" align="bottom"></p>
+</center>
+<br/><br/>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="why-voting">Why Voting? </h2>
+
+<p>The idea behind boosting, and voting as well can be phrased as follows:
+<b>Can a group of people somehow arrive at highly
+reasoned decisions, despite the weak judgement of the individual
+members?</b>
+</p>
+
+<p>The aim is to create a good classifier by combining several weak classifiers.
+<b>A weak classifier is a classifier which is able to produce results that are only slightly better than guessing at random.</b>
+</p>
+
+<p>The basic approach is to apply repeatedly (in boosting this is done in an iterative way) a weak classifier to modifications of the data.
+In voting we simply apply the law of large numbers while in boosting we give more weight to misclassified data in
+each iteration.  
+</p>
+
+<p>Decision trees play an important role as our weak classifier. They serve as the basic method. </p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="tossing-coins">Tossing coins </h2>
+
+<p>The simplest case is a so-called voting ensemble. To illustrate this,
+think of yourself tossing coins with a biased outcome of 51 per cent
+for heads and 49% for tails. With only few tosses,
+you may not clearly see this distribution for heads and tails. However, after some
+thousands of tosses, there will be a clear majority of heads.  With 2000 tosses
+you should see approximately 1020 heads and 980 tails.
+</p>
+
+<p>We can then state that the outcome is a clear majority of heads. If
+you do this ten thousand times, it is easy to see that there is a 97%
+likelihood of a majority of heads.
+</p>
+
+<p>Another example would be to collect all polls before an
+election. Different polls may show different likelihoods for a
+candidate winning with say a majority  of the popular vote. The majority vote
+would then consist in many polls indicating that this candidate will
+actually win.
+</p>
+
+<p>The example here shows how we can implement the coin tossing case,
+clealry demostrating that after some tosses we see the <a href="https://en.wikipedia.org/wiki/Law_of_large_numbers" target="_blank">law of large</a>
+numbers kicking in.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="standard-imports-first">Standard imports first </h2>
+
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #228B22"># Common imports</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> Image 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> export_graphviz
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler, OneHotEncoder
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.compose</span> <span style="color: #8B008B; font-weight: bold">import</span> ColumnTransformer
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> Image 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">pydot</span> <span style="color: #8B008B; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
+
+<span style="color: #228B22"># Where to save the figures and data files</span>
+PROJECT_ROOT_DIR = <span style="color: #CD5555">&quot;Results&quot;</span>
+FIGURE_ID = <span style="color: #CD5555">&quot;Results/FigureFiles&quot;</span>
+DATA_ID = <span style="color: #CD5555">&quot;DataFiles/&quot;</span>
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(PROJECT_ROOT_DIR):
+    os.mkdir(PROJECT_ROOT_DIR)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(FIGURE_ID):
+    os.makedirs(FIGURE_ID)
+
+<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(DATA_ID):
+    os.makedirs(DATA_ID)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">image_path</span>(fig_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(FIGURE_ID, fig_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">data_path</span>(dat_id):
+    <span style="color: #8B008B; font-weight: bold">return</span> os.path.join(DATA_ID, dat_id)
+
+<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">save_fig</span>(fig_id):
+    plt.savefig(image_path(fig_id) + <span style="color: #CD5555">&quot;.png&quot;</span>, <span style="color: #658b00">format</span>=<span style="color: #CD5555">&#39;png&#39;</span>)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="simple-voting-example-head-or-tail">Simple Voting Example, head or tail </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #228B22"># Common imports</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">matplotlib.colors</span> <span style="color: #8B008B; font-weight: bold">import</span> ListedColormap
 plt.rcParams[<span style="color: #CD5555">&#39;axes.labelsize&#39;</span>] = <span style="color: #B452CD">14</span>
 plt.rcParams[<span style="color: #CD5555">&#39;xtick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
 plt.rcParams[<span style="color: #CD5555">&#39;ytick.labelsize&#39;</span>] = <span style="color: #B452CD">12</span>
 
+heads_proba = <span style="color: #B452CD">0.51</span>
+coin_tosses = (np.random.rand(<span style="color: #B452CD">10000</span>, <span style="color: #B452CD">10</span>) &lt; heads_proba).astype(np.int32)
+cumulative_heads_ratio = np.cumsum(coin_tosses, axis=<span style="color: #B452CD">0</span>) / np.arange(<span style="color: #B452CD">1</span>, <span style="color: #B452CD">10001</span>).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
+plt.figure(figsize=(<span style="color: #B452CD">8</span>,<span style="color: #B452CD">3.5</span>))
+plt.plot(cumulative_heads_ratio)
+plt.plot([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">10000</span>], [<span style="color: #B452CD">0.51</span>, <span style="color: #B452CD">0.51</span>], <span style="color: #CD5555">&quot;k--&quot;</span>, linewidth=<span style="color: #B452CD">2</span>, label=<span style="color: #CD5555">&quot;51%&quot;</span>)
+plt.plot([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">10000</span>], [<span style="color: #B452CD">0.5</span>, <span style="color: #B452CD">0.5</span>], <span style="color: #CD5555">&quot;k-&quot;</span>, label=<span style="color: #CD5555">&quot;50%&quot;</span>)
+plt.xlabel(<span style="color: #CD5555">&quot;Number of coin tosses&quot;</span>)
+plt.ylabel(<span style="color: #CD5555">&quot;Heads ratio&quot;</span>)
+plt.legend(loc=<span style="color: #CD5555">&quot;lower right&quot;</span>)
+plt.axis([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">10000</span>, <span style="color: #B452CD">0.42</span>, <span style="color: #B452CD">0.58</span>])
+save_fig(<span style="color: #CD5555">&quot;votingsimple&quot;</span>)
+plt.show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
 
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="using-the-voting-classifier">Using the Voting Classifier </h2>
+
+<p>We can use the voting classifier on other data sets, here the exciting binary case of two distinct objects using the make moons functionality of <b>Scikit-Learn</b>.</p>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> make_moons
+
+X, y = make_moons(n_samples=<span style="color: #B452CD">500</span>, noise=<span style="color: #B452CD">0.30</span>, random_state=<span style="color: #B452CD">42</span>)
+X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=<span style="color: #B452CD">42</span>)
+
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> RandomForestClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> VotingClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> LogisticRegression
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets
+
+log_clf = LogisticRegression(solver=<span style="color: #CD5555">&quot;liblinear&quot;</span>, random_state=<span style="color: #B452CD">42</span>)
+rnd_clf = RandomForestClassifier(n_estimators=<span style="color: #B452CD">10</span>, random_state=<span style="color: #B452CD">42</span>)
+svm_clf = SVC(gamma=<span style="color: #CD5555">&quot;auto&quot;</span>, random_state=<span style="color: #B452CD">42</span>)
+
+voting_clf = VotingClassifier(
+    estimators=[(<span style="color: #CD5555">&#39;lr&#39;</span>, log_clf), (<span style="color: #CD5555">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #CD5555">&#39;svc&#39;</span>, svm_clf)],
+    voting=<span style="color: #CD5555">&#39;hard&#39;</span>)
+
+voting_clf.fit(X_train, y_train)
+
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #8B008B; font-weight: bold">for</span> clf <span style="color: #8B008B">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    <span style="color: #658b00">print</span>(clf.<span style="color: #00688B">__class__</span>.<span style="color: #00688B">__name__</span>, accuracy_score(y_test, y_pred))
+
+log_clf = LogisticRegression(solver=<span style="color: #CD5555">&quot;liblinear&quot;</span>, random_state=<span style="color: #B452CD">42</span>)
+rnd_clf = RandomForestClassifier(n_estimators=<span style="color: #B452CD">10</span>, random_state=<span style="color: #B452CD">42</span>)
+svm_clf = SVC(gamma=<span style="color: #CD5555">&quot;auto&quot;</span>, probability=<span style="color: #8B008B; font-weight: bold">True</span>, random_state=<span style="color: #B452CD">42</span>)
+voting_clf = VotingClassifier(
+    estimators=[(<span style="color: #CD5555">&#39;lr&#39;</span>, log_clf), (<span style="color: #CD5555">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #CD5555">&#39;svc&#39;</span>, svm_clf)],
+    voting=<span style="color: #CD5555">&#39;soft&#39;</span>)
+voting_clf.fit(X_train, y_train)
+
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #8B008B; font-weight: bold">for</span> clf <span style="color: #8B008B">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    <span style="color: #658b00">print</span>(clf.<span style="color: #00688B">__class__</span>.<span style="color: #00688B">__name__</span>, accuracy_score(y_test, y_pred))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
 
 
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="voting-and-bagging">Voting and Bagging </h2>
+
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> make_moons
+
+X, y = make_moons(n_samples=<span style="color: #B452CD">500</span>, noise=<span style="color: #B452CD">0.30</span>, random_state=<span style="color: #B452CD">42</span>)
+X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=<span style="color: #B452CD">42</span>)
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> RandomForestClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> VotingClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> LogisticRegression
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
+
+log_clf = LogisticRegression(random_state=<span style="color: #B452CD">42</span>)
+rnd_clf = RandomForestClassifier(random_state=<span style="color: #B452CD">42</span>)
+svm_clf = SVC(random_state=<span style="color: #B452CD">42</span>)
+
+voting_clf = VotingClassifier(
+    estimators=[(<span style="color: #CD5555">&#39;lr&#39;</span>, log_clf), (<span style="color: #CD5555">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #CD5555">&#39;svc&#39;</span>, svm_clf)],
+    voting=<span style="color: #CD5555">&#39;hard&#39;</span>)
+voting_clf.fit(X_train, y_train)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #8B008B; font-weight: bold">for</span> clf <span style="color: #8B008B">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    <span style="color: #658b00">print</span>(clf.<span style="color: #00688B">__class__</span>.<span style="color: #00688B">__name__</span>, accuracy_score(y_test, y_pred))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;">log_clf = LogisticRegression(random_state=<span style="color: #B452CD">42</span>)
+rnd_clf = RandomForestClassifier(random_state=<span style="color: #B452CD">42</span>)
+svm_clf = SVC(probability=<span style="color: #8B008B; font-weight: bold">True</span>, random_state=<span style="color: #B452CD">42</span>)
+
+voting_clf = VotingClassifier(
+    estimators=[(<span style="color: #CD5555">&#39;lr&#39;</span>, log_clf), (<span style="color: #CD5555">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #CD5555">&#39;svc&#39;</span>, svm_clf)],
+    voting=<span style="color: #CD5555">&#39;soft&#39;</span>)
+voting_clf.fit(X_train, y_train)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #8B008B; font-weight: bold">for</span> clf <span style="color: #8B008B">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    <span style="color: #658b00">print</span>(clf.<span style="color: #00688B">__class__</span>.<span style="color: #00688B">__name__</span>, accuracy_score(y_test, y_pred))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="bagging">Bagging </h2>
+
+<p>The <b>plain</b> decision trees suffer from high
+variance. This means that if we split the training data into two parts
+at random, and fit a decision tree to both halves, the results that we
+get could be quite different. In contrast, a procedure with low
+variance will yield similar results if applied repeatedly to distinct
+data sets; linear regression tends to have low variance, if the ratio
+of \( n \) to \( p \) is moderately large. 
+</p>
+
+<p><b>Bootstrap aggregation</b>, or just <b>bagging</b>, is a
+general-purpose procedure for reducing the variance of a statistical
+learning method. 
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="more-bagging">More bagging </h2>
+
+<p>Bagging typically results in improved accuracy
+over prediction using a single tree. Unfortunately, however, it can be
+difficult to interpret the resulting model. Recall that one of the
+advantages of decision trees is the attractive and easily interpreted
+diagram that results.
+</p>
+
+<p>However, when we bag a large number of trees, it is no longer
+possible to represent the resulting statistical learning procedure
+using a single tree, and it is no longer clear which variables are
+most important to the procedure. Thus, bagging improves prediction
+accuracy at the expense of interpretability.  Although the collection
+of bagged trees is much more difficult to interpret than a single
+tree, one can obtain an overall summary of the importance of each
+predictor using the MSE (for bagging regression trees) or the Gini
+index (for bagging classification trees). In the case of bagging
+regression trees, we can record the total amount that the MSE is
+decreased due to splits over a given predictor, averaged over all \( B \) possible
+trees. A large value indicates an important predictor. Similarly, in
+the context of bagging classification trees, we can add up the total
+amount that the Gini index  is decreased by splits over a given
+predictor, averaged over all \( B \) trees.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="making-your-own-bootstrap-changing-the-level-of-the-decision-tree">Making your own Bootstrap: Changing the Level of the Decision Tree </h2>
+
+<p>Let us bring up our good old boostrap example from the linear regression lectures. We change the linerar regression algorithm with
+a decision tree wth different depths and perform a bootstrap aggregate (in this case we perform as many bootstraps as data points \( n \)).
+</p>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.pipeline</span> <span style="color: #8B008B; font-weight: bold">import</span> make_pipeline
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.utils</span> <span style="color: #8B008B; font-weight: bold">import</span> resample
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeRegressor
+
+n = <span style="color: #B452CD">100</span>
+n_boostraps = <span style="color: #B452CD">100</span>
+maxdepth = <span style="color: #B452CD">8</span>
+
+<span style="color: #228B22"># Make data set.</span>
+x = np.linspace(-<span style="color: #B452CD">3</span>, <span style="color: #B452CD">3</span>, n).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
+y = np.exp(-x**<span style="color: #B452CD">2</span>) + <span style="color: #B452CD">1.5</span> * np.exp(-(x-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>)+ np.random.normal(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0.1</span>, x.shape)
+error = np.zeros(maxdepth)
+bias = np.zeros(maxdepth)
+variance = np.zeros(maxdepth)
+polydegree = np.zeros(maxdepth)
+X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=<span style="color: #B452CD">0.2</span>)
 
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.pipeline</span> <span style="color: #8B008B; font-weight: bold">import</span> Pipeline
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> LinearSVC
+scaler = StandardScaler()
+scaler.fit(X_train)
+X_train_scaled = scaler.transform(X_train)
+X_test_scaled = scaler.transform(X_test)
 
+<span style="color: #228B22"># we produce a simple tree first as benchmark</span>
+simpletree = DecisionTreeRegressor(max_depth=<span style="color: #B452CD">3</span>) 
+simpletree.fit(X_train_scaled, y_train)
+simpleprediction = simpletree.predict(X_test_scaled)
+<span style="color: #8B008B; font-weight: bold">for</span> degree <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(<span style="color: #B452CD">1</span>,maxdepth):
+    model = DecisionTreeRegressor(max_depth=degree) 
+    y_pred = np.empty((y_test.shape[<span style="color: #B452CD">0</span>], n_boostraps))
+    <span style="color: #8B008B; font-weight: bold">for</span> i <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(n_boostraps):
+        x_, y_ = resample(X_train_scaled, y_train)
+        model.fit(x_, y_)
+        y_pred[:, i] = model.predict(X_test_scaled)<span style="color: #228B22">#.ravel()</span>
 
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> make_moons
-X, y = make_moons(n_samples=<span style="color: #B452CD">100</span>, noise=<span style="color: #B452CD">0.15</span>, random_state=<span style="color: #B452CD">42</span>)
-
-<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">plot_dataset</span>(X, y, axes):
-    plt.plot(X[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">0</span>], X[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">0</span>], <span style="color: #CD5555">&quot;bs&quot;</span>)
-    plt.plot(X[:, <span style="color: #B452CD">0</span>][y==<span style="color: #B452CD">1</span>], X[:, <span style="color: #B452CD">1</span>][y==<span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;g^&quot;</span>)
-    plt.axis(axes)
-    plt.grid(<span style="color: #8B008B; font-weight: bold">True</span>, which=<span style="color: #CD5555">&#39;both&#39;</span>)
-    plt.xlabel(<span style="color: #CD5555">r&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-    plt.ylabel(<span style="color: #CD5555">r&quot;$x_2$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>, rotation=<span style="color: #B452CD">0</span>)
-
-plot_dataset(X, y, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plt.show()
-
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> make_moons
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.pipeline</span> <span style="color: #8B008B; font-weight: bold">import</span> Pipeline
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> PolynomialFeatures
-
-polynomial_svm_clf = Pipeline([
-        (<span style="color: #CD5555">&quot;poly_features&quot;</span>, PolynomialFeatures(degree=<span style="color: #B452CD">3</span>)),
-        (<span style="color: #CD5555">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #CD5555">&quot;svm_clf&quot;</span>, LinearSVC(C=<span style="color: #B452CD">10</span>, loss=<span style="color: #CD5555">&quot;hinge&quot;</span>, random_state=<span style="color: #B452CD">42</span>))
-    ])
-
-polynomial_svm_clf.fit(X, y)
-
-<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">plot_predictions</span>(clf, axes):
-    x0s = np.linspace(axes[<span style="color: #B452CD">0</span>], axes[<span style="color: #B452CD">1</span>], <span style="color: #B452CD">100</span>)
-    x1s = np.linspace(axes[<span style="color: #B452CD">2</span>], axes[<span style="color: #B452CD">3</span>], <span style="color: #B452CD">100</span>)
-    x0, x1 = np.meshgrid(x0s, x1s)
-    X = np.c_[x0.ravel(), x1.ravel()]
-    y_pred = clf.predict(X).reshape(x0.shape)
-    y_decision = clf.decision_function(X).reshape(x0.shape)
-    plt.contourf(x0, x1, y_pred, cmap=plt.cm.brg, alpha=<span style="color: #B452CD">0.2</span>)
-    plt.contourf(x0, x1, y_decision, cmap=plt.cm.brg, alpha=<span style="color: #B452CD">0.1</span>)
-
-plot_predictions(polynomial_svm_clf, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plot_dataset(X, y, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-
-plt.show()
-
-
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
-
-poly_kernel_svm_clf = Pipeline([
-        (<span style="color: #CD5555">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #CD5555">&quot;svm_clf&quot;</span>, SVC(kernel=<span style="color: #CD5555">&quot;poly&quot;</span>, degree=<span style="color: #B452CD">3</span>, coef0=<span style="color: #B452CD">1</span>, C=<span style="color: #B452CD">5</span>))
-    ])
-poly_kernel_svm_clf.fit(X, y)
-
-poly100_kernel_svm_clf = Pipeline([
-        (<span style="color: #CD5555">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #CD5555">&quot;svm_clf&quot;</span>, SVC(kernel=<span style="color: #CD5555">&quot;poly&quot;</span>, degree=<span style="color: #B452CD">10</span>, coef0=<span style="color: #B452CD">100</span>, C=<span style="color: #B452CD">5</span>))
-    ])
-poly100_kernel_svm_clf.fit(X, y)
-
-plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
-
-plt.subplot(<span style="color: #B452CD">121</span>)
-plot_predictions(poly_kernel_svm_clf, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plot_dataset(X, y, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plt.title(<span style="color: #CD5555">r&quot;$d=3, r=1, C=5$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
-
-plt.subplot(<span style="color: #B452CD">122</span>)
-plot_predictions(poly100_kernel_svm_clf, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plot_dataset(X, y, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-plt.title(<span style="color: #CD5555">r&quot;$d=10, r=100, C=5$&quot;</span>, fontsize=<span style="color: #B452CD">18</span>)
-
-plt.show()
-
-<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">gaussian_rbf</span>(x, landmark, gamma):
-    <span style="color: #8B008B; font-weight: bold">return</span> np.exp(-gamma * np.linalg.norm(x - landmark, axis=<span style="color: #B452CD">1</span>)**<span style="color: #B452CD">2</span>)
-
-gamma = <span style="color: #B452CD">0.3</span>
-
-x1s = np.linspace(-<span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">200</span>).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
-x2s = gaussian_rbf(x1s, -<span style="color: #B452CD">2</span>, gamma)
-x3s = gaussian_rbf(x1s, <span style="color: #B452CD">1</span>, gamma)
-
-XK = np.c_[gaussian_rbf(X1D, -<span style="color: #B452CD">2</span>, gamma), gaussian_rbf(X1D, <span style="color: #B452CD">1</span>, gamma)]
-yk = np.array([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>])
-
-plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">4</span>))
-
-plt.subplot(<span style="color: #B452CD">121</span>)
-plt.grid(<span style="color: #8B008B; font-weight: bold">True</span>, which=<span style="color: #CD5555">&#39;both&#39;</span>)
-plt.axhline(y=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.scatter(x=[-<span style="color: #B452CD">2</span>, <span style="color: #B452CD">1</span>], y=[<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0</span>], s=<span style="color: #B452CD">150</span>, alpha=<span style="color: #B452CD">0.5</span>, c=<span style="color: #CD5555">&quot;red&quot;</span>)
-plt.plot(X1D[:, <span style="color: #B452CD">0</span>][yk==<span style="color: #B452CD">0</span>], np.zeros(<span style="color: #B452CD">4</span>), <span style="color: #CD5555">&quot;bs&quot;</span>)
-plt.plot(X1D[:, <span style="color: #B452CD">0</span>][yk==<span style="color: #B452CD">1</span>], np.zeros(<span style="color: #B452CD">5</span>), <span style="color: #CD5555">&quot;g^&quot;</span>)
-plt.plot(x1s, x2s, <span style="color: #CD5555">&quot;g--&quot;</span>)
-plt.plot(x1s, x3s, <span style="color: #CD5555">&quot;b:&quot;</span>)
-plt.gca().get_yaxis().set_ticks([<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0.25</span>, <span style="color: #B452CD">0.5</span>, <span style="color: #B452CD">0.75</span>, <span style="color: #B452CD">1</span>])
-plt.xlabel(<span style="color: #CD5555">r&quot;$x_1$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.ylabel(<span style="color: #CD5555">r&quot;Similarity&quot;</span>, fontsize=<span style="color: #B452CD">14</span>)
-plt.annotate(<span style="color: #CD5555">r&#39;$\mathbf{x}$&#39;</span>,
-             xy=(X1D[<span style="color: #B452CD">3</span>, <span style="color: #B452CD">0</span>], <span style="color: #B452CD">0</span>),
-             xytext=(-<span style="color: #B452CD">0.5</span>, <span style="color: #B452CD">0.20</span>),
-             ha=<span style="color: #CD5555">&quot;center&quot;</span>,
-             arrowprops=<span style="color: #658b00">dict</span>(facecolor=<span style="color: #CD5555">&#39;black&#39;</span>, shrink=<span style="color: #B452CD">0.1</span>),
-             fontsize=<span style="color: #B452CD">18</span>,
-            )
-plt.text(-<span style="color: #B452CD">2</span>, <span style="color: #B452CD">0.9</span>, <span style="color: #CD5555">&quot;$x_2$&quot;</span>, ha=<span style="color: #CD5555">&quot;center&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.text(<span style="color: #B452CD">1</span>, <span style="color: #B452CD">0.9</span>, <span style="color: #CD5555">&quot;$x_3$&quot;</span>, ha=<span style="color: #CD5555">&quot;center&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.axis([-<span style="color: #B452CD">4.5</span>, <span style="color: #B452CD">4.5</span>, -<span style="color: #B452CD">0.1</span>, <span style="color: #B452CD">1.1</span>])
-
-plt.subplot(<span style="color: #B452CD">122</span>)
-plt.grid(<span style="color: #8B008B; font-weight: bold">True</span>, which=<span style="color: #CD5555">&#39;both&#39;</span>)
-plt.axhline(y=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.axvline(x=<span style="color: #B452CD">0</span>, color=<span style="color: #CD5555">&#39;k&#39;</span>)
-plt.plot(XK[:, <span style="color: #B452CD">0</span>][yk==<span style="color: #B452CD">0</span>], XK[:, <span style="color: #B452CD">1</span>][yk==<span style="color: #B452CD">0</span>], <span style="color: #CD5555">&quot;bs&quot;</span>)
-plt.plot(XK[:, <span style="color: #B452CD">0</span>][yk==<span style="color: #B452CD">1</span>], XK[:, <span style="color: #B452CD">1</span>][yk==<span style="color: #B452CD">1</span>], <span style="color: #CD5555">&quot;g^&quot;</span>)
-plt.xlabel(<span style="color: #CD5555">r&quot;$x_2$&quot;</span>, fontsize=<span style="color: #B452CD">20</span>)
-plt.ylabel(<span style="color: #CD5555">r&quot;$x_3$  &quot;</span>, fontsize=<span style="color: #B452CD">20</span>, rotation=<span style="color: #B452CD">0</span>)
-plt.annotate(<span style="color: #CD5555">r&#39;$\phi\left(\mathbf{x}\right)$&#39;</span>,
-             xy=(XK[<span style="color: #B452CD">3</span>, <span style="color: #B452CD">0</span>], XK[<span style="color: #B452CD">3</span>, <span style="color: #B452CD">1</span>]),
-             xytext=(<span style="color: #B452CD">0.65</span>, <span style="color: #B452CD">0.50</span>),
-             ha=<span style="color: #CD5555">&quot;center&quot;</span>,
-             arrowprops=<span style="color: #658b00">dict</span>(facecolor=<span style="color: #CD5555">&#39;black&#39;</span>, shrink=<span style="color: #B452CD">0.1</span>),
-             fontsize=<span style="color: #B452CD">18</span>,
-            )
-plt.plot([-<span style="color: #B452CD">0.1</span>, <span style="color: #B452CD">1.1</span>], [<span style="color: #B452CD">0.57</span>, -<span style="color: #B452CD">0.1</span>], <span style="color: #CD5555">&quot;r--&quot;</span>, linewidth=<span style="color: #B452CD">3</span>)
-plt.axis([-<span style="color: #B452CD">0.1</span>, <span style="color: #B452CD">1.1</span>, -<span style="color: #B452CD">0.1</span>, <span style="color: #B452CD">1.1</span>])
-    
-plt.subplots_adjust(right=<span style="color: #B452CD">1</span>)
-
-plt.show()
-
-
-x1_example = X1D[<span style="color: #B452CD">3</span>, <span style="color: #B452CD">0</span>]
-<span style="color: #8B008B; font-weight: bold">for</span> landmark <span style="color: #8B008B">in</span> (-<span style="color: #B452CD">2</span>, <span style="color: #B452CD">1</span>):
-    k = gaussian_rbf(np.array([[x1_example]]), np.array([[landmark]]), gamma)
-    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Phi({}, {}) = {}&quot;</span>.format(x1_example, landmark, k))
-
-rbf_kernel_svm_clf = Pipeline([
-        (<span style="color: #CD5555">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #CD5555">&quot;svm_clf&quot;</span>, SVC(kernel=<span style="color: #CD5555">&quot;rbf&quot;</span>, gamma=<span style="color: #B452CD">5</span>, C=<span style="color: #B452CD">0.001</span>))
-    ])
-rbf_kernel_svm_clf.fit(X, y)
-
-
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
-
-gamma1, gamma2 = <span style="color: #B452CD">0.1</span>, <span style="color: #B452CD">5</span>
-C1, C2 = <span style="color: #B452CD">0.001</span>, <span style="color: #B452CD">1000</span>
-hyperparams = (gamma1, C1), (gamma1, C2), (gamma2, C1), (gamma2, C2)
-
-svm_clfs = []
-<span style="color: #8B008B; font-weight: bold">for</span> gamma, C <span style="color: #8B008B">in</span> hyperparams:
-    rbf_kernel_svm_clf = Pipeline([
-            (<span style="color: #CD5555">&quot;scaler&quot;</span>, StandardScaler()),
-            (<span style="color: #CD5555">&quot;svm_clf&quot;</span>, SVC(kernel=<span style="color: #CD5555">&quot;rbf&quot;</span>, gamma=gamma, C=C))
-        ])
-    rbf_kernel_svm_clf.fit(X, y)
-    svm_clfs.append(rbf_kernel_svm_clf)
-
-plt.figure(figsize=(<span style="color: #B452CD">11</span>, <span style="color: #B452CD">7</span>))
-
-<span style="color: #8B008B; font-weight: bold">for</span> i, svm_clf <span style="color: #8B008B">in</span> <span style="color: #658b00">enumerate</span>(svm_clfs):
-    plt.subplot(<span style="color: #B452CD">221</span> + i)
-    plot_predictions(svm_clf, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-    plot_dataset(X, y, [-<span style="color: #B452CD">1.5</span>, <span style="color: #B452CD">2.5</span>, -<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1.5</span>])
-    gamma, C = hyperparams[i]
-    plt.title(<span style="color: #CD5555">r&quot;$\gamma = {}, C = {}$&quot;</span>.format(gamma, C), fontsize=<span style="color: #B452CD">16</span>)
-
+    polydegree[degree] = degree
+    error[degree] = np.mean( np.mean((y_test - y_pred)**<span style="color: #B452CD">2</span>, axis=<span style="color: #B452CD">1</span>, keepdims=<span style="color: #8B008B; font-weight: bold">True</span>) )
+    bias[degree] = np.mean( (y_test - np.mean(y_pred, axis=<span style="color: #B452CD">1</span>, keepdims=<span style="color: #8B008B; font-weight: bold">True</span>))**<span style="color: #B452CD">2</span> )
+    variance[degree] = np.mean( np.var(y_pred, axis=<span style="color: #B452CD">1</span>, keepdims=<span style="color: #8B008B; font-weight: bold">True</span>) )
+    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;Polynomial degree:&#39;</span>, degree)
+    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;Error:&#39;</span>, error[degree])
+    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;Bias^2:&#39;</span>, bias[degree])
+    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;Var:&#39;</span>, variance[degree])
+    <span style="color: #658b00">print</span>(<span style="color: #CD5555">&#39;{} &gt;= {} + {} = {}&#39;</span>.format(error[degree], bias[degree], variance[degree], bias[degree]+variance[degree]))
+ 
+mse_simpletree= np.mean( np.mean((y_test - simpleprediction)**<span style="color: #B452CD">2</span>))
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Simple tree:&quot;</span>,mse_simpletree)
+plt.xlim(<span style="color: #B452CD">1</span>,maxdepth)
+plt.plot(polydegree, error, label=<span style="color: #CD5555">&#39;MSE&#39;</span>)
+plt.plot(polydegree, bias, label=<span style="color: #CD5555">&#39;bias&#39;</span>)
+plt.plot(polydegree, variance, label=<span style="color: #CD5555">&#39;Variance&#39;</span>)
+plt.legend()
+save_fig(<span style="color: #CD5555">&quot;baggingboot&quot;</span>)
 plt.show()
 </pre>
 </div>
@@ -1253,40 +2117,494 @@ plt.show()
 
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="mathematical-optimization-of-convex-functions">Mathematical optimization of convex functions </h2>
+<h2 id="random-forests">Random forests </h2>
 
-<p>A mathematical (quadratic) optimization problem, or just optimization problem, has the form</p>
-$$
-\begin{align*}
-    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\boldsymbol{\lambda}^T\boldsymbol{P}\boldsymbol{\lambda}+\boldsymbol{q}^T\boldsymbol{\lambda},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{\lambda} \preceq \boldsymbol{h} \wedge  \boldsymbol{A}\boldsymbol{\lambda}=f.
-\end{align*}
-$$
-
-<p>subject to some constraints for say a selected set \( i=1,2,\dots, n \).
-In our case we are optimizing with respect to the Lagrangian multipliers \( \lambda_i \), and the
-vector \( \boldsymbol{\lambda}=[\lambda_1, \lambda_2,\dots, \lambda_n] \) is the optimization variable we are dealing with.
+<p>Random forests provide an improvement over bagged trees by way of a
+small tweak that decorrelates the trees. 
 </p>
 
-<p>In our case we are particularly interested in a class of optimization problems called convex optmization problems. 
-In our discussion on gradient descent methods we discussed at length the definition of a convex function. 
+<p>As in bagging, we build a
+number of decision trees on bootstrapped training samples. But when
+building these decision trees, each time a split in a tree is
+considered, a random sample of \( m \) predictors is chosen as split
+candidates from the full set of \( p \) predictors. The split is allowed to
+use only one of those \( m \) predictors. 
 </p>
 
-<p>Convex optimization problems play a central role in applied mathematics and we recommend strongly <a href="http://web.stanford.edu/~boyd/cvxbook/" target="_blank">Boyd and Vandenberghe's text on the topics</a>.</p>
+<p>A fresh sample of \( m \) predictors is
+taken at each split, and typically we choose 
+</p>
+
+$$
+m\approx \sqrt{p}.
+$$
+
+<p>In building a random forest, at
+each split in the tree, the algorithm is not even allowed to consider
+a majority of the available predictors. 
+</p>
+
+<p>The reason for this is rather clever. Suppose that there is one very
+strong predictor in the data set, along with a number of other
+moderately strong predictors. Then in the collection of bagged
+variable importance random forest trees, most or all of the trees will
+use this strong predictor in the top split. Consequently, all of the
+bagged trees will look quite similar to each other. Hence the
+predictions from the bagged trees will be highly correlated.
+Unfortunately, averaging many highly correlated quantities does not
+lead to as large of a reduction in variance as averaging many
+uncorrelated quantities. In particular, this means that bagging will
+not lead to a substantial reduction in variance over a single tree in
+this setting.
+</p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="how-do-we-solve-these-problems">How do we solve these problems? </h2>
+<h2 id="random-forest-algorithm">Random Forest Algorithm </h2>
+<p>The algorithm described here can be applied to both classification and regression problems.</p>
 
-<p>If we use Python as programming language and wish to venture beyond
-<b>scikit-learn</b>, <b>tensorflow</b> and similar software which makes our
-lives so much easier, we need to dive into the wonderful world of
-quadratic programming. We can, if we wish, solve the minimization
-problem using say standard gradient methods or conjugate gradient
-methods. However, these methods tend to exhibit a rather slow
-converge. So, welcome to the promised land of quadratic programming.
+<p>We will grow of forest of say \( B \) trees.</p>
+<ol>
+<li> For \( b=1:B \)</li>
+<ul>
+  <li> Draw a bootstrap sample from the training data organized in our \( \boldsymbol{X} \) matrix.</li>
+  <li> We grow then a random forest tree \( T_b \) based on the bootstrapped data by repeating the steps outlined till we reach the maximum node size is reached</li>
+<ol>
+   <li> we select \( m \le p \) variables at random from the \( p \) predictors/features</li>
+   <li> pick the best split point among the \( m \) features using for example the CART algorithm and create a new node</li>
+   <li> split the node into daughter nodes</li>
+</ol>
+</ul>
+<li> Output then the ensemble of trees \( \{T_b\}_1^{B} \) and make predictions for either a regression type of problem or a classification type of problem.</li> 
+</ol>
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="random-forests-compared-with-other-methods-on-the-cancer-data">Random Forests Compared with other Methods on the Cancer Data </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span>  train_test_split 
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> load_breast_cancer
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.svm</span> <span style="color: #8B008B; font-weight: bold">import</span> SVC
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> LogisticRegression
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.tree</span> <span style="color: #8B008B; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> BaggingClassifier
+
+<span style="color: #228B22"># Load the data</span>
+cancer = load_breast_cancer()
+
+X_train, X_test, y_train, y_test = train_test_split(cancer.data,cancer.target,random_state=<span style="color: #B452CD">0</span>)
+<span style="color: #658b00">print</span>(X_train.shape)
+<span style="color: #658b00">print</span>(X_test.shape)
+<span style="color: #228B22">#define methods</span>
+<span style="color: #228B22"># Logistic Regression</span>
+logreg = LogisticRegression(solver=<span style="color: #CD5555">&#39;lbfgs&#39;</span>)
+<span style="color: #228B22"># Support vector machine</span>
+svm = SVC(gamma=<span style="color: #CD5555">&#39;auto&#39;</span>, C=<span style="color: #B452CD">100</span>)
+<span style="color: #228B22"># Decision Trees</span>
+deep_tree_clf = DecisionTreeClassifier(max_depth=<span style="color: #8B008B; font-weight: bold">None</span>)
+<span style="color: #228B22">#Scale the data</span>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler
+scaler = StandardScaler()
+scaler.fit(X_train)
+X_train_scaled = scaler.transform(X_train)
+X_test_scaled = scaler.transform(X_test)
+<span style="color: #228B22"># Logistic Regression</span>
+logreg.fit(X_train_scaled, y_train)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Test set accuracy Logistic Regression with scaled data: {:.2f}&quot;</span>.format(logreg.score(X_test_scaled,y_test)))
+<span style="color: #228B22"># Support Vector Machine</span>
+svm.fit(X_train_scaled, y_train)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Test set accuracy SVM with scaled data: {:.2f}&quot;</span>.format(logreg.score(X_test_scaled,y_test)))
+<span style="color: #228B22"># Decision Trees</span>
+deep_tree_clf.fit(X_train_scaled, y_train)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Test set accuracy with Decision Trees and scaled data: {:.2f}&quot;</span>.format(deep_tree_clf.score(X_test_scaled,y_test)))
+
+
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> RandomForestClassifier
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> LabelEncoder
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> cross_validate
+<span style="color: #228B22"># Data set not specificied</span>
+<span style="color: #228B22">#Instantiate the model with 500 trees and entropy as splitting criteria</span>
+Random_Forest_model = RandomForestClassifier(n_estimators=<span style="color: #B452CD">500</span>,criterion=<span style="color: #CD5555">&quot;entropy&quot;</span>)
+Random_Forest_model.fit(X_train_scaled, y_train)
+<span style="color: #228B22">#Cross validation</span>
+accuracy = cross_validate(Random_Forest_model,X_test_scaled,y_test,cv=<span style="color: #B452CD">10</span>)[<span style="color: #CD5555">&#39;test_score&#39;</span>]
+<span style="color: #658b00">print</span>(accuracy)
+<span style="color: #658b00">print</span>(<span style="color: #CD5555">&quot;Test set accuracy with Random Forests and scaled data: {:.2f}&quot;</span>.format(Random_Forest_model.score(X_test_scaled,y_test)))
+
+
+<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">scikitplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">skplt</span>
+y_pred = Random_Forest_model.predict(X_test_scaled)
+skplt.metrics.plot_confusion_matrix(y_test, y_pred, normalize=<span style="color: #8B008B; font-weight: bold">True</span>)
+plt.show()
+y_probas = Random_Forest_model.predict_proba(X_test_scaled)
+skplt.metrics.plot_roc(y_test, y_probas)
+plt.show()
+skplt.metrics.plot_cumulative_gain(y_test, y_probas)
+plt.show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+<p>Recall that the cumulative gains curve shows the percentage of the
+overall number of cases in a given category <em>gained</em> by targeting a
+percentage of the total number of cases.
 </p>
 
-<p>The functions we need are contained in the <a href="https://cvxopt.org/userguide/coneprog.html" target="_blank">quadratic programming library CVXOPT</a> and we need to import it together with <b>numpy</b> as</p>
+<p>Similarly, the receiver operating characteristic curve, or ROC curve,
+displays the diagnostic ability of a binary classifier system as its
+discrimination threshold is varied. It plots the true positive rate against the false positive rate.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="compare-bagging-on-trees-with-random-forests">Compare  Bagging on Trees with Random Forests  </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;">bag_clf = BaggingClassifier(
+    DecisionTreeClassifier(splitter=<span style="color: #CD5555">&quot;random&quot;</span>, max_leaf_nodes=<span style="color: #B452CD">16</span>, random_state=<span style="color: #B452CD">42</span>),
+    n_estimators=<span style="color: #B452CD">500</span>, max_samples=<span style="color: #B452CD">1.0</span>, bootstrap=<span style="color: #8B008B; font-weight: bold">True</span>, n_jobs=-<span style="color: #B452CD">1</span>, random_state=<span style="color: #B452CD">42</span>)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #eeeedd">
+  <pre style="line-height: 125%;">bag_clf.fit(X_train, y_train)
+y_pred = bag_clf.predict(X_test)
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> RandomForestClassifier
+rnd_clf = RandomForestClassifier(n_estimators=<span style="color: #B452CD">500</span>, max_leaf_nodes=<span style="color: #B452CD">16</span>, n_jobs=-<span style="color: #B452CD">1</span>, random_state=<span style="color: #B452CD">42</span>)
+rnd_clf.fit(X_train, y_train)
+y_pred_rf = rnd_clf.predict(X_test)
+np.sum(y_pred == y_pred_rf) / <span style="color: #658b00">len</span>(y_pred) 
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="boosting-a-bird-s-eye-view">Boosting, a Bird's Eye View </h2>
+
+<p>The basic idea is to combine weak classifiers in order to create a good
+classifier. With a weak classifier we often intend a classifier which
+produces results which are only slightly better than we would get by
+random guesses.
+</p>
+
+<p>This is done by applying in an iterative way a weak (or a standard
+classifier like decision trees) to modify the data. In each iteration
+we emphasize those observations which are misclassified by weighting
+them with a factor.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="what-is-boosting-additive-modelling-iterative-fitting">What is boosting? Additive Modelling/Iterative Fitting </h2>
+
+<p>Boosting is a way of fitting an additive expansion in a set of
+elementary basis functions like for example some simple polynomials.
+Assume for example that we have a function
+</p>
+$$
+f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m),
+$$
+
+<p>where \( \beta_m \) are the expansion parameters to be determined in a
+minimization process and \( b(x;\gamma_m) \) are some simple functions of
+the multivariable parameter \( x \) which is characterized by the
+parameters \( \gamma_m \).
+</p>
+
+<p>As an example, consider the Sigmoid function we used in logistic
+regression. In that case, we can translate the function
+\( b(x;\gamma_m) \) into the Sigmoid function
+</p>
+
+$$
+\sigma(t) = \frac{1}{1+\exp{(-t)}},
+$$
+
+<p>where \( t=\gamma_0+\gamma_1 x \) and the parameters \( \gamma_0 \) and
+\( \gamma_1 \) were determined by the Logistic Regression fitting
+algorithm.
+</p>
+
+<p>As another example, consider the cost function we defined for linear regression</p>
+$$
+C(\boldsymbol{y},\boldsymbol{f}) = \frac{1}{n} \sum_{i=0}^{n-1}(y_i-f(x_i))^2.
+$$
+
+<p>In this case the function \( f(x) \) was replaced by the design matrix
+\( \boldsymbol{X} \) and the unknown linear regression parameters \( \boldsymbol{\beta} \),
+that is \( \boldsymbol{f}=\boldsymbol{X}\boldsymbol{\beta} \). In linear regression we can 
+simply invert a matrix and obtain the parameters \( \beta \) by
+</p>
+
+$$
+\boldsymbol{\beta}=\left(\boldsymbol{X}^T\boldsymbol{X}\right)^{-1}\boldsymbol{X}^T\boldsymbol{y}.
+$$
+
+<p>In iterative fitting or additive modeling, we minimize the cost function with respect to the parameters \( \beta_m \) and \( \gamma_m \).</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="iterative-fitting-regression-and-squared-error-cost-function">Iterative Fitting, Regression and Squared-error Cost Function </h2>
+
+<p>The way we proceed is as follows (here we specialize to the squared-error cost function)</p>
+
+<ol>
+<li> Establish a cost function, here \( C(\boldsymbol{y},\boldsymbol{f}) = \frac{1}{n} \sum_{i=0}^{n-1}(y_i-f_M(x_i))^2 \) with \( f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m) \).</li>
+<li> Initialize with a guess \( f_0(x) \). It could be one or even zero or some random numbers.</li>
+<li> For \( m=1:M \)
+<ol type="a"></li>
+ <li> minimize \( \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2 \) wrt \( \gamma \) and \( \beta \)</li>
+ <li> This gives the optimal values \( \beta_m \) and \( \gamma_m \)</li>
+ <li> Determine then the new values \( f_m(x)=f_{m-1}(x) +\beta_m b(x;\gamma_m) \)</li>
+</ol>
+</ol>
+<p>We could use any of the algorithms we have discussed till now. If we
+use trees, \( \gamma \) parameterizes the split variables and split points
+at the internal nodes, and the predictions at the terminal nodes.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="squared-error-example-and-iterative-fitting">Squared-Error Example and Iterative Fitting </h2>
+
+<p>To better understand what happens, let us develop the steps for the iterative fitting using the above squared error function.</p>
+
+<p>For simplicity we assume also that our functions \( b(x;\gamma)=1+\gamma x \). </p>
+
+<p>This means that for every iteration \( m \), we need to optimize</p>
+
+$$
+(\beta_m,\gamma_m) = \mathrm{argmin}_{\beta,\lambda}\hspace{0.1cm} \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2=\sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta(1+\gamma x_i))^2.
+$$
+
+<p>We start our iteration by simply setting \( f_0(x)=0 \). 
+Taking the derivatives  with respect to \( \beta \) and \( \gamma \) we obtain
+</p>
+$$
+\frac{\partial {\cal C}}{\partial \beta} = -2\sum_{i}(1+\gamma x_i)(y_i-\beta(1+\gamma x_i))=0,
+$$
+
+<p>and</p>
+$$
+\frac{\partial {\cal C}}{\partial \gamma} =-2\sum_{i}\beta x_i(y_i-\beta(1+\gamma x_i))=0.
+$$
+
+<p>We can then rewrite these equations as (defining \( \boldsymbol{w}=\boldsymbol{e}+\gamma \boldsymbol{x}) \) with \( \boldsymbol{e} \) being the unit vector)</p>
+$$
+\gamma \boldsymbol{w}^T(\boldsymbol{y}-\beta\gamma \boldsymbol{w})=0,
+$$
+
+<p>which gives us \( \beta = \boldsymbol{w}^T\boldsymbol{y}/(\boldsymbol{w}^T\boldsymbol{w}) \). Similarly we have </p>
+$$
+\beta\gamma \boldsymbol{x}^T(\boldsymbol{y}-\beta(1+\gamma \boldsymbol{x}))=0,
+$$
+
+<p>which leads to \( \gamma =(\boldsymbol{x}^T\boldsymbol{y}-\beta\boldsymbol{x}^T\boldsymbol{e})/(\beta\boldsymbol{x}^T\boldsymbol{x}) \).  Inserting
+for \( \beta \) gives us an equation for \( \gamma \). This is a non-linear equation in the unknown \( \gamma \) and has to be solved numerically. 
+</p>
+
+<p>The solution to these two equations gives us in turn \( \beta_1 \) and \( \gamma_1 \) leading to the new expression for \( f_1(x) \) as
+\( f_1(x) = \beta_1(1+\gamma_1x) \). Doing this \( M \) times results in our final estimate for the function \( f \). 
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="iterative-fitting-classification-and-adaboost">Iterative Fitting, Classification and AdaBoost </h2>
+
+<p>Let us consider a binary classification problem with two outcomes \( y_i \in \{-1,1\} \) and \( i=0,1,2,\dots,n-1 \) as our set of
+observations. We define a classification function \( G(x) \) which produces a prediction taking one or the other of the two values 
+\( \{-1,1\} \).
+</p>
+
+<p>The error rate of the training sample is then</p>
+
+$$
+\mathrm{\overline{err}}=\frac{1}{n} \sum_{i=0}^{n-1} I(y_i\ne G(x_i)). 
+$$
+
+<p>The iterative procedure starts with defining a weak classifier whose
+error rate is barely better than random guessing.  The iterative
+procedure in boosting is to sequentially apply a  weak
+classification algorithm to repeatedly modified versions of the data
+producing a sequence of weak classifiers \( G_m(x) \).
+</p>
+
+<p>Here we will express our  function \( f(x) \) in terms of \( G(x) \). That is</p>
+$$
+f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m),
+$$
+
+<p>will be a function of </p>
+$$
+G_M(x) = \mathrm{sign} \sum_{i=1}^M \alpha_m G_m(x).
+$$
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="adaptive-boosting-adaboost">Adaptive Boosting, AdaBoost </h2>
+
+<p>In our iterative procedure we define thus</p>
+$$
+f_m(x) = f_{m-1}(x)+\beta_mG_m(x).
+$$
+
+<p>The simplest possible cost function which leads (also simple from a computational point of view) to the AdaBoost algorithm is the
+exponential cost/loss function defined as
+</p>
+$$
+C(\boldsymbol{y},\boldsymbol{f}) = \sum_{i=0}^{n-1}\exp{(-y_i(f_{m-1}(x_i)+\beta G(x_i))}.
+$$
+
+<p>We optimize \( \beta \) and \( G \) for each value of \( m=1:M \) as we did in the regression case.
+This is normally done in two steps. Let us however first rewrite the cost function as 
+</p>
+
+$$
+C(\boldsymbol{y},\boldsymbol{f}) = \sum_{i=0}^{n-1}w_i^{m}\exp{(-y_i\beta G(x_i))},
+$$
+
+<p>where we have defined \( w_i^m= \exp{(-y_if_{m-1}(x_i))} \).</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="building-up-adaboost">Building up AdaBoost </h2>
+
+<p>First, for any \( \beta > 0 \), we optimize \( G \) by setting</p>
+$$
+G_m(x) = \mathrm{sign} \sum_{i=0}^{n-1} w_i^m I(y_i \ne G_(x_i)),
+$$
+
+<p>which is the classifier that minimizes the weighted error rate in predicting \( y \).</p>
+
+<p>We can do this by rewriting</p>
+$$
+\exp{-(\beta)}\sum_{y_i=G(x_i)}w_i^m+\exp{(\beta)}\sum_{y_i\ne G(x_i)}w_i^m,
+$$
+
+<p>which can be rewritten as</p>
+$$
+(\exp{(\beta)}-\exp{-(\beta)})\sum_{i=0}^{n-1}w_i^mI(y_i\ne G(x_i))+\exp{(-\beta)}\sum_{i=0}^{n-1}w_i^m=0,
+$$
+
+<p>which leads to</p>
+$$
+\beta_m = \frac{1}{2}\log{\frac{1-\mathrm{\overline{err}}}{\mathrm{\overline{err}}}},
+$$
+
+<p>where we have redefined the error as </p>
+$$
+\mathrm{\overline{err}}_m=\frac{1}{n}\frac{\sum_{i=0}^{n-1}w_i^mI(y_i\ne G(x_i)}{\sum_{i=0}^{n-1}w_i^m},
+$$
+
+<p>which leads to an update of</p>
+$$
+f_m(x) = f_{m-1}(x) +\beta_m G_m(x).
+$$
+
+<p>This leads to the new weights</p>
+$$
+w_i^{m+1} = w_i^m \exp{(-y_i\beta_m G_m(x_i))}
+$$
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="adaptive-boosting-adaboost-basic-algorithm">Adaptive boosting: AdaBoost, Basic Algorithm </h2>
+
+<p>The algorithm here is rather straightforward. Assume that our weak
+classifier is a decision tree and we consider a binary set of outputs
+with \( y_i \in \{-1,1\} \) and \( i=0,1,2,\dots,n-1 \) as our set of
+observations. Our design matrix is given in terms of the
+feature/predictor vectors
+\( \boldsymbol{X}=[\boldsymbol{x}_0\boldsymbol{x}_1\dots\boldsymbol{x}_{p-1}] \). Finally, we define also a
+classifier determined by our data via a function \( G(x) \). This function tells us how well we are able to classify our outputs/targets \( \boldsymbol{y} \). 
+</p>
+
+<p>We have already defined the misclassification error \( \mathrm{err} \) as</p>
+$$
+\mathrm{err}=\frac{1}{n}\sum_{i=0}^{n-1}I(y_i\ne G(x_i)),
+$$
+
+<p>where the function \( I() \) is one if we misclassify and zero if we classify correctly. </p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="basic-steps-of-adaboost">Basic Steps of AdaBoost </h2>
+
+<p>With the above definitions we are now ready to set up the algorithm for AdaBoost.
+The basic idea is to set up weights which will be used to scale the correctly classified and the misclassified cases.
+</p>
+<ol>
+<li> We start by initializing all weights to \( w_i = 1/n \), with \( i=0,1,2,\dots n-1 \). It is easy to see that we must have \( \sum_{i=0}^{n-1}w_i = 1 \).</li>
+<li> We rewrite the misclassification error as</li> 
+</ol>
+$$
+\mathrm{\overline{err}}_m=\frac{\sum_{i=0}^{n-1}w_i^m I(y_i\ne G(x_i))}{\sum_{i=0}^{n-1}w_i},
+$$
+
+<ol>
+<li> Then we start looping over all attempts at classifying, namely we start an iterative process for \( m=1:M \), where \( M \) is the final number of classifications. Our given classifier could for example be a plain decision tree.
+<ol type="a"></li>
+ <li> Fit then a given classifier to the training set using the weights \( w_i \).</li>
+ <li> Compute then \( \mathrm{err} \) and figure out which events are classified properly and which are classified wrongly.</li>
+ <li> Define a quantity \( \alpha_{m} = \log{(1-\mathrm{\overline{err}}_m)/\mathrm{\overline{err}}_m} \)</li>
+ <li> Set the new weights to \( w_i = w_i\times \exp{(\alpha_m I(y_i\ne G(x_i)} \).</li>
+</ol>
+<li> Compute the new classifier \( G(x)= \sum_{i=0}^{n-1}\alpha_m I(y_i\ne G(x_i) \).</li>
+</ol>
+<p>For the iterations with \( m \le 2 \) the weights are modified
+individually at each steps. The observations which were misclassified
+at iteration \( m-1 \) have a weight which is larger than those which were
+classified properly. As this proceeds, the observations which were
+difficult to classifiy correctly are given a larger influence. Each
+new classification step \( m \) is then forced to concentrate on those
+observations that are missed in the previous iterations.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="adaboost-examples">AdaBoost Examples </h2>
+
+<p>Using <b>Scikit-Learn</b> it is easy to apply the adaptive boosting algorithm, as done here.</p>
 
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
@@ -1295,94 +2613,27 @@ converge. So, welcome to the promised land of quadratic programming.
     <div class="inner_cell">
       <div class="input_area">
         <div class="highlight" style="background: #eeeedd">
-  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span>
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">cvxopt</span>
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
+  <pre style="line-height: 125%;"><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> AdaBoostClassifier
 
-<p>This will make our life much easier. You don't need t0 write your own optimizer.</p>
+ada_clf = AdaBoostClassifier(
+    DecisionTreeClassifier(max_depth=<span style="color: #B452CD">1</span>), n_estimators=<span style="color: #B452CD">200</span>,
+    algorithm=<span style="color: #CD5555">&quot;SAMME.R&quot;</span>, learning_rate=<span style="color: #B452CD">0.5</span>, random_state=<span style="color: #B452CD">42</span>)
+ada_clf.fit(X_train, y_train)
 
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="a-simple-example">A simple example </h2>
+<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.ensemble</span> <span style="color: #8B008B; font-weight: bold">import</span> AdaBoostClassifier
 
-<p>We remind ourselves about the general problem we want to solve</p>
-$$
-\begin{align*}
-    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}\boldsymbol{x}^T\boldsymbol{P}\boldsymbol{x}+\boldsymbol{q}^T\boldsymbol{x},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm} to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{x} \preceq \boldsymbol{h} \wedge  \boldsymbol{A}\boldsymbol{x}=f.
-\end{align*}
-$$
-
-<p>Let us show how to perform the optmization using a simple case. Assume we want to optimize the following problem</p>
-$$
-\begin{align*}
-    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}x^2+5x+3y \\ \nonumber
-    &\mathrm{subject}\hspace{0.5cm} \mathrm{to} \\ \nonumber
-    &x, y \geq 0 \\ \nonumber
-    &x+3y  \geq 15 \\ \nonumber
-    &2x+5y  \leq  100 \\ \nonumber
-    &3x+4y  \leq  80.  \\ \nonumber
-\end{align*}
-$$
-
-<p>The minimization problem can be rewritten in terms of vectors and matrices as (with \( x \) and \( y \) being the unknowns)</p>
-$$
-\frac{1}{2}\begin{bmatrix} x\\ y \end{bmatrix}^T   \begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}  \begin{bmatrix} x \\ y \end{bmatrix}  + \begin{bmatrix}3\\ 4  \end{bmatrix}^T \begin{bmatrix}x \\ y  \end{bmatrix}.  
-$$
-
-<p>Similarly, we can now set up the inequalities (we need to change \( \geq \) to \( \leq \) by multiplying with \( -1 \) on bot sides) as the following matrix-vector equation</p>
-$$
-\begin{bmatrix} -1 & 0 \\ 0 & -1 \\ -1 & -3 \\ 2 & 5 \\ 3 & 4\end{bmatrix}\begin{bmatrix} x \\ y\end{bmatrix} \preceq \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
-$$
-
-<p>We have collapsed all the inequalities into a single matrix \( \boldsymbol{G} \). We see also that our matrix </p>
-$$
-\boldsymbol{P} =\begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}
-$$
-
-<p>is clearly positive semi-definite (all eigenvalues larger or equal zero). 
-Finally, the vector \( \boldsymbol{h} \) is defined as 
-</p>
-$$
-\boldsymbol{h} = \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
-$$
-
-<p>Since we don't have any equalities the matrix \( \boldsymbol{A} \) is set to zero
-The following code solves the equations for us
-</p>
-
-<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #eeeedd">
-  <pre style="line-height: 125%;"><span style="color: #228B22"># Import the necessary packages</span>
-<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span>
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">cvxopt</span> <span style="color: #8B008B; font-weight: bold">import</span> matrix
-<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">cvxopt</span> <span style="color: #8B008B; font-weight: bold">import</span> solvers
-P = matrix(numpy.diag([<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>]), tc=<span style="color: #CD5555">&#39;d&#39;</span>)
-q = matrix(numpy.array([<span style="color: #B452CD">3</span>,<span style="color: #B452CD">4</span>]), tc=<span style="color: #CD5555">&#39;d&#39;</span>)
-G = matrix(numpy.array([[-<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>],[<span style="color: #B452CD">0</span>,-<span style="color: #B452CD">1</span>],[-<span style="color: #B452CD">1</span>,-<span style="color: #B452CD">3</span>],[<span style="color: #B452CD">2</span>,<span style="color: #B452CD">5</span>],[<span style="color: #B452CD">3</span>,<span style="color: #B452CD">4</span>]]), tc=<span style="color: #CD5555">&#39;d&#39;</span>)
-h = matrix(numpy.array([<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>,-<span style="color: #B452CD">15</span>,<span style="color: #B452CD">100</span>,<span style="color: #B452CD">80</span>]), tc=<span style="color: #CD5555">&#39;d&#39;</span>)
-<span style="color: #228B22"># Construct the QP, invoke solver</span>
-sol = solvers.qp(P,q,G,h)
-<span style="color: #228B22"># Extract optimal value and solution</span>
-sol[<span style="color: #CD5555">&#39;x&#39;</span>] 
-sol[<span style="color: #CD5555">&#39;primal objective&#39;</span>]
-<span style="color: #658b00">print</span>(sol[<span style="color: #CD5555">&#39;x&#39;</span>] )
+ada_clf = AdaBoostClassifier(
+    DecisionTreeClassifier(max_depth=<span style="color: #B452CD">1</span>), n_estimators=<span style="color: #B452CD">200</span>,
+    algorithm=<span style="color: #CD5555">&quot;SAMME.R&quot;</span>, learning_rate=<span style="color: #B452CD">0.5</span>, random_state=<span style="color: #B452CD">42</span>)
+ada_clf.fit(X_train_scaled, y_train)
+y_pred = ada_clf.predict(X_test_scaled)
+skplt.metrics.plot_confusion_matrix(y_test, y_pred, normalize=<span style="color: #8B008B; font-weight: bold">True</span>)
+plt.show()
+y_probas = ada_clf.predict_proba(X_test_scaled)
+skplt.metrics.plot_roc(y_test, y_probas)
+plt.show()
+skplt.metrics.plot_cumulative_gain(y_test, y_probas)
+plt.show()
 </pre>
 </div>
       </div>
@@ -1399,33 +2650,9 @@ sol[<span style="color: #CD5555">&#39;primal objective&#39;</span>]
 </div>
 
 
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="back-to-the-more-realistic-cases">Back to the more realistic cases </h2>
-
-<p>We are now ready to return to our setup of the optmization problem for a more realistic case. Introducing the <b>slack</b> parameter \( C \) we have</p>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1K(\boldsymbol{x}_1,\boldsymbol{x}_1) & y_1y_2K(\boldsymbol{x}_1,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_1,\boldsymbol{x}_n) \\
-y_2y_1K(\boldsymbol{x}_2,\boldsymbol{x}_1) & y_2y_2K(\boldsymbol{x}_2,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_2,\boldsymbol{x}_n) \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1K(\boldsymbol{x}_n,\boldsymbol{x}_1) & y_ny_2K(\boldsymbol{x}_n\boldsymbol{x}_2) & \dots & \dots & y_ny_nK(\boldsymbol{x}_n,\boldsymbol{x}_n) \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{I}\boldsymbol{\lambda}, 
-$$
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
-With  the slack constants this leads to the additional constraint \( 0\leq \lambda_i \leq C \).
-</p>
-
-<p>Using the <b>CVXOPT</b> library, the matrix \( P \) would then be defined by the above matrix while the KKT conditions would all be collected by the matrix \( G \).</p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="support-vector-machines-for-regression">Support vector machines for regression </h2>
-
-<p>Material will be added here.</p>
 <!-- ------------------- end of main content --------------- -->
 <center style="font-size:80%">
-<!-- copyright --> &copy; 1999-2022, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
+<!-- copyright --> &copy; 1999-2023, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
 </center>
 </body>
 </html>
diff --git a/doc/pub/week46/html/week46.html b/doc/pub/week46/html/week46.html
index 338f9ed7f..c55285389 100644
--- a/doc/pub/week46/html/week46.html
+++ b/doc/pub/week46/html/week46.html
@@ -8,8 +8,8 @@ doconce format html week46.do.txt --pygments_html_style=default --html_style=blo
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 <meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 46: Support Vector Machines and Project 3.">
-<title>Week 46: Support Vector Machines and Project 3.</title>
+<meta name="description" content="Week 46: Decision Trees, Ensemble methods  and Random Forests">
+<title>Week 46: Decision Trees, Ensemble methods  and Random Forests</title>
 <style type="text/css">
 /* bloodish style */
 body {
@@ -140,68 +140,190 @@ div.toc p,a {
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Overview of week 46', 2, None, 'overview-of-week-46'),
-              ('Support Vector Machines, overarching aims',
+ 'sections': [('Plan for week 46', 2, None, 'plan-for-week-46'),
+              ('Decision trees, overarching aims',
                2,
                None,
-               'support-vector-machines-overarching-aims'),
-              ('Hyperplanes and all that', 2, None, 'hyperplanes-and-all-that'),
-              ('What is a hyperplane?', 2, None, 'what-is-a-hyperplane'),
-              ('A $p$-dimensional space of features',
+               'decision-trees-overarching-aims'),
+              ('Basics of a tree', 2, None, 'basics-of-a-tree'),
+              ('A Sketch of a Tree, Regression problem',
                2,
                None,
-               'a-p-dimensional-space-of-features'),
-              ('The two-dimensional case', 2, None, 'the-two-dimensional-case'),
-              ('Getting into the details', 2, None, 'getting-into-the-details'),
-              ('First attempt at a minimization approach',
+               'a-sketch-of-a-tree-regression-problem'),
+              ('A Sketch of a Tree, Classification  problem',
                2,
                None,
-               'first-attempt-at-a-minimization-approach'),
-              ('Solving the equations', 2, None, 'solving-the-equations'),
-              ('Code Example', 2, None, 'code-example'),
-              ('Problems with the Simpler Approach',
+               'a-sketch-of-a-tree-classification-problem'),
+              ('A typical Decision Tree with its pertinent Jargon, '
+               'Classification Problem',
                2,
                None,
-               'problems-with-the-simpler-approach'),
-              ('A better approach', 2, None, 'a-better-approach'),
-              ('A quick Reminder on Lagrangian Multipliers',
+               'a-typical-decision-tree-with-its-pertinent-jargon-classification-problem'),
+              ('General Features', 2, None, 'general-features'),
+              ('How do we set it up?', 2, None, 'how-do-we-set-it-up'),
+              ('Decision trees and Regression',
                2,
                None,
-               'a-quick-reminder-on-lagrangian-multipliers'),
-              ('Adding the Multiplier', 2, None, 'adding-the-multiplier'),
-              ('Setting up the Problem', 2, None, 'setting-up-the-problem'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ('The last steps', 2, None, 'the-last-steps'),
-              ('A soft classifier', 2, None, 'a-soft-classifier'),
-              ('Soft optmization problem', 2, None, 'soft-optmization-problem'),
-              ('Kernels and non-linearity',
+               'decision-trees-and-regression'),
+              ('Building a tree, regression',
                2,
                None,
-               'kernels-and-non-linearity'),
-              ('The equations', 2, None, 'the-equations'),
-              ('The problem to solve', 2, None, 'the-problem-to-solve'),
-              ("Different kernels and Mercer's theorem",
+               'building-a-tree-regression'),
+              ('A top-down approach, recursive binary splitting',
                2,
                None,
-               'different-kernels-and-mercer-s-theorem'),
-              ('The moons example', 2, None, 'the-moons-example'),
-              ('Mathematical optimization of convex functions',
+               'a-top-down-approach-recursive-binary-splitting'),
+              ('Making a tree', 2, None, 'making-a-tree'),
+              ('Pruning the tree', 2, None, 'pruning-the-tree'),
+              ('Cost complexity pruning', 2, None, 'cost-complexity-pruning'),
+              ('Schematic Regression Procedure',
                2,
                None,
-               'mathematical-optimization-of-convex-functions'),
-              ('How do we solve these problems?',
+               'schematic-regression-procedure'),
+              ('A Classification Tree', 2, None, 'a-classification-tree'),
+              ('Growing a classification tree',
                2,
                None,
-               'how-do-we-solve-these-problems'),
-              ('A simple example', 2, None, 'a-simple-example'),
-              ('Back to the more realistic cases',
+               'growing-a-classification-tree'),
+              ('Classification tree, how to split nodes',
                2,
                None,
-               'back-to-the-more-realistic-cases'),
-              ('Support vector machines for regression',
+               'classification-tree-how-to-split-nodes'),
+              ('Visualizing the Tree, Classification',
                2,
                None,
-               'support-vector-machines-for-regression')]}
+               'visualizing-the-tree-classification'),
+              ('Visualizing the Tree, The Moons',
+               2,
+               None,
+               'visualizing-the-tree-the-moons'),
+              ('Other ways of visualizing the trees',
+               2,
+               None,
+               'other-ways-of-visualizing-the-trees'),
+              ('Printing out as text', 2, None, 'printing-out-as-text'),
+              ('Algorithms for Setting up Decision Trees',
+               2,
+               None,
+               'algorithms-for-setting-up-decision-trees'),
+              ('The CART algorithm for Classification',
+               2,
+               None,
+               'the-cart-algorithm-for-classification'),
+              ('The CART algorithm for Regression',
+               2,
+               None,
+               'the-cart-algorithm-for-regression'),
+              ('Why binary splits?', 2, None, 'why-binary-splits'),
+              ('Computing a Tree using the Gini Index',
+               2,
+               None,
+               'computing-a-tree-using-the-gini-index'),
+              ('The Table', 2, None, 'the-table'),
+              ('Computing the various Gini Indices',
+               2,
+               None,
+               'computing-the-various-gini-indices'),
+              ('Computing the various Gini Indices, Hours slept',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-slept'),
+              ('Computing the various Gini Indices, Hours studied',
+               2,
+               None,
+               'computing-the-various-gini-indices-hours-studied'),
+              ('A possible code using Scikit-Learn',
+               2,
+               None,
+               'a-possible-code-using-scikit-learn'),
+              ('Further example: Computing the Gini index',
+               2,
+               None,
+               'further-example-computing-the-gini-index'),
+              ('Simple Python Code to read in Data and perform Classification',
+               2,
+               None,
+               'simple-python-code-to-read-in-data-and-perform-classification'),
+              ('Computing the Gini Factor',
+               2,
+               None,
+               'computing-the-gini-factor'),
+              ('Regression trees', 2, None, 'regression-trees'),
+              ('Final regressor code', 2, None, 'final-regressor-code'),
+              ('Pros and cons of trees, pros',
+               2,
+               None,
+               'pros-and-cons-of-trees-pros'),
+              ('Disadvantages', 2, None, 'disadvantages'),
+              ('Ensemble Methods: From a Single Tree to Many Trees and Extreme '
+               'Boosting, Meet the Jungle of Methods',
+               2,
+               None,
+               'ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods'),
+              ('An Overview of Ensemble Methods',
+               2,
+               None,
+               'an-overview-of-ensemble-methods'),
+              ('Why Voting?', 2, None, 'why-voting'),
+              ('Tossing coins', 2, None, 'tossing-coins'),
+              ('Standard imports first', 2, None, 'standard-imports-first'),
+              ('Simple Voting Example, head or tail',
+               2,
+               None,
+               'simple-voting-example-head-or-tail'),
+              ('Using the Voting Classifier',
+               2,
+               None,
+               'using-the-voting-classifier'),
+              ('Voting and Bagging', 2, None, 'voting-and-bagging'),
+              ('Bagging', 2, None, 'bagging'),
+              ('More bagging', 2, None, 'more-bagging'),
+              ('Making your own Bootstrap: Changing the Level of the Decision '
+               'Tree',
+               2,
+               None,
+               'making-your-own-bootstrap-changing-the-level-of-the-decision-tree'),
+              ('Random forests', 2, None, 'random-forests'),
+              ('Random Forest Algorithm', 2, None, 'random-forest-algorithm'),
+              ('Random Forests Compared with other Methods on the Cancer Data',
+               2,
+               None,
+               'random-forests-compared-with-other-methods-on-the-cancer-data'),
+              ('Compare  Bagging on Trees with Random Forests',
+               2,
+               None,
+               'compare-bagging-on-trees-with-random-forests'),
+              ("Boosting, a Bird's Eye View",
+               2,
+               None,
+               'boosting-a-bird-s-eye-view'),
+              ('What is boosting? Additive Modelling/Iterative Fitting',
+               2,
+               None,
+               'what-is-boosting-additive-modelling-iterative-fitting'),
+              ('Iterative Fitting, Regression and Squared-error Cost Function',
+               2,
+               None,
+               'iterative-fitting-regression-and-squared-error-cost-function'),
+              ('Squared-Error Example and Iterative Fitting',
+               2,
+               None,
+               'squared-error-example-and-iterative-fitting'),
+              ('Iterative Fitting, Classification and AdaBoost',
+               2,
+               None,
+               'iterative-fitting-classification-and-adaboost'),
+              ('Adaptive Boosting, AdaBoost',
+               2,
+               None,
+               'adaptive-boosting-adaboost'),
+              ('Building up AdaBoost', 2, None, 'building-up-adaboost'),
+              ('Adaptive boosting: AdaBoost, Basic Algorithm',
+               2,
+               None,
+               'adaptive-boosting-adaboost-basic-algorithm'),
+              ('Basic Steps of AdaBoost', 2, None, 'basic-steps-of-adaboost'),
+              ('AdaBoost Examples', 2, None, 'adaboost-examples')]}
 end of tocinfo -->
 
 <body>
@@ -223,7 +345,7 @@ MathJax.Hub.Config({
 
 <!-- ------------------- main content ---------------------- -->
 <center>
-<h1>Week 46: Support Vector Machines and Project 3. </h1>
+<h1>Week 46: Decision Trees, Ensemble methods  and Random Forests</h1>
 </center>  <!-- document title -->
 
 <!-- author(s): Morten Hjorth-Jensen -->
@@ -239,704 +361,138 @@ MathJax.Hub.Config({
 </center>
 <br>
 <center>
-<h4>Nov 19, 2022</h4>
+<h4>Week 46, November 13-17</h4>
 </center> <!-- date -->
 <br>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="overview-of-week-46">Overview of week 46 </h2>
-
-<ul>
-<li> <b>Thursday</b>: Support Vector Machines, discussion of project 3</li>
-<ul>
-  <li> <a href="https://youtu.be/F3CkH-opbdY" target="_blank">Video of lecture</a></li>
-</ul>
-<li> <b>Friday</b>: Support vector machines and start Principal Component Analysis</li>
-<ul>
-  <li> <a href="https://youtu.be/BbupEEvMXtg" target="_blank">Video of lecture</a></li>
-</ul>
-</ul>
-<div class="alert alert-block alert-block alert-text-normal">
-<b>Reading</b>
-<p>
-<p>Reading recommendations:</p>
-<ol>
- <li> See lecture notes for week 46 at <a href="https://compphysics.github.io/MachineLearning/doc/web/course.html." target="_blank"><tt>https://compphysics.github.io/MachineLearning/doc/web/course.html.</tt></a></li>
- <li> Hastie et al chapter 12</li>
- <li> Bishop chapter 7.1 and 7.2</li>
-</ol>
-</div>
-
+<h2 id="plan-for-week-46">Plan for week 46 </h2>
 
 <div class="alert alert-block alert-block alert-text-normal">
-<b>Videos</b>
+<b>Active learning sessions on Tuesday and Wednesday</b>
 <p>
-<ol>
-<li> <a href="https://www.youtube.com/watch?v=efR1C6CvhmE&ab_channel=StatQuestwithJoshStarmer" target="_blank">Overview video on Support Vector Machines</a></li>
-<li> See also <a href="https://www.youtube.com/watch?v=N1vOgolbjSc&ab_channel=AliceZhao" target="_blank">this video</a>.</li>
-</ol>
+<ul>
+  <li> Work and Discussion of project 2</li>
+  <li> Discussion of project 3 as well</li>
+</ul>
+</div>
+  
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Material for the lecture on Thursday November 16, 2023</b>
+<p>
+<ul>
+  <li> Thursday: Basics of decision trees, classification and regression algorithms and ensemble models</li> 
+  <li> Readings and Videos:</li>
+<ul>
+    <li> These lecture notes</li>
+    <li> <a href="https://youtu.be/" target="_blank">Video of lecture to be added</a></li>
+    <li> <a href="https://www.youtube.com/watch?v=RmajweUFKvM&ab_channel=Simplilearn" target="_blank">Video on Decision trees</a></li>
+    <li> Decision Trees: Geron's chapter 6 covers decision trees while ensemble models, voting and bagging are discussed in chapter 7. See also lecture from <a href="https://www.uio.no/studier/emner/matnat/math/STK-IN4300/h20/slides/lecture_7.pdf" target="_blank">STK-IN4300, lecture 7</a>. Chapter 9.2 of Hastie et al contains also a good discussion.</li>
+</ul>
+</ul>
 </div>
 
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="support-vector-machines-overarching-aims">Support Vector Machines, overarching aims  </h2>
+<h2 id="decision-trees-overarching-aims">Decision trees, overarching aims  </h2>
 
-<p>A Support Vector Machine (SVM) is a very powerful and versatile
-Machine Learning method, capable of performing linear or nonlinear
-classification, regression, and even outlier detection. It is one of
-the most popular models in Machine Learning, and anyone interested in
-Machine Learning should have it in their toolbox. SVMs are
-particularly well suited for classification of complex but small-sized or
-medium-sized datasets.  
+<p>We start here with the most basic algorithm, the so-called decision
+tree. With this basic algorithm we can in turn build more complex
+networks, spanning from homogeneous and heterogenous forests (bagging,
+random forests and more) to one of the most popular supervised
+algorithms nowadays, the extreme gradient boosting, or just
+XGBoost. But let us start with the simplest possible ingredient.
 </p>
 
-<p>The case with two well-separated classes only can be understood in an
-intuitive way in terms of lines in a two-dimensional space separating
-the two classes (see figure below).
+<p>Decision trees are supervised learning algorithms used for both,
+classification and regression tasks.
 </p>
 
-<p>The basic mathematics behind the SVM is however less familiar to most of us. 
-It relies on the definition of hyperplanes and the
-definition of a <b>margin</b> which separates classes (in case of
-classification problems) of variables. It is also used for regression
-problems.
+<p>The main idea of decision trees
+is to find those descriptive features which contain the most
+<b>information</b> regarding the target feature and then split the dataset
+along the values of these features such that the target feature values
+for the resulting underlying datasets are as pure as possible.
 </p>
 
-<p>With SVMs we distinguish between hard margin and soft margins. The
-latter introduces a so-called softening parameter to be discussed
-below.  We distinguish also between linear and non-linear
-approaches. The latter are the most frequent ones since it is rather
-unlikely that we can separate classes easily by say straight lines.
+<p>The descriptive features which reproduce best the target/output features are normally  said
+to be the most informative ones. The process of finding the <b>most
+informative</b> feature is done until we accomplish a stopping criteria
+where we then finally end up in so called <b>leaf nodes</b>. 
 </p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="hyperplanes-and-all-that">Hyperplanes and all that </h2>
+<h2 id="basics-of-a-tree">Basics of a tree </h2>
 
-<p>The theory behind support vector machines (SVM hereafter) is based on
-the mathematical description of so-called hyperplanes. Let us start
-with a two-dimensional case. This will also allow us to introduce our
-first SVM examples. These will be tailored to the case of two specific
-classes, as displayed in the figure here based on the usage of the petal data.
+<p>A decision tree is typically divided into a <b>root node</b>, the <b>interior nodes</b>,
+and the final <b>leaf nodes</b> or just <b>leaves</b>. These entities are then connected by so-called <b>branches</b>.
 </p>
 
-<p>We assume here that our data set can be well separated into two
-domains, where a straight line does the job in the separating the two
-classes. Here the two classes are represented by either squares or
-circles.
-</p>
-
-<!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #f8f8f8">
-  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn</span> <span style="color: #008000; font-weight: bold">import</span> datasets
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC, LinearSVC
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.linear_model</span> <span style="color: #008000; font-weight: bold">import</span> SGDClassifier
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> StandardScaler
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib</span>
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;axes.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">14</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;xtick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;ytick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
-
-
-iris <span style="color: #666666">=</span> datasets<span style="color: #666666">.</span>load_iris()
-X <span style="color: #666666">=</span> iris[<span style="color: #BA2121">&quot;data&quot;</span>][:, (<span style="color: #666666">2</span>, <span style="color: #666666">3</span>)]  <span style="color: #408080; font-style: italic"># petal length, petal width</span>
-y <span style="color: #666666">=</span> iris[<span style="color: #BA2121">&quot;target&quot;</span>]
-
-setosa_or_versicolor <span style="color: #666666">=</span> (y <span style="color: #666666">==</span> <span style="color: #666666">0</span>) <span style="color: #666666">|</span> (y <span style="color: #666666">==</span> <span style="color: #666666">1</span>)
-X <span style="color: #666666">=</span> X[setosa_or_versicolor]
-y <span style="color: #666666">=</span> y[setosa_or_versicolor]
-
-
-
-C <span style="color: #666666">=</span> <span style="color: #666666">5</span>
-alpha <span style="color: #666666">=</span> <span style="color: #666666">1</span> <span style="color: #666666">/</span> (C <span style="color: #666666">*</span> <span style="color: #008000">len</span>(X))
-
-lin_clf <span style="color: #666666">=</span> LinearSVC(loss<span style="color: #666666">=</span><span style="color: #BA2121">&quot;hinge&quot;</span>, C<span style="color: #666666">=</span>C, random_state<span style="color: #666666">=42</span>)
-svm_clf <span style="color: #666666">=</span> SVC(kernel<span style="color: #666666">=</span><span style="color: #BA2121">&quot;linear&quot;</span>, C<span style="color: #666666">=</span>C)
-sgd_clf <span style="color: #666666">=</span> SGDClassifier(loss<span style="color: #666666">=</span><span style="color: #BA2121">&quot;hinge&quot;</span>, learning_rate<span style="color: #666666">=</span><span style="color: #BA2121">&quot;constant&quot;</span>, eta0<span style="color: #666666">=0.001</span>, alpha<span style="color: #666666">=</span>alpha,
-                        max_iter<span style="color: #666666">=100000</span>, random_state<span style="color: #666666">=42</span>)
-
-scaler <span style="color: #666666">=</span> StandardScaler()
-X_scaled <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>fit_transform(X)
-
-lin_clf<span style="color: #666666">.</span>fit(X_scaled, y)
-svm_clf<span style="color: #666666">.</span>fit(X_scaled, y)
-sgd_clf<span style="color: #666666">.</span>fit(X_scaled, y)
-
-<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;LinearSVC:                   &quot;</span>, lin_clf<span style="color: #666666">.</span>intercept_, lin_clf<span style="color: #666666">.</span>coef_)
-<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;SVC:                         &quot;</span>, svm_clf<span style="color: #666666">.</span>intercept_, svm_clf<span style="color: #666666">.</span>coef_)
-<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;SGDClassifier(alpha=</span><span style="color: #BB6688; font-weight: bold">{:.5f}</span><span style="color: #BA2121">):&quot;</span><span style="color: #666666">.</span>format(sgd_clf<span style="color: #666666">.</span>alpha), sgd_clf<span style="color: #666666">.</span>intercept_, sgd_clf<span style="color: #666666">.</span>coef_)
-
-<span style="color: #408080; font-style: italic"># Compute the slope and bias of each decision boundary</span>
-w1 <span style="color: #666666">=</span> <span style="color: #666666">-</span>lin_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">0</span>]<span style="color: #666666">/</span>lin_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-b1 <span style="color: #666666">=</span> <span style="color: #666666">-</span>lin_clf<span style="color: #666666">.</span>intercept_[<span style="color: #666666">0</span>]<span style="color: #666666">/</span>lin_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-w2 <span style="color: #666666">=</span> <span style="color: #666666">-</span>svm_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">0</span>]<span style="color: #666666">/</span>svm_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-b2 <span style="color: #666666">=</span> <span style="color: #666666">-</span>svm_clf<span style="color: #666666">.</span>intercept_[<span style="color: #666666">0</span>]<span style="color: #666666">/</span>svm_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-w3 <span style="color: #666666">=</span> <span style="color: #666666">-</span>sgd_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">0</span>]<span style="color: #666666">/</span>sgd_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-b3 <span style="color: #666666">=</span> <span style="color: #666666">-</span>sgd_clf<span style="color: #666666">.</span>intercept_[<span style="color: #666666">0</span>]<span style="color: #666666">/</span>sgd_clf<span style="color: #666666">.</span>coef_[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>]
-
-<span style="color: #408080; font-style: italic"># Transform the decision boundary lines back to the original scale</span>
-line1 <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>inverse_transform([[<span style="color: #666666">-10</span>, <span style="color: #666666">-10</span> <span style="color: #666666">*</span> w1 <span style="color: #666666">+</span> b1], [<span style="color: #666666">10</span>, <span style="color: #666666">10</span> <span style="color: #666666">*</span> w1 <span style="color: #666666">+</span> b1]])
-line2 <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>inverse_transform([[<span style="color: #666666">-10</span>, <span style="color: #666666">-10</span> <span style="color: #666666">*</span> w2 <span style="color: #666666">+</span> b2], [<span style="color: #666666">10</span>, <span style="color: #666666">10</span> <span style="color: #666666">*</span> w2 <span style="color: #666666">+</span> b2]])
-line3 <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>inverse_transform([[<span style="color: #666666">-10</span>, <span style="color: #666666">-10</span> <span style="color: #666666">*</span> w3 <span style="color: #666666">+</span> b3], [<span style="color: #666666">10</span>, <span style="color: #666666">10</span> <span style="color: #666666">*</span> w3 <span style="color: #666666">+</span> b3]])
-
-<span style="color: #408080; font-style: italic"># Plot all three decision boundaries</span>
-plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">4</span>))
-plt<span style="color: #666666">.</span>plot(line1[:, <span style="color: #666666">0</span>], line1[:, <span style="color: #666666">1</span>], <span style="color: #BA2121">&quot;k:&quot;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;LinearSVC&quot;</span>)
-plt<span style="color: #666666">.</span>plot(line2[:, <span style="color: #666666">0</span>], line2[:, <span style="color: #666666">1</span>], <span style="color: #BA2121">&quot;b--&quot;</span>, linewidth<span style="color: #666666">=2</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;SVC&quot;</span>)
-plt<span style="color: #666666">.</span>plot(line3[:, <span style="color: #666666">0</span>], line3[:, <span style="color: #666666">1</span>], <span style="color: #BA2121">&quot;r-&quot;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;SGDClassifier&quot;</span>)
-plt<span style="color: #666666">.</span>plot(X[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==1</span>], X[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==1</span>], <span style="color: #BA2121">&quot;bs&quot;</span>) <span style="color: #408080; font-style: italic"># label=&quot;Iris-Versicolor&quot;</span>
-plt<span style="color: #666666">.</span>plot(X[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==0</span>], X[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==0</span>], <span style="color: #BA2121">&quot;yo&quot;</span>) <span style="color: #408080; font-style: italic"># label=&quot;Iris-Setosa&quot;</span>
-plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">&quot;Petal length&quot;</span>, fontsize<span style="color: #666666">=14</span>)
-plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">&quot;Petal width&quot;</span>, fontsize<span style="color: #666666">=14</span>)
-plt<span style="color: #666666">.</span>legend(loc<span style="color: #666666">=</span><span style="color: #BA2121">&quot;upper center&quot;</span>, fontsize<span style="color: #666666">=14</span>)
-plt<span style="color: #666666">.</span>axis([<span style="color: #666666">0</span>, <span style="color: #666666">5.5</span>, <span style="color: #666666">0</span>, <span style="color: #666666">2</span>])
-
-plt<span style="color: #666666">.</span>show()
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
-
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="what-is-a-hyperplane">What is a hyperplane? </h2>
-
-<p>The aim of the SVM algorithm is to find a hyperplane in a
-\( p \)-dimensional space, where \( p \) is the number of features that
-distinctly classifies the data points.
-</p>
-
-<p>In a \( p \)-dimensional space, a hyperplane is what we call an affine subspace of dimension of \( p-1 \).
-As an example, in two dimension, a hyperplane is simply as straight line while in three dimensions it is 
-a two-dimensional subspace, or stated simply, a plane. 
-</p>
-
-<p>In two dimensions, with the variables \( x_1 \) and \( x_2 \), the hyperplane is defined as</p>
-$$
-b+w_1x_1+w_2x_2=0,
-$$
-
-<p>where \( b \) is the intercept and \( w_1 \) and \( w_2 \) define the elements of a vector orthogonal to the line 
-\( b+w_1x_1+w_2x_2=0 \). 
-In two dimensions we define the vectors \( \boldsymbol{x} =[x1,x2] \) and \( \boldsymbol{w}=[w1,w2] \). 
-We can then rewrite the above equation as 
-</p>
-
-$$
-\boldsymbol{x}^T\boldsymbol{w}+b=0.
-$$
-
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="a-p-dimensional-space-of-features">A \( p \)-dimensional space of features </h2>
-
-<p>We limit ourselves to two classes of outputs \( y_i \) and assign these classes the values \( y_i = \pm 1 \). 
-In a \( p \)-dimensional space of say \( p \) features we have a hyperplane defines as 
-</p>
-$$
-b+wx_1+w_2x_2+\dots +w_px_p=0.
-$$
-
-<p>If we define a 
-matrix \( \boldsymbol{X}=\left[\boldsymbol{x}_1,\boldsymbol{x}_2,\dots, \boldsymbol{x}_p\right] \)
-of dimension \( n\times p \), where \( n \) represents the observations for each feature and each vector \( x_i \) is a column vector of the matrix \( \boldsymbol{X} \), 
-</p>
-$$
-\boldsymbol{x}_i = \begin{bmatrix} x_{i1} \\ x_{i2} \\ \dots \\ \dots \\ x_{ip} \end{bmatrix}.
-$$
-
-<p>If the above condition is not met for a given vector \( \boldsymbol{x}_i \) we have </p>
-$$
-b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} >0,
-$$
-
-<p>if our output \( y_i=1 \).
-In this case we say that \( \boldsymbol{x}_i \) lies on one of the sides of the hyperplane and if 
-</p>
-$$
-b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} < 0,
-$$
-
-<p>for the class of observations \( y_i=-1 \), 
-then \( \boldsymbol{x}_i \) lies on the other side. 
-</p>
-
-<p>Equivalently, for the two classes of observations we have </p>
-$$
-y_i\left(b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip}\right) > 0. 
-$$
-
-<p>When we try to separate hyperplanes, if it exists, we can use it to construct a natural classifier: a test observation is assigned a given class depending on which side of the hyperplane it is located.</p>
-
-<!-- !split  -->
-<h2 id="the-two-dimensional-case">The two-dimensional case </h2>
-
-<p>Let us try to develop our intuition about SVMs by limiting ourselves to a two-dimensional
-plane.  To separate the two classes of data points, there are many
-possible lines (hyperplanes if you prefer a more strict naming)  
-that could be chosen. Our objective is to find a
-plane that has the maximum margin, i.e the maximum distance between
-data points of both classes. Maximizing the margin distance provides
-some reinforcement so that future data points can be classified with
-more confidence.
-</p>
-
-<p>What a linear classifier attempts to accomplish is to split the
-feature space into two half spaces by placing a hyperplane between the
-data points.  This hyperplane will be our decision boundary.  All
-points on one side of the plane will belong to class one and all points
-on the other side of the plane will belong to the second class two.
-</p>
-
-<p>Unfortunately there are many ways in which we can place a hyperplane
-to divide the data.  Below is an example of two candidate hyperplanes
-for our data sample.
+<p>The leaf nodes
+contain the predictions we will make for new query instances presented
+to our trained model. This is possible since the model has 
+learned the underlying structure of the training data and hence can,
+given some assumptions, make predictions about the target feature value
+(class) of unseen query instances.
 </p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="getting-into-the-details">Getting into the details </h2>
+<h2 id="a-sketch-of-a-tree-regression-problem">A Sketch of a Tree, Regression problem  </h2>
 
-<p>Let us define the function</p>
-$$
-f(x) = \boldsymbol{w}^T\boldsymbol{x}+b = 0,
-$$
-
-<p>as the function that determines the line \( L \) that separates two classes (our two features), see the figure here. </p>
-
-<p>Any point defined by \( \boldsymbol{x}_i \) and \( \boldsymbol{x}_2 \) on the line \( L \) will satisfy \( \boldsymbol{w}^T(\boldsymbol{x}_1-\boldsymbol{x}_2)=0 \). </p>
-
-<p>The signed distance \( \delta \) from any point defined by a vector \( \boldsymbol{x} \) and a point \( \boldsymbol{x}_0 \) on the line \( L \) is then</p>
-$$
-\delta = \frac{1}{\vert\vert \boldsymbol{w}\vert\vert}(\boldsymbol{w}^T\boldsymbol{x}+b).
-$$
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
 
+<!-- FIGURE: [DataFiles/Regsimpletree.png, width=600 frac=0.8] -->
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="first-attempt-at-a-minimization-approach">First attempt at a minimization approach </h2>
-
-<p>How do we find the parameter \( b \) and the vector \( \boldsymbol{w} \)? What we could
-do is to define a cost function which now contains the set of all
-misclassified points \( M \) and attempt to minimize this function
-</p>
-
-$$
-C(\boldsymbol{w},b) = -\sum_{i\in M} y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b).
-$$
-
-<p>We could now for example define all values \( y_i =1 \) as misclassified in case we have \( \boldsymbol{w}^T\boldsymbol{x}_i+b < 0 \) and the opposite if we have \( y_i=-1 \). Taking the derivatives gives us</p>
-$$
-\frac{\partial C}{\partial b} = -\sum_{i\in M} y_i,
-$$
-
-<p>and </p>
-$$
-\frac{\partial C}{\partial \boldsymbol{w}} = -\sum_{i\in M} y_ix_i.
-$$
+<h2 id="a-sketch-of-a-tree-classification-problem">A Sketch of a Tree, Classification  problem  </h2>
 
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+<!-- FIGURE: [DataFiles/Classimpletree.png, width=600 frac=0.8] -->
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="solving-the-equations">Solving the equations </h2>
+<h2 id="a-typical-decision-tree-with-its-pertinent-jargon-classification-problem">A typical Decision Tree with its pertinent Jargon, Classification Problem </h2>
 
-<p>We can now use the Newton-Raphson method or different variants of the gradient descent family (from plain gradient descent to various stochastic gradient descent approaches) to solve the equations</p>
-$$
-b \leftarrow b +\eta \frac{\partial C}{\partial b},
-$$
+<br/><br/>
+<center>
+<p><img src="DataFiles/cancer.png" width="600" align="bottom"></p>
+</center>
+<br/><br/>
 
-<p>and</p>
-$$
-\boldsymbol{w} \leftarrow \boldsymbol{w} +\eta \frac{\partial C}{\partial \boldsymbol{w}},
-$$
-
-<p>where \( \eta \) is our by now well-known learning rate. </p>
+<p>This tree was produced using the Wisconsin cancer data (discussed here as well, see code examples below) using <b>Scikit-Learn</b>'s decision tree classifier. Here we have used the so-called <b>gini</b> index (see below) to split the various branches.</p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="code-example">Code Example </h2>
+<h2 id="general-features">General Features </h2>
 
-<p>The equations we discussed above can be coded rather easily (the
-framework is similar to what we developed for logistic
-regression). We are going to set up a simple case with two classes only and we want to find a line which separates them the best possible way.
-</p>
+<p>The overarching approach to decision trees is a top-down approach.</p>
 
-<!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #f8f8f8">
-  <pre style="line-height: 125%;">
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
-
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="problems-with-the-simpler-approach">Problems with the Simpler Approach </h2>
-
-<p>There are however problems with this approach, although it looks
-pretty straightforward to implement. When running the above code, we see that we can easily end up with many diffeent lines which separate the two classes.
-</p>
-
-<p>For small
-gaps between the entries, we may also end up needing many iterations
-before the solutions converge and if the data cannot be separated
-properly into two distinct classes, we may not experience a converge
-at all.
+<ul>
+<li> A leaf provides the classification of a given instance.</li>
+<li> A node specifies a test of some attribute of the instance.</li>
+<li> A branch corresponds to a possible values of an attribute.</li>
+<li> An instance is classified by starting at the root node of the tree, testing the attribute specified by this node, then moving down the tree branch corresponding to the value of the attribute in the given example.</li>
+</ul>
+<p>This process is then repeated for the subtree rooted at the new
+node.
 </p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="a-better-approach">A better approach </h2>
+<h2 id="how-do-we-set-it-up">How do we set it up? </h2>
 
-<p>A better approach is rather to try to define a large margin between
-the two classes (if they are well separated from the beginning).
+<p>In simplified terms, the process of training a decision tree and
+predicting the target features of query instances is as follows:
 </p>
 
-<p>Thus, we wish to find a margin \( M \) with \( \boldsymbol{w} \) normalized to
-\( \vert\vert \boldsymbol{w}\vert\vert =1 \) subject to the condition
-</p>
-
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, p. 
-$$
-
-<p>All points are thus at a signed distance from the decision boundary defined by the line \( L \). The parameters \( b \) and \( w_1 \) and \( w_2 \) define this line. </p>
-
-<p>We seek thus the largest value \( M \) defined by</p>
-$$
-\frac{1}{\vert \vert \boldsymbol{w}\vert\vert}y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, n, 
-$$
-
-<p>or just </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq M\vert \vert \boldsymbol{w}\vert\vert \hspace{0.1cm}\forall i. 
-$$
-
-<p>If we scale the equation so that \( \vert \vert \boldsymbol{w}\vert\vert = 1/M \), we have to find the minimum of 
-\( \boldsymbol{w}^T\boldsymbol{w}=\vert \vert \boldsymbol{w}\vert\vert \) (the norm) subject to the condition
-</p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) \geq 1 \hspace{0.1cm}\forall i.
-$$
-
-<p>We have thus defined our margin as the invers of the norm of
-\( \boldsymbol{w} \). We want to minimize the norm in order to have a as large as
-possible margin \( M \). Before we proceed, we need to remind ourselves
-about Lagrangian multipliers.
-</p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="a-quick-reminder-on-lagrangian-multipliers">A quick Reminder on Lagrangian Multipliers </h2>
-
-<p>Consider a function of three independent variables \( f(x,y,z) \) . For the function \( f \) to be an
-extreme we have
-</p>
-$$
-df=0.
-$$
-
-<p>A necessary and sufficient condition is</p>
-$$
-\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
-$$
-
-<p>due to</p>
-$$
-df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz.
-$$
-
-<p>In many problems the variables \( x,y,z \) are often subject to constraints (such as those above for the margin)
-so that they are no longer all independent. It is possible at least in principle to use each 
-constraint to eliminate one variable
-and to proceed with a new and smaller set of independent varables.
-</p>
-
-<p>The use of so-called Lagrangian  multipliers is an alternative technique  when the elimination
-of variables is incovenient or undesirable.  Assume that we have an equation of constraint on 
-the variables \( x,y,z \)
-</p>
-$$
-\phi(x,y,z) = 0,
-$$
-
-<p> resulting in</p>
-$$
-d\phi = \frac{\partial \phi}{\partial x}dx+\frac{\partial \phi}{\partial y}dy+\frac{\partial \phi}{\partial z}dz =0.
-$$
-
-<p>Now we cannot set anymore</p>
-$$
-\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
-$$
-
-<p>if \( df=0 \) is wanted
-because there are now only two independent variables!  Assume \( x \) and \( y \) are the independent 
-variables.
-Then \( dz \) is no longer arbitrary.
-</p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="adding-the-multiplier">Adding the Multiplier </h2>
-
-<p>However, we can add to</p>
-$$
-df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz,
-$$
-
-<p>a multiplum of \( d\phi \), viz. \( \lambda d\phi \), resulting  in</p>
-$$
-df+\lambda d\phi = (\frac{\partial f}{\partial z}+\lambda
-\frac{\partial \phi}{\partial x})dx+(\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y})dy+
-(\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z})dz =0.
-$$
-
-<p>Our multiplier is chosen so that</p>
-$$
-\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z} =0.
-$$
-
-<p>We need to remember that we took \( dx \) and \( dy \) to be arbitrary and thus we must have</p>
-$$
-\frac{\partial f}{\partial x}+\lambda\frac{\partial \phi}{\partial x} =0,
-$$
-
-<p>and</p>
-$$
-\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y} =0.
-$$
-
-<p>When all these equations are satisfied, \( df=0 \).  We have four unknowns, \( x,y,z \) and
-\( \lambda \). Actually we want only \( x,y,z \), \( \lambda \) needs not to be determined, 
-it is therefore often called
-Lagrange's undetermined multiplier.
-If we have a set of constraints \( \phi_k \) we have the equations
-</p>
-$$
-\frac{\partial f}{\partial x_i}+\sum_k\lambda_k\frac{\partial \phi_k}{\partial x_i} =0.
-$$
-
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="setting-up-the-problem">Setting up the Problem </h2>
-<p>In order to solve the above problem, we define the following Lagrangian function to be minimized </p>
-$$
-{\cal L}(\lambda,b,\boldsymbol{w})=\frac{1}{2}\boldsymbol{w}^T\boldsymbol{w}-\sum_{i=1}^n\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)-1\right],
-$$
-
-<p>where \( \lambda_i \) is a so-called Lagrange multiplier subject to the condition \( \lambda_i \geq 0 \).</p>
-
-<p>Taking the derivatives  with respect to \( b \) and \( \boldsymbol{w} \) we obtain </p>
-$$
-\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
-$$
-
-<p>and </p>
-$$
-\frac{\partial {\cal L}}{\partial \boldsymbol{w}} = 0 = \boldsymbol{w}-\sum_{i} \lambda_iy_i\boldsymbol{x}_i.
-$$
-
-<p>Inserting these constraints into the equation for \( {\cal L} \) we obtain</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-
-<p>subject to the constraints \( \lambda_i\geq 0 \) and \( \sum_i\lambda_iy_i=0 \). 
-We must in addition satisfy the <a href="https://en.wikipedia.org/wiki/Karush%E2%80%93Kuhn%E2%80%93Tucker_conditions" target="_blank">Karush-Kuhn-Tucker</a> (KKT) condition
-</p>
-$$
-\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -1\right] \hspace{0.1cm}\forall i.
-$$
-
 <ol>
-<li> If \( \lambda_i > 0 \), then \( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1 \) and we say that \( x_i \) is on the boundary.</li>
-<li> If \( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)> 1 \), we say \( x_i \) is not on the boundary and we set \( \lambda_i=0 \).</li> 
+<li> Present a dataset containing of a number of training instances characterized by a number of descriptive features and a target feature</li>
+<li> Train the decision tree model by continuously splitting the target feature along the values of the descriptive features using a measure of information gain during the training process</li>
+<li> Grow the tree until we accomplish a stopping criteria create leaf nodes which represent the <em>predictions</em> we want to make for new query instances</li>
+<li> Show query instances to the tree and run down the tree until we arrive at leaf nodes</li>
 </ol>
-<p>When \( \lambda_i > 0 \), the vectors \( \boldsymbol{x}_i \) are called support vectors. They are the vectors closest to the line (or hyperplane) and define the margin \( M \). </p>
+<p>Then we are essentially done!</p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="the-problem-to-solve">The problem to solve </h2>
-
-<p>We can rewrite </p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-
-<p>and its constraints in terms of a matrix-vector problem where we minimize w.r.t. \( \lambda \) the following problem</p>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1\boldsymbol{x}_1^T\boldsymbol{x}_1 & y_1y_2\boldsymbol{x}_1^T\boldsymbol{x}_2 & \dots & \dots & y_1y_n\boldsymbol{x}_1^T\boldsymbol{x}_n \\
-y_2y_1\boldsymbol{x}_2^T\boldsymbol{x}_1 & y_2y_2\boldsymbol{x}_2^T\boldsymbol{x}_2 & \dots & \dots & y_1y_n\boldsymbol{x}_2^T\boldsymbol{x}_n \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1\boldsymbol{x}_n^T\boldsymbol{x}_1 & y_ny_2\boldsymbol{x}_n^T\boldsymbol{x}_2 & \dots & \dots & y_ny_n\boldsymbol{x}_n^T\boldsymbol{x}_n \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{1}\boldsymbol{\lambda}, 
-$$
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
-</p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="the-last-steps">The last steps </h2>
-
-<p>Solving the above problem, yields the values of \( \lambda_i \).
-To find the coefficients of your hyperplane we need simply to compute 
-</p>
-$$
-\boldsymbol{w}=\sum_{i} \lambda_iy_i\boldsymbol{x}_i.
-$$
-
-<p>With our vector \( \boldsymbol{w} \) we can in turn find the value of the intercept \( b \) (here in two dimensions) via </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1,
-$$
-
-<p>resulting in</p>
-$$
-b = \frac{1}{y_i}-\boldsymbol{w}^T\boldsymbol{x}_i,
-$$
-
-<p>or if we write it out in terms of the support vectors only, with \( N_s \) being their number,  we have</p>
-$$
-b = \frac{1}{N_s}\sum_{j\in N_s}\left(y_j-\sum_{i=1}^n\lambda_iy_i\boldsymbol{x}_i^T\boldsymbol{x}_j\right).
-$$
-
-<p>With our hyperplane coefficients we can use our classifier to assign any observation by simply using </p>
-$$
-y_i = \mathrm{sign}(\boldsymbol{w}^T\boldsymbol{x}_i+b).
-$$
-
-<p>Below we discuss how to find the optimal values of \( \lambda_i \). Before we proceed however, we discuss now the so-called soft classifier. </p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="a-soft-classifier">A soft classifier </h2>
-
-<p>Till now, the margin is strictly defined by the support vectors. This defines what is called a hard classifier, that is the margins are well defined.</p>
-
-<p>Suppose now that classes overlap in feature space, as shown in the
-figure here. One way to deal with this problem before we define the
-so-called <b>kernel approach</b>, is to allow a kind of slack in the sense
-that we allow some points to be on the wrong side of the margin.
-</p>
-
-<p>We introduce thus the so-called <b>slack</b> variables \( \boldsymbol{\xi} =[\xi_1,x_2,\dots,x_n] \) and 
-modify our previous equation
-</p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1,
-$$
-
-<p>to </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1-\xi_i,
-$$
-
-<p>with the requirement \( \xi_i\geq 0 \). The total violation is now \( \sum_i\xi \). 
-The value \( \xi_i \) in the constraint the last constraint corresponds to the  amount by which the prediction
-\( y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1 \) is on the wrong side of its margin. Hence by bounding the sum \( \sum_i \xi_i \),
-we bound the total amount by which predictions fall on the wrong side of their margins.
-</p>
-
-<p>Misclassifications occur when \( \xi_i > 1 \). Thus bounding the total sum by some value \( C \) bounds in turn the total number of
-misclassifications.
-</p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="soft-optmization-problem">Soft optmization problem </h2>
-
-<p>This has in turn the consequences that we change our optmization problem to finding the minimum of </p>
-$$
-{\cal L}=\frac{1}{2}\boldsymbol{w}^T\boldsymbol{w}-\sum_{i=1}^n\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)-(1-\xi_)\right]+C\sum_{i=1}^n\xi_i-\sum_{i=1}^n\gamma_i\xi_i,
-$$
-
-<p>subject to </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b)=1-\xi_i \hspace{0.1cm}\forall i,
-$$
-
-<p>with the requirement \( \xi_i\geq 0 \).</p>
-
-<p>Taking the derivatives  with respect to \( b \) and \( \boldsymbol{w} \) we obtain </p>
-$$
-\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
-$$
-
-<p>and </p>
-$$
-\frac{\partial {\cal L}}{\partial \boldsymbol{w}} = 0 = \boldsymbol{w}-\sum_{i} \lambda_iy_i\boldsymbol{x}_i, 
-$$
-
-<p>and</p>
-$$
-\lambda_i = C-\gamma_i \hspace{0.1cm}\forall i.
-$$
-
-<p>Inserting these constraints into the equation for \( {\cal L} \) we obtain the same equation as before</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{x}_j,
-$$
-
-<p>but now subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \) and \( 0\leq\lambda_i \leq C \). 
-We must in addition satisfy the Karush-Kuhn-Tucker condition which now reads
-</p>
-$$
-\lambda_i\left[y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -(1-\xi_)\right]=0 \hspace{0.1cm}\forall i,
-$$
-
-$$
-\gamma_i\xi_i = 0,
-$$
-
-<p>and </p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{x}_i+b) -(1-\xi_) \geq 0 \hspace{0.1cm}\forall i.
-$$
-
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="kernels-and-non-linearity">Kernels and non-linearity </h2>
-
-<p>The cases we have studied till now, were all characterized by two classes
-with a close to linear separability. The classifiers we have described
-so far find linear boundaries in our input feature space. It is
-possible to make our procedure more flexible by exploring the feature
-space using other basis expansions such as higher-order polynomials,
-wavelets, splines etc.
-</p>
-
-<p>If our feature space is not easy to separate, as shown in the figure
-here, we can achieve a better separation by introducing more complex
-basis functions. The ideal would be, as shown in the next figure, to, via a specific transformation to 
-obtain a separation between the classes which is almost linear. 
-</p>
-
-<p>The change of basis, from \( x\rightarrow z=\phi(x) \) leads to the same type of equations to be solved, except that
-we need to introduce for example a polynomial transformation to a two-dimensional training set.
-</p>
-
+<h2 id="decision-trees-and-regression">Decision trees and Regression  </h2>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
 <div class="cell border-box-sizing code_cell rendered">
@@ -945,51 +501,444 @@ we need to introduce for example a polynomial transformation to a two-dimensiona
       <div class="input_area">
         <div class="highlight" style="background: #f8f8f8">
   <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> PolynomialFeatures
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.linear_model</span> <span style="color: #008000; font-weight: bold">import</span> LinearRegression
+
+steps<span style="color: #666666">=250</span>
+
+distance<span style="color: #666666">=0</span>
+x<span style="color: #666666">=0</span>
+distance_list<span style="color: #666666">=</span>[]
+steps_list<span style="color: #666666">=</span>[]
+<span style="color: #008000; font-weight: bold">while</span> x<span style="color: #666666">&lt;</span>steps:
+    distance<span style="color: #666666">+=</span>np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>randint(<span style="color: #666666">-1</span>,<span style="color: #666666">2</span>)
+    distance_list<span style="color: #666666">.</span>append(distance)
+    x<span style="color: #666666">+=1</span>
+    steps_list<span style="color: #666666">.</span>append(x)
+plt<span style="color: #666666">.</span>plot(steps_list,distance_list, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;green&#39;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;Random Walk Data&quot;</span>)
+
+steps_list<span style="color: #666666">=</span>np<span style="color: #666666">.</span>asarray(steps_list)
+distance_list<span style="color: #666666">=</span>np<span style="color: #666666">.</span>asarray(distance_list)
+
+X<span style="color: #666666">=</span>steps_list[:,np<span style="color: #666666">.</span>newaxis]
+
+<span style="color: #408080; font-style: italic">#Polynomial fits</span>
+
+<span style="color: #408080; font-style: italic">#Degree 2</span>
+poly_features<span style="color: #666666">=</span>PolynomialFeatures(degree<span style="color: #666666">=2</span>, include_bias<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">False</span>)
+X_poly<span style="color: #666666">=</span>poly_features<span style="color: #666666">.</span>fit_transform(X)
+
+lin_reg<span style="color: #666666">=</span>LinearRegression()
+poly_fit<span style="color: #666666">=</span>lin_reg<span style="color: #666666">.</span>fit(X_poly,distance_list)
+b<span style="color: #666666">=</span>lin_reg<span style="color: #666666">.</span>coef_
+c<span style="color: #666666">=</span>lin_reg<span style="color: #666666">.</span>intercept_
+<span style="color: #008000">print</span> (<span style="color: #BA2121">&quot;2nd degree coefficients:&quot;</span>)
+<span style="color: #008000">print</span> (<span style="color: #BA2121">&quot;zero power: &quot;</span>,c)
+<span style="color: #008000">print</span> (<span style="color: #BA2121">&quot;first power: &quot;</span>, b[<span style="color: #666666">0</span>])
+<span style="color: #008000">print</span> (<span style="color: #BA2121">&quot;second power: &quot;</span>,b[<span style="color: #666666">1</span>])
+
+z <span style="color: #666666">=</span> np<span style="color: #666666">.</span>arange(<span style="color: #666666">0</span>, steps, <span style="color: #666666">.01</span>)
+z_mod<span style="color: #666666">=</span>b[<span style="color: #666666">1</span>]<span style="color: #666666">*</span>z<span style="color: #666666">**2+</span>b[<span style="color: #666666">0</span>]<span style="color: #666666">*</span>z<span style="color: #666666">+</span>c
+
+fit_mod<span style="color: #666666">=</span>b[<span style="color: #666666">1</span>]<span style="color: #666666">*</span>X<span style="color: #666666">**2+</span>b[<span style="color: #666666">0</span>]<span style="color: #666666">*</span>X<span style="color: #666666">+</span>c
+plt<span style="color: #666666">.</span>plot(z, z_mod, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;r&#39;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;2nd Degree Fit&quot;</span>)
+plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">&quot;Polynomial Regression&quot;</span>)
+
+plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">&quot;Steps&quot;</span>)
+plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">&quot;Distance&quot;</span>)
+
+<span style="color: #408080; font-style: italic">#Degree 10</span>
+poly_features10<span style="color: #666666">=</span>PolynomialFeatures(degree<span style="color: #666666">=10</span>, include_bias<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">False</span>)
+X_poly10<span style="color: #666666">=</span>poly_features10<span style="color: #666666">.</span>fit_transform(X)
+
+poly_fit10<span style="color: #666666">=</span>lin_reg<span style="color: #666666">.</span>fit(X_poly10,distance_list)
+
+y_plot<span style="color: #666666">=</span>poly_fit10<span style="color: #666666">.</span>predict(X_poly10)
+plt<span style="color: #666666">.</span>plot(X, y_plot, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;black&#39;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;10th Degree Fit&quot;</span>)
+
+plt<span style="color: #666666">.</span>legend()
+plt<span style="color: #666666">.</span>show()
+
+
+<span style="color: #408080; font-style: italic">#Decision Tree Regression</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeRegressor
+regr_1<span style="color: #666666">=</span>DecisionTreeRegressor(max_depth<span style="color: #666666">=2</span>)
+regr_2<span style="color: #666666">=</span>DecisionTreeRegressor(max_depth<span style="color: #666666">=5</span>)
+regr_3<span style="color: #666666">=</span>DecisionTreeRegressor(max_depth<span style="color: #666666">=7</span>)
+regr_1<span style="color: #666666">.</span>fit(X, distance_list)
+regr_2<span style="color: #666666">.</span>fit(X, distance_list)
+regr_3<span style="color: #666666">.</span>fit(X, distance_list)
+
+X_test <span style="color: #666666">=</span> np<span style="color: #666666">.</span>arange(<span style="color: #666666">0.0</span>, steps, <span style="color: #666666">0.01</span>)[:, np<span style="color: #666666">.</span>newaxis]
+y_1 <span style="color: #666666">=</span> regr_1<span style="color: #666666">.</span>predict(X_test)
+y_2 <span style="color: #666666">=</span> regr_2<span style="color: #666666">.</span>predict(X_test)
+y_3<span style="color: #666666">=</span>regr_3<span style="color: #666666">.</span>predict(X_test)
+
+<span style="color: #408080; font-style: italic"># Plot the results</span>
+plt<span style="color: #666666">.</span>figure()
+plt<span style="color: #666666">.</span>scatter(X, distance_list, s<span style="color: #666666">=2.5</span>, c<span style="color: #666666">=</span><span style="color: #BA2121">&quot;black&quot;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;data&quot;</span>)
+plt<span style="color: #666666">.</span>plot(X_test, y_1, color<span style="color: #666666">=</span><span style="color: #BA2121">&quot;red&quot;</span>,
+         label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;max_depth=2&quot;</span>, linewidth<span style="color: #666666">=2</span>)
+plt<span style="color: #666666">.</span>plot(X_test, y_2, color<span style="color: #666666">=</span><span style="color: #BA2121">&quot;green&quot;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;max_depth=5&quot;</span>, linewidth<span style="color: #666666">=2</span>)
+plt<span style="color: #666666">.</span>plot(X_test, y_3, color<span style="color: #666666">=</span><span style="color: #BA2121">&quot;m&quot;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;max_depth=7&quot;</span>, linewidth<span style="color: #666666">=2</span>)
+
+plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">&quot;Data&quot;</span>)
+plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">&quot;Darget&quot;</span>)
+plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">&quot;Decision Tree Regression&quot;</span>)
+plt<span style="color: #666666">.</span>legend()
+plt<span style="color: #666666">.</span>show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="building-a-tree-regression">Building a tree, regression </h2>
+
+<p>There are mainly two steps</p>
+<ol>
+<li> We split the predictor space (the set of possible values \( x_1,x_2,\dots, x_p \)) into \( J \) distinct and non-non-overlapping regions, \( R_1,R_2,\dots,R_J \).</li>  
+<li> For every observation that falls into the region \( R_j \) , we make the same prediction, which is simply the mean of the response values for the training observations in \( R_j \).</li>
+</ol>
+<p>How do we construct the regions \( R_1,\dots,R_J \)?  In theory, the
+regions could have any shape. However, we choose to divide the
+predictor space into high-dimensional rectangles, or boxes, for
+simplicity and for ease of interpretation of the resulting predictive
+model. The goal is to find boxes \( R_1,\dots,R_J \) that minimize the
+MSE, given by
+</p>
+
+$$
+\sum_{j=1}^J\sum_{i\in R_j}(y_i-\overline{y}_{R_j})^2,
+$$
+
+<p>where \( \overline{y}_{R_j} \)  is the mean response for the training observations 
+within box \( j \). 
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="a-top-down-approach-recursive-binary-splitting">A top-down approach, recursive binary splitting </h2>
+
+<p>Unfortunately, it is computationally infeasible to consider every
+possible partition of the feature space into \( J \) boxes.  The common
+strategy is to take a top-down approach
+</p>
+
+<p>The approach is top-down because it begins at the top of the tree (all
+observations belong to a single region) and then successively splits
+the predictor space; each split is indicated via two new branches
+further down on the tree. It is greedy because at each step of the
+tree-building process, the best split is made at that particular step,
+rather than looking ahead and picking a split that will lead to a
+better tree in some future step.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="making-a-tree">Making a tree </h2>
+
+<p>In order to implement the recursive binary splitting we start by selecting
+the predictor \( x_j \) and a cutpoint \( s \) that splits the predictor space into two regions \( R_1 \) and \( R_2 \)
+</p>
+$$
+\left\{X\vert x_j < s\right\},
+$$
+
+<p>and</p>
+$$
+\left\{X\vert x_j \geq s\right\},
+$$
+
+<p>so that we obtain the lowest MSE, that is</p>
+$$
+\sum_{i:x_i\in R_j}(y_i-\overline{y}_{R_1})^2+\sum_{i:x_i\in R_2}(y_i-\overline{y}_{R_2})^2,
+$$
+
+<p>which we want to minimize by considering all predictors
+\( x_1,x_2,\dots,x_p \).  We consider also all possible values of \( s \) for
+each predictor. These values could be determined by randomly assigned
+numbers or by starting at the midpoint and then proceed till we find
+an optimal value.
+</p>
+
+<p>For any \( j \) and \( s \), we define the pair of half-planes where
+\( \overline{y}_{R_1} \) is the mean response for the training
+observations in \( R_1(j,s) \), and \( \overline{y}_{R_2} \) is the mean
+response for the training observations in \( R_2(j,s) \).
+</p>
+
+<p>Finding the values of \( j \) and \( s \) that minimize the above equation can be
+done quite quickly, especially when the number of features \( p \) is not
+too large.
+</p>
+
+<p>Next, we repeat the process, looking
+for the best predictor and best cutpoint in order to split the data
+further so as to minimize the MSE within each of the resulting
+regions. However, this time, instead of splitting the entire predictor
+space, we split one of the two previously identified regions. We now
+have three regions. Again, we look to split one of these three regions
+further, so as to minimize the MSE. The process continues until a
+stopping criterion is reached; for instance, we may continue until no
+region contains more than five observations.
+</p>
+
+<!-- !split  -->
+<h2 id="pruning-the-tree">Pruning the tree </h2>
+
+<p>The above procedure is rather straightforward, but leads often to
+overfitting and unnecessarily large and complicated trees. The basic
+idea is to grow a large tree \( T_0 \) and then prune it back in order to
+obtain a subtree. A smaller tree with fewer splits (fewer regions) can
+lead to smaller variance and better interpretation at the cost of a
+little more bias.
+</p>
+
+<p>The so-called Cost complexity pruning algorithm gives us a
+way to do just this. Rather than considering every possible subtree,
+we consider a sequence of trees indexed by a nonnegative tuning
+parameter \( \alpha \).
+</p>
+
+<p>Read more at the following <a href="https://scikit-learn.org/stable/auto_examples/tree/plot_cost_complexity_pruning.html#sphx-glr-auto-examples-tree-plot-cost-complexity-pruning-py" target="_blank">Scikit-Learn link on pruning</a>.</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="cost-complexity-pruning">Cost complexity pruning </h2>
+
+<p>For each value of \( \alpha \)  there corresponds a subtree \( T \in T_0 \) such that</p>
+$$
+\sum_{m=1}^{\overline{T}}\sum_{i:x_i\in R_m}(y_i-\overline{y}_{R_m})^2+\alpha\overline{T},
+$$
+
+<p>is as small as possible. Here \( \overline{T} \) is 
+the number of terminal nodes of the tree \( T \) , \( R_m \) is the
+rectangle (i.e. the subset of predictor space)  corresponding to the \( m \)-th terminal node.
+</p>
+
+<p>The tuning parameter \( \alpha \) controls a trade-off between the subtree&#8217;s
+complexity and its fit to the training data. When \( \alpha = 0 \), then the
+subtree \( T \) will simply equal \( T_0 \), 
+because then the above equation just measures the
+training error. 
+However, as \( \alpha \) increases, there is a price to pay for
+having a tree with many terminal nodes. The above equation will
+tend to be minimized for a smaller subtree. 
+</p>
+
+<p>It turns out that as we increase \( \alpha \) from zero
+branches get pruned from the tree in a nested and predictable fashion,
+so obtaining the whole sequence of subtrees as a function of \( \alpha \) is
+easy. We can select a value of \( \alpha \) using a validation set or using
+cross-validation. We then return to the full data set and obtain the
+subtree corresponding to \( \alpha \). 
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="schematic-regression-procedure">Schematic Regression Procedure </h2>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Building a Regression Tree</b>
+<p>
+
+<ol>
+<li> Use recursive binary splitting to grow a large tree on the training data, stopping only when each terminal node has fewer than some minimum number of observations.</li>
+<li> Apply cost complexity pruning to the large tree in order to obtain a sequence of best subtrees, as a function of \( \alpha \).</li>
+<li> Use for example \( K \)-fold cross-validation to choose \( \alpha \). Divide the training observations into \( K \) folds. For each \( k=1,2,\dots,K \) we:</li> 
+<ul>
+  <li> repeat steps 1 and 2 on all but the \( k \)-th fold of the training data.</li> 
+  <li> Then we valuate the mean squared prediction error on the data in the left-out \( k \)-th fold, as a function of \( \alpha \).</li>
+  <li> Finally  we average the results for each value of \( \alpha \), and pick \( \alpha \) to minimize the average error.</li>
+</ul>
+<li> Return the subtree from Step 2 that corresponds to the chosen value of \( \alpha \).</li> 
+</ol>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="a-classification-tree">A Classification Tree </h2>
+
+<p>A classification tree is very similar to a regression tree, except
+that it is used to predict a qualitative response rather than a
+quantitative one. Recall that for a regression tree, the predicted
+response for an observation is given by the mean response of the
+training observations that belong to the same terminal node. In
+contrast, for a classification tree, we predict that each observation
+belongs to the most commonly occurring class of training observations
+in the region to which it belongs. In interpreting the results of a
+classification tree, we are often interested not only in the class
+prediction corresponding to a particular terminal node region, but
+also in the class proportions among the training observations that
+fall into that region.  
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="growing-a-classification-tree">Growing a classification tree </h2>
+
+<p>The task of growing a
+classification tree is quite similar to the task of growing a
+regression tree. Just as in the regression setting, we use recursive
+binary splitting to grow a classification tree. However, in the
+classification setting, the MSE cannot be used as a criterion for making
+the binary splits.  A natural alternative to MSE is the <b>classification
+error rate</b>. Since we plan to assign an observation in a given region
+to the most commonly occurring error rate class of training
+observations in that region, the classification error rate is simply
+the fraction of the training observations in that region that do not
+belong to the most common class. 
+</p>
+
+<p>When building a classification tree, either the Gini index or the
+entropy are typically used to evaluate the quality of a particular
+split, since these two approaches are more sensitive to node purity
+than is the classification error rate. 
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="classification-tree-how-to-split-nodes">Classification tree, how to split nodes </h2>
+
+<p>If our targets are the outcome of a classification process that takes
+for example \( k=1,2,\dots,K \) values, the only thing we need to think of
+is to set up the splitting criteria for each node.
+</p>
+
+<p>We define a PDF \( p_{mk} \) that represents the number of observations of
+a class \( k \) in a region \( R_m \) with \( N_m \) observations. We represent
+this likelihood function in terms of the proportion \( I(y_i=k) \) of
+observations of this class in the region \( R_m \) as
+</p>
+
+$$
+p_{mk} = \frac{1}{N_m}\sum_{x_i\in R_m}I(y_i=k).
+$$
+
+<p>We let \( p_{mk} \) represent the majority class of observations in region
+\( m \). The three most common ways of splitting a node are given by
+</p>
+
+<ul>
+<li> Misclassification error</li> 
+</ul>
+$$
+p_{mk} = \frac{1}{N_m}\sum_{x_i\in R_m}I(y_i\ne k) = 1-p_{mk}.
+$$
+
+<ul>
+<li> Gini index \( g \)</li>
+</ul>
+$$
+g = \sum_{k=1}^K p_{mk}(1-p_{mk}).
+$$
+
+<ul>
+<li> Information entropy or just entropy \( s \)</li>
+</ul>
+$$
+s = -\sum_{k=1}^K p_{mk}\log{p_{mk}}.
+$$
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="visualizing-the-tree-classification">Visualizing the Tree, Classification </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">os</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> load_breast_cancer
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span> train_test_split
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.metrics</span> <span style="color: #008000; font-weight: bold">import</span> confusion_matrix
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> export_graphviz
+
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">IPython.display</span> <span style="color: #008000; font-weight: bold">import</span> Image 
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">pydot</span> <span style="color: #008000; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">pandas</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">pd</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
+
+
+cancer <span style="color: #666666">=</span> load_breast_cancer()
+X <span style="color: #666666">=</span> pd<span style="color: #666666">.</span>DataFrame(cancer<span style="color: #666666">.</span>data, columns<span style="color: #666666">=</span>cancer<span style="color: #666666">.</span>feature_names)
+<span style="color: #008000">print</span>(X)
+y <span style="color: #666666">=</span> pd<span style="color: #666666">.</span>Categorical<span style="color: #666666">.</span>from_codes(cancer<span style="color: #666666">.</span>target, cancer<span style="color: #666666">.</span>target_names)
+y <span style="color: #666666">=</span> pd<span style="color: #666666">.</span>get_dummies(y)
+<span style="color: #008000">print</span>(y)
+X_train, X_test, y_train, y_test <span style="color: #666666">=</span> train_test_split(X, y, random_state<span style="color: #666666">=1</span>)
+tree_clf <span style="color: #666666">=</span> DecisionTreeClassifier(max_depth<span style="color: #666666">=5</span>)
+tree_clf<span style="color: #666666">.</span>fit(X_train, y_train)
+
+export_graphviz(
+    tree_clf,
+    out_file<span style="color: #666666">=</span><span style="color: #BA2121">&quot;DataFiles/cancer.dot&quot;</span>,
+    feature_names<span style="color: #666666">=</span>cancer<span style="color: #666666">.</span>feature_names,
+    class_names<span style="color: #666666">=</span>cancer<span style="color: #666666">.</span>target_names,
+    rounded<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>,
+    filled<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>
+)
+cmd <span style="color: #666666">=</span> <span style="color: #BA2121">&#39;dot -Tpng DataFiles/cancer.dot -o DataFiles/cancer.png&#39;</span>
+os<span style="color: #666666">.</span>system(cmd)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="visualizing-the-tree-the-moons">Visualizing the Tree, The Moons  </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #408080; font-style: italic"># Common imports</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span>  train_test_split 
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> make_moons
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> export_graphviz
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">pydot</span> <span style="color: #008000; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">pandas</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">pd</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">os</span>
 
 np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>seed(<span style="color: #666666">42</span>)
+X, y <span style="color: #666666">=</span> make_moons(n_samples<span style="color: #666666">=100</span>, noise<span style="color: #666666">=0.25</span>, random_state<span style="color: #666666">=53</span>)
+X_train, X_test, y_train, y_test <span style="color: #666666">=</span> train_test_split(X,y,random_state<span style="color: #666666">=0</span>)
+tree_clf <span style="color: #666666">=</span> DecisionTreeClassifier(max_depth<span style="color: #666666">=5</span>)
+tree_clf<span style="color: #666666">.</span>fit(X_train, y_train)
 
-<span style="color: #408080; font-style: italic"># To plot pretty figures</span>
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib</span>
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;axes.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">14</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;xtick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
-plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;ytick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
-
-
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn</span> <span style="color: #008000; font-weight: bold">import</span> datasets
-
-
-
-X1D <span style="color: #666666">=</span> np<span style="color: #666666">.</span>linspace(<span style="color: #666666">-4</span>, <span style="color: #666666">4</span>, <span style="color: #666666">9</span>)<span style="color: #666666">.</span>reshape(<span style="color: #666666">-1</span>, <span style="color: #666666">1</span>)
-X2D <span style="color: #666666">=</span> np<span style="color: #666666">.</span>c_[X1D, X1D<span style="color: #666666">**2</span>]
-y <span style="color: #666666">=</span> np<span style="color: #666666">.</span>array([<span style="color: #666666">0</span>, <span style="color: #666666">0</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">0</span>, <span style="color: #666666">0</span>])
-
-plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">4</span>))
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">121</span>)
-plt<span style="color: #666666">.</span>grid(<span style="color: #008000; font-weight: bold">True</span>, which<span style="color: #666666">=</span><span style="color: #BA2121">&#39;both&#39;</span>)
-plt<span style="color: #666666">.</span>axhline(y<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>plot(X1D[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==0</span>], np<span style="color: #666666">.</span>zeros(<span style="color: #666666">4</span>), <span style="color: #BA2121">&quot;bs&quot;</span>)
-plt<span style="color: #666666">.</span>plot(X1D[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==1</span>], np<span style="color: #666666">.</span>zeros(<span style="color: #666666">5</span>), <span style="color: #BA2121">&quot;g^&quot;</span>)
-plt<span style="color: #666666">.</span>gca()<span style="color: #666666">.</span>get_yaxis()<span style="color: #666666">.</span>set_ticks([])
-plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">r&quot;$x_1$&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>axis([<span style="color: #666666">-4.5</span>, <span style="color: #666666">4.5</span>, <span style="color: #666666">-0.2</span>, <span style="color: #666666">0.2</span>])
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">122</span>)
-plt<span style="color: #666666">.</span>grid(<span style="color: #008000; font-weight: bold">True</span>, which<span style="color: #666666">=</span><span style="color: #BA2121">&#39;both&#39;</span>)
-plt<span style="color: #666666">.</span>axhline(y<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>axvline(x<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>plot(X2D[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==0</span>], X2D[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==0</span>], <span style="color: #BA2121">&quot;bs&quot;</span>)
-plt<span style="color: #666666">.</span>plot(X2D[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==1</span>], X2D[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==1</span>], <span style="color: #BA2121">&quot;g^&quot;</span>)
-plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">r&quot;$x_1$&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">r&quot;$x_2$&quot;</span>, fontsize<span style="color: #666666">=20</span>, rotation<span style="color: #666666">=0</span>)
-plt<span style="color: #666666">.</span>gca()<span style="color: #666666">.</span>get_yaxis()<span style="color: #666666">.</span>set_ticks([<span style="color: #666666">0</span>, <span style="color: #666666">4</span>, <span style="color: #666666">8</span>, <span style="color: #666666">12</span>, <span style="color: #666666">16</span>])
-plt<span style="color: #666666">.</span>plot([<span style="color: #666666">-4.5</span>, <span style="color: #666666">4.5</span>], [<span style="color: #666666">6.5</span>, <span style="color: #666666">6.5</span>], <span style="color: #BA2121">&quot;r--&quot;</span>, linewidth<span style="color: #666666">=3</span>)
-plt<span style="color: #666666">.</span>axis([<span style="color: #666666">-4.5</span>, <span style="color: #666666">4.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">17</span>])
-plt<span style="color: #666666">.</span>subplots_adjust(right<span style="color: #666666">=1</span>)
-plt<span style="color: #666666">.</span>show()
+export_graphviz(
+    tree_clf,
+    out_file<span style="color: #666666">=</span><span style="color: #BA2121">&quot;DataFiles/moons.dot&quot;</span>,
+    rounded<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>,
+    filled<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>
+)
+cmd <span style="color: #666666">=</span> <span style="color: #BA2121">&#39;dot -Tpng DataFiles/moons.dot -o DataFiles/moons.png&#39;</span>
+os<span style="color: #666666">.</span>system(cmd)
 </pre>
 </div>
       </div>
@@ -1007,117 +956,231 @@ plt<span style="color: #666666">.</span>show()
 
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="the-equations">The equations </h2>
+<h2 id="other-ways-of-visualizing-the-trees">Other ways of visualizing the trees </h2>
 
-<p>Suppose we define a polynomial transformation of degree two only (we continue to live in a plane with \( x_i \) and \( y_i \) as variables)</p>
-$$
-z = \phi(x_i) =\left(x_i^2, y_i^2, \sqrt{2}x_iy_i\right).
-$$
+<p><b>Scikit-Learn</b> has also another way to visualize the trees which is very useful, here with the Iris data.</p>
 
-<p>With our new basis, the equations we solved earlier are basically the same, that is we have now (without the slack option for simplicity)</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{z}_i^T\boldsymbol{z}_j,
-$$
 
-<p>subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \), and for the support vectors</p>
-$$
-y_i(\boldsymbol{w}^T\boldsymbol{z}_i+b)= 1 \hspace{0.1cm}\forall i,
-$$
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> load_iris
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn</span> <span style="color: #008000; font-weight: bold">import</span> tree
+X, y <span style="color: #666666">=</span> load_iris(return_X_y<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>)
+tree_clf <span style="color: #666666">=</span> tree<span style="color: #666666">.</span>DecisionTreeClassifier()
+tree_clf <span style="color: #666666">=</span> tree_clf<span style="color: #666666">.</span>fit(X, y)
+<span style="color: #408080; font-style: italic"># and then plot the tree</span>
+tree<span style="color: #666666">.</span>plot_tree(tree_clf) 
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
 
-<p>from which we also find \( b \).
-To compute \( \boldsymbol{z}_i^T\boldsymbol{z}_j \) we define the kernel \( K(\boldsymbol{x}_i,\boldsymbol{x}_j) \) as
-</p>
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=\boldsymbol{z}_i^T\boldsymbol{z}_j= \phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j).
-$$
-
-<p>For the above example, the kernel reads</p>
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=[x_i^2, y_i^2, \sqrt{2}x_iy_i]^T\begin{bmatrix} x_j^2 \\ y_j^2 \\ \sqrt{2}x_jy_j \end{bmatrix}=x_i^2x_j^2+2x_ix_jy_iy_j+y_i^2y_j^2.
-$$
-
-<p>We note that this is nothing but the dot product of the two original
-vectors \( (\boldsymbol{x}_i^T\boldsymbol{x}_j)^2 \). Instead of thus computing the
-product in the Lagrangian of \( \boldsymbol{z}_i^T\boldsymbol{z}_j \) we simply compute
-the dot product \( (\boldsymbol{x}_i^T\boldsymbol{x}_j)^2 \).
-</p>
-
-<p>This leads to the so-called
-kernel trick and the result leads to the same as if we went through
-the trouble of performing the transformation
-\( \phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j) \) during the SVM calculations.
-</p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="the-problem-to-solve">The problem to solve </h2>
-<p>Using our definition of the kernel We can rewrite again the Lagrangian</p>
-$$
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\boldsymbol{x}_i^T\boldsymbol{z}_j,
-$$
+<h2 id="printing-out-as-text">Printing out as text </h2>
 
-<p>subject to the constraints \( \lambda_i\geq 0 \), \( \sum_i\lambda_iy_i=0 \) in terms of a convex optimization problem</p>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1K(\boldsymbol{x}_1,\boldsymbol{x}_1) & y_1y_2K(\boldsymbol{x}_1,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_1,\boldsymbol{x}_n) \\
-y_2y_1K(\boldsymbol{x}_2,\boldsymbol{x}_1) & y_2y_2(\boldsymbol{x}_2,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_2,\boldsymbol{x}_n) \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1K(\boldsymbol{x}_n,\boldsymbol{x}_1) & y_ny_2K(\boldsymbol{x}_n\boldsymbol{x}_2) & \dots & \dots & y_ny_nK(\boldsymbol{x}_n,\boldsymbol{x}_n) \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{1}\boldsymbol{\lambda}, 
-$$
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
-If we add the slack constants this leads to the additional constraint \( 0\leq \lambda_i \leq C \).
+<p>Alternatively, the tree can also be exported in textual format with the function exporttext.
+This method doesn&#8217;t require the installation of external libraries and is more compact:
 </p>
 
-<p>We can rewrite this (see the solutions below) in terms of a convex optimization problem of the type</p>
-$$
-\begin{align*}
-    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\boldsymbol{\lambda}^T\boldsymbol{P}\boldsymbol{\lambda}+\boldsymbol{q}^T\boldsymbol{\lambda},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{\lambda} \preceq \boldsymbol{h} \hspace{0.2cm} \wedge \boldsymbol{A}\boldsymbol{\lambda}=f.
-\end{align*}
-$$
 
-<p>Below we discuss how to solve these equations. Here we note that the matrix \( \boldsymbol{P} \) has matrix elements \( p_{ij}=y_iy_jK(\boldsymbol{x}_i,\boldsymbol{x}_j) \).
-Given a kernel \( K \) and the targets \( y_i \) this matrix is easy to set up. The constraint \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \) leads to \( f=0 \) and \( \boldsymbol{A}=\boldsymbol{y} \). How to set up the matrix \( \boldsymbol{G} \) is discussed later. Here note that the inequalities \( 0\leq \lambda_i \leq C \) can be split up into
-\( 0\leq \lambda_i \) and \( \lambda_i \leq C \). These two inequalities define then the matrix \( \boldsymbol{G} \) and the vector \( \boldsymbol{h} \).
-</p>
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> load_iris
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> export_text
+iris <span style="color: #666666">=</span> load_iris()
+decision_tree <span style="color: #666666">=</span> DecisionTreeClassifier(random_state<span style="color: #666666">=0</span>, max_depth<span style="color: #666666">=2</span>)
+decision_tree <span style="color: #666666">=</span> decision_tree<span style="color: #666666">.</span>fit(iris<span style="color: #666666">.</span>data, iris<span style="color: #666666">.</span>target)
+r <span style="color: #666666">=</span> export_text(decision_tree, feature_names<span style="color: #666666">=</span>iris[<span style="color: #BA2121">&#39;feature_names&#39;</span>])
+<span style="color: #008000">print</span>(r)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="different-kernels-and-mercer-s-theorem">Different kernels and Mercer's theorem </h2>
+<h2 id="algorithms-for-setting-up-decision-trees">Algorithms for Setting up Decision Trees </h2>
 
-<p>There are several popular kernels being used. These are</p>
+<p>Two algorithms stand out in the set up of decision trees:</p>
 <ol>
-<li> Linear: \( K(\boldsymbol{x},\boldsymbol{y})=\boldsymbol{x}^T\boldsymbol{y} \),</li>
-<li> Polynomial: \( K(\boldsymbol{x},\boldsymbol{y})=(\boldsymbol{x}^T\boldsymbol{y}+\gamma)^d \),</li>
-<li> Gaussian Radial Basis Function: \( K(\boldsymbol{x},\boldsymbol{y})=\exp{\left(-\gamma\vert\vert\boldsymbol{x}-\boldsymbol{y}\vert\vert^2\right)} \),</li>
-<li> Tanh: \( K(\boldsymbol{x},\boldsymbol{y})=\tanh{(\boldsymbol{x}^T\boldsymbol{y}+\gamma)} \),</li>
+<li> The CART (Classification And Regression Tree) algorithm for both classification and regression</li>
+<li> The ID3 algorithm based on the computation of the information gain for classification</li>
 </ol>
-<p>and many other ones.</p>
-
-<p>An important theorem for us is <a href="https://en.wikipedia.org/wiki/Mercer%27s_theorem" target="_blank">Mercer's
-theorem</a>.  The
-theorem states that if a kernel function \( K \) is symmetric, continuous
-and leads to a positive semi-definite matrix \( \boldsymbol{P} \) then there
-exists a function \( \phi \) that maps \( \boldsymbol{x}_i \) and \( \boldsymbol{x}_j \) into
-another space (possibly with much higher dimensions) such that
-</p>
-
-$$
-K(\boldsymbol{x}_i,\boldsymbol{x}_j)=\phi(\boldsymbol{x}_i)^T\phi(\boldsymbol{x}_j).
-$$
-
-<p>So you can use \( K \) as a kernel since you know \( \phi \) exists, even if
-you don&#8217;t know what \( \phi \) is. 
-</p>
-
-<p>Note that some frequently used kernels (such as the Sigmoid kernel)
-don&#8217;t respect all of Mercer&#8217;s conditions, yet they generally work well
-in practice.
+<p>We discuss both algorithms with applications here. The popular library
+<b>Scikit-Learn</b> uses the CART algorithm. For classification problems
+you can use either the <b>gini</b> index or the <b>entropy</b> to split a tree
+in two branches.
 </p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="the-moons-example">The moons example </h2>
+<h2 id="the-cart-algorithm-for-classification">The CART algorithm for Classification </h2>
+
+<p>For classification, the CART algorithm splits the data set in two subsets using a single feature \( k \) and a threshold \( t_k \).
+This could be for example a threshold set by a number below a certain circumference of a malign tumor.
+</p>
+
+<p>How do we find these two quantities?
+We search for the pair \( (k,t_k) \) that produces the purest subset using for example the <b>gini</b> factor \( G \).
+The cost function it tries to minimize is then
+</p>
+$$
+C(k,t_k) = \frac{m_{\mathrm{left}}}{m}G_{\mathrm{left}}+ \frac{m_{\mathrm{right}}}{m}G_{\mathrm{right}},
+$$
+
+<p>where \( G_{\mathrm{left/right}} \) measures the impurity of the left/right subset  and \( m_{\mathrm{left/right}} \)
+ is the number of instances in the left/right subset
+</p>
+
+<p>Once it has successfully split the training set in two, it splits the subsets using the same logic, then the subsubsets
+and so on, recursively. It stops recursing once it reaches the maximum depth (defined by the
+\( max\_depth \) hyperparameter), or if it cannot find a split that will reduce impurity. A few other
+hyperparameters control additional stopping conditions such as the \( min\_samples\_split \),
+\( min\_samples\_leaf \), \( min\_weight\_fraction\_leaf \), and \( max\_leaf\_nodes \).
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="the-cart-algorithm-for-regression">The CART algorithm for Regression </h2>
+
+<p>The CART algorithm for regression works is similar to the one for classification except that instead of trying to split the
+training set in a way that minimizes say the <b>gini</b> or <b>entropy</b> impurity, it now tries to split the training set in a way that minimizes our well-known mean-squared error (MSE). The cost function is now
+</p>
+$$
+C(k,t_k) = \frac{m_{\mathrm{left}}}{m}\mathrm{MSE}_{\mathrm{left}}+ \frac{m_{\mathrm{right}}}{m}\mathrm{MSE}_{\mathrm{right}}.
+$$
+
+<p>Here the MSE for a specific node is defined as</p>
+$$
+\mathrm{MSE}_{\mathrm{node}}=\frac{1}{m_\mathrm{node}}\sum_{i\in \mathrm{node}}(\overline{y}_{\mathrm{node}}-y_i)^2,
+$$
+
+<p>with</p>
+$$
+\overline{y}_{\mathrm{node}}=\frac{1}{m_\mathrm{node}}\sum_{i\in \mathrm{node}}y_i,
+$$
+
+<p>the mean value of all observations in a specific node.</p>
+
+<p>Without any regularization, the regression task for decision trees, 
+just like for classification tasks, is  prone to overfitting.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="why-binary-splits">Why binary splits? </h2>
+
+<p>It is custom to split to a tree uising binary splits. The reason is
+that multiway splits fragment the data too quickly, leaving
+insufficient data at the next level down.  Multiway splits can be
+achieved by a series of binary split and this is normally preferred.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="computing-a-tree-using-the-gini-index">Computing a Tree using the Gini Index </h2>
+
+<p>Consider the following example with attributes/features and two
+possible outcomes (classes)  for each attribute.  Assume we wish to find some
+correlations between the average grade of a student as function of the
+number of hours studied and hours slept. We want also to correlate the
+grade in a given course with the general trend, whether the students
+recently has gotten grades below average or above.
+</p>
+
+<p>We have three features/attributes</p>
+<ol>
+<li> Trend of average grades before present course, classified as either below  or above  the average grade of the whole class</li>  
+<li> The number of hours studies, classified again as either higher (more than 3 hours per day) or lower . Here we have used a standard for one \( ECTS \) which is scaled to 25-30 hours of work for a semester which lasts 18 weeks, with 15 weeks of lectures and 3 weeks for exams, assuming a total of 30 ECTS per semester.</li> 
+<li> The number of hours slept as high for more than \( 8 \) hours and below  for less than 8 hours of sleep, classified again as either high or low</li>
+<li> The final grade whether it is above or below average</li>
+</ol>
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="the-table">The Table </h2>
+
+<table class="dotable" border="1">
+<thead>
+<tr><th align="center">Grade Trend</th> <th align="center">Hours slept</th> <th align="center">Hours Studied</th> <th align="center">Grade</th> </tr>
+</thead>
+<tbody>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   High           </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Below          </td> <td align="center">   Low            </td> <td align="center">   High             </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   Low            </td> <td align="center">   Low              </td> <td align="center">   Below    </td> </tr>
+<tr><td align="center">   Above          </td> <td align="center">   High           </td> <td align="center">   High             </td> <td align="center">   Above    </td> </tr>
+</tbody>
+</table>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="computing-the-various-gini-indices">Computing the various Gini Indices </h2>
+
+<p>In computations we will translate all classes into numbers. Being
+these binary classes, they can easily be split into ones and zeros.
+</p>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Gini index for Average trend</b>
+<p>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="computing-the-various-gini-indices-hours-slept">Computing the various Gini Indices, Hours slept </h2>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Gini index for hour slept</b>
+<p>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="computing-the-various-gini-indices-hours-studied">Computing the various Gini Indices, Hours studied </h2>
+
+<div class="alert alert-block alert-block alert-text-normal">
+<b>Gini index for hour studied</b>
+<p>
+<a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf" target="_blank">See handwritten notes November 3</a>
+</div>
+
+
+<p>For final tree, see the above handwritten notes</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="a-possible-code-using-scikit-learn">A possible code using Scikit-Learn </h2>
+
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
 <div class="cell border-box-sizing code_cell rendered">
@@ -1125,193 +1188,994 @@ in practice.
     <div class="inner_cell">
       <div class="input_area">
         <div class="highlight" style="background: #f8f8f8">
-  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">__future__</span> <span style="color: #008000; font-weight: bold">import</span> division, print_function, unicode_literals
-
+  <pre style="line-height: 125%;"><span style="color: #408080; font-style: italic"># Common imports</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
-np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>seed(<span style="color: #666666">42</span>)
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">pandas</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">pd</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span> train_test_split
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> export_graphviz
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> StandardScaler, OneHotEncoder
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.compose</span> <span style="color: #008000; font-weight: bold">import</span> ColumnTransformer
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">IPython.display</span> <span style="color: #008000; font-weight: bold">import</span> Image 
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">pydot</span> <span style="color: #008000; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">os</span>
 
+<span style="color: #408080; font-style: italic"># Where to save the figures and data files</span>
+PROJECT_ROOT_DIR <span style="color: #666666">=</span> <span style="color: #BA2121">&quot;Results&quot;</span>
+FIGURE_ID <span style="color: #666666">=</span> <span style="color: #BA2121">&quot;Results/FigureFiles&quot;</span>
+DATA_ID <span style="color: #666666">=</span> <span style="color: #BA2121">&quot;DataFiles/&quot;</span>
+
+<span style="color: #008000; font-weight: bold">if</span> <span style="color: #AA22FF; font-weight: bold">not</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>exists(PROJECT_ROOT_DIR):
+    os<span style="color: #666666">.</span>mkdir(PROJECT_ROOT_DIR)
+
+<span style="color: #008000; font-weight: bold">if</span> <span style="color: #AA22FF; font-weight: bold">not</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>exists(FIGURE_ID):
+    os<span style="color: #666666">.</span>makedirs(FIGURE_ID)
+
+<span style="color: #008000; font-weight: bold">if</span> <span style="color: #AA22FF; font-weight: bold">not</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>exists(DATA_ID):
+    os<span style="color: #666666">.</span>makedirs(DATA_ID)
+
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">image_path</span>(fig_id):
+    <span style="color: #008000; font-weight: bold">return</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>join(FIGURE_ID, fig_id)
+
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">data_path</span>(dat_id):
+    <span style="color: #008000; font-weight: bold">return</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>join(DATA_ID, dat_id)
+
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">save_fig</span>(fig_id):
+    plt<span style="color: #666666">.</span>savefig(image_path(fig_id) <span style="color: #666666">+</span> <span style="color: #BA2121">&quot;.png&quot;</span>, <span style="color: #008000">format</span><span style="color: #666666">=</span><span style="color: #BA2121">&#39;png&#39;</span>)
+
+infile <span style="color: #666666">=</span> <span style="color: #008000">open</span>(data_path(<span style="color: #BA2121">&quot;grades.csv&quot;</span>),<span style="color: #BA2121">&#39;r&#39;</span>)
+
+<span style="color: #408080; font-style: italic"># Read the experimental data with Pandas</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">IPython.display</span> <span style="color: #008000; font-weight: bold">import</span> display
+grades <span style="color: #666666">=</span> pd<span style="color: #666666">.</span>read_csv(infile)
+grades <span style="color: #666666">=</span> pd<span style="color: #666666">.</span>DataFrame(grades)
+display(grades)
+<span style="color: #408080; font-style: italic"># Features and targets</span>
+X <span style="color: #666666">=</span> grades<span style="color: #666666">.</span>loc[:, grades<span style="color: #666666">.</span>columns <span style="color: #666666">!=</span> <span style="color: #BA2121">&#39;Grade&#39;</span>]<span style="color: #666666">.</span>values
+y <span style="color: #666666">=</span> grades<span style="color: #666666">.</span>loc[:, grades<span style="color: #666666">.</span>columns <span style="color: #666666">==</span> <span style="color: #BA2121">&#39;Grade&#39;</span>]<span style="color: #666666">.</span>values
+<span style="color: #008000">print</span>(X)
+<span style="color: #408080; font-style: italic"># Then do a Classification tree</span>
+tree_clf <span style="color: #666666">=</span> DecisionTreeClassifier(max_depth<span style="color: #666666">=2</span>)
+tree_clf<span style="color: #666666">.</span>fit(X, y)
+<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;Train set accuracy with Decision Tree: </span><span style="color: #BB6688; font-weight: bold">{:.2f}</span><span style="color: #BA2121">&quot;</span><span style="color: #666666">.</span>format(tree_clf<span style="color: #666666">.</span>score(X,y)))
+<span style="color: #408080; font-style: italic">#transfer to a decision tree graph</span>
+export_graphviz(
+    tree_clf,
+    out_file<span style="color: #666666">=</span><span style="color: #BA2121">&quot;DataFiles/grade.dot&quot;</span>,
+    rounded<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>,
+    filled<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>
+)
+cmd <span style="color: #666666">=</span> <span style="color: #BA2121">&#39;dot -Tpng DataFiles/grade.dot -o DataFiles/grades.png&#39;</span>
+os<span style="color: #666666">.</span>system(cmd)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="further-example-computing-the-gini-index">Further example: Computing the Gini index </h2>
+
+<p>The next example we will look at is a classical one in many Machine
+Learning applications. Based on various meteorological features, we
+have several so-called attributes which decide whether we at the end
+will do some outdoor activity like skiing, going for a bike ride etc
+etc.  The table here contains the feautures <b>outlook</b>, <b>temperature</b>,
+<b>humidity</b> and <b>wind</b>.  The target or output is whether we ride
+(True=1) or whether we do something else that day (False=0). The
+attributes for each feature are then sunny, overcast and rain for the
+outlook, hot, cold and mild for temperature, high and normal for
+humidity and weak and strong for wind.
+</p>
+
+<p>The table here summarizes the various attributes and</p>
+<table class="dotable" border="1">
+<thead>
+<tr><th align="center">Day</th> <th align="center">Outlook </th> <th align="center">Temperature</th> <th align="center">Humidity</th> <th align="center"> Wind </th> <th align="center">Ride</th> </tr>
+</thead>
+<tbody>
+<tr><td align="center">   1      </td> <td align="center">   Sunny       </td> <td align="center">   Hot            </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   0       </td> </tr>
+<tr><td align="center">   2      </td> <td align="center">   Sunny       </td> <td align="center">   Hot            </td> <td align="center">   High        </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   3      </td> <td align="center">   Overcast    </td> <td align="center">   Hot            </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   4      </td> <td align="center">   Rain        </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   5      </td> <td align="center">   Rain        </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   6      </td> <td align="center">   Rain        </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Strong    </td> <td align="center">   0       </td> </tr>
+<tr><td align="center">   7      </td> <td align="center">   Overcast    </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   8      </td> <td align="center">   Sunny       </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Weak      </td> <td align="center">   0       </td> </tr>
+<tr><td align="center">   9      </td> <td align="center">   Sunny       </td> <td align="center">   Cool           </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   10     </td> <td align="center">   Rain        </td> <td align="center">   Mild           </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   11     </td> <td align="center">   Sunny       </td> <td align="center">   Mild           </td> <td align="center">   Normal      </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   12     </td> <td align="center">   Overcast    </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Strong    </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   13     </td> <td align="center">   Overcast    </td> <td align="center">   Hot            </td> <td align="center">   Normal      </td> <td align="center">   Weak      </td> <td align="center">   1       </td> </tr>
+<tr><td align="center">   14     </td> <td align="center">   Rain        </td> <td align="center">   Mild           </td> <td align="center">   High        </td> <td align="center">   Strong    </td> <td align="center">   0       </td> </tr>
+</tbody>
+</table>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="simple-python-code-to-read-in-data-and-perform-classification">Simple Python Code to read in Data and perform Classification </h2>
+
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #408080; font-style: italic"># Common imports</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">pandas</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">pd</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span> train_test_split
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> export_graphviz
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> StandardScaler, OneHotEncoder
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.compose</span> <span style="color: #008000; font-weight: bold">import</span> ColumnTransformer
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">IPython.display</span> <span style="color: #008000; font-weight: bold">import</span> Image 
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">pydot</span> <span style="color: #008000; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">os</span>
+
+<span style="color: #408080; font-style: italic"># Where to save the figures and data files</span>
+PROJECT_ROOT_DIR <span style="color: #666666">=</span> <span style="color: #BA2121">&quot;Results&quot;</span>
+FIGURE_ID <span style="color: #666666">=</span> <span style="color: #BA2121">&quot;Results/FigureFiles&quot;</span>
+DATA_ID <span style="color: #666666">=</span> <span style="color: #BA2121">&quot;DataFiles/&quot;</span>
+
+<span style="color: #008000; font-weight: bold">if</span> <span style="color: #AA22FF; font-weight: bold">not</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>exists(PROJECT_ROOT_DIR):
+    os<span style="color: #666666">.</span>mkdir(PROJECT_ROOT_DIR)
+
+<span style="color: #008000; font-weight: bold">if</span> <span style="color: #AA22FF; font-weight: bold">not</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>exists(FIGURE_ID):
+    os<span style="color: #666666">.</span>makedirs(FIGURE_ID)
+
+<span style="color: #008000; font-weight: bold">if</span> <span style="color: #AA22FF; font-weight: bold">not</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>exists(DATA_ID):
+    os<span style="color: #666666">.</span>makedirs(DATA_ID)
+
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">image_path</span>(fig_id):
+    <span style="color: #008000; font-weight: bold">return</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>join(FIGURE_ID, fig_id)
+
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">data_path</span>(dat_id):
+    <span style="color: #008000; font-weight: bold">return</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>join(DATA_ID, dat_id)
+
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">save_fig</span>(fig_id):
+    plt<span style="color: #666666">.</span>savefig(image_path(fig_id) <span style="color: #666666">+</span> <span style="color: #BA2121">&quot;.png&quot;</span>, <span style="color: #008000">format</span><span style="color: #666666">=</span><span style="color: #BA2121">&#39;png&#39;</span>)
+
+infile <span style="color: #666666">=</span> <span style="color: #008000">open</span>(data_path(<span style="color: #BA2121">&quot;rideclass.csv&quot;</span>),<span style="color: #BA2121">&#39;r&#39;</span>)
+
+<span style="color: #408080; font-style: italic"># Read the experimental data with Pandas</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">IPython.display</span> <span style="color: #008000; font-weight: bold">import</span> display
+ridedata <span style="color: #666666">=</span> pd<span style="color: #666666">.</span>read_csv(infile,names <span style="color: #666666">=</span> (<span style="color: #BA2121">&#39;Outlook&#39;</span>,<span style="color: #BA2121">&#39;Temperature&#39;</span>,<span style="color: #BA2121">&#39;Humidity&#39;</span>,<span style="color: #BA2121">&#39;Wind&#39;</span>,<span style="color: #BA2121">&#39;Ride&#39;</span>))
+ridedata <span style="color: #666666">=</span> pd<span style="color: #666666">.</span>DataFrame(ridedata)
+
+<span style="color: #408080; font-style: italic"># Features and targets</span>
+X <span style="color: #666666">=</span> ridedata<span style="color: #666666">.</span>loc[:, ridedata<span style="color: #666666">.</span>columns <span style="color: #666666">!=</span> <span style="color: #BA2121">&#39;Ride&#39;</span>]<span style="color: #666666">.</span>values
+y <span style="color: #666666">=</span> ridedata<span style="color: #666666">.</span>loc[:, ridedata<span style="color: #666666">.</span>columns <span style="color: #666666">==</span> <span style="color: #BA2121">&#39;Ride&#39;</span>]<span style="color: #666666">.</span>values
+
+<span style="color: #408080; font-style: italic"># Create the encoder.</span>
+encoder <span style="color: #666666">=</span> OneHotEncoder(handle_unknown<span style="color: #666666">=</span><span style="color: #BA2121">&quot;ignore&quot;</span>)
+<span style="color: #408080; font-style: italic"># Assume for simplicity all features are categorical.</span>
+encoder<span style="color: #666666">.</span>fit(X)    
+<span style="color: #408080; font-style: italic"># Apply the encoder.</span>
+X <span style="color: #666666">=</span> encoder<span style="color: #666666">.</span>transform(X)
+<span style="color: #008000">print</span>(X)
+<span style="color: #408080; font-style: italic"># Then do a Classification tree</span>
+tree_clf <span style="color: #666666">=</span> DecisionTreeClassifier(max_depth<span style="color: #666666">=2</span>)
+tree_clf<span style="color: #666666">.</span>fit(X, y)
+<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;Train set accuracy with Decision Tree: </span><span style="color: #BB6688; font-weight: bold">{:.2f}</span><span style="color: #BA2121">&quot;</span><span style="color: #666666">.</span>format(tree_clf<span style="color: #666666">.</span>score(X,y)))
+<span style="color: #408080; font-style: italic">#transfer to a decision tree graph</span>
+export_graphviz(
+    tree_clf,
+    out_file<span style="color: #666666">=</span><span style="color: #BA2121">&quot;DataFiles/ride.dot&quot;</span>,
+    rounded<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>,
+    filled<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>
+)
+cmd <span style="color: #666666">=</span> <span style="color: #BA2121">&#39;dot -Tpng DataFiles/cancer.dot -o DataFiles/cancer.png&#39;</span>
+os<span style="color: #666666">.</span>system(cmd)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="computing-the-gini-factor">Computing the Gini Factor  </h2>
+
+<p>The above functions (gini, entropy and misclassification error) are
+important components of the so-called CART algorithm. We will discuss
+this algorithm below after we have discussed the information gain
+algorithm ID3.
+</p>
+
+<p>In the example here we have converted all our attributes into numerical values \( 0,1,2 \) etc.</p>
+
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #408080; font-style: italic"># Split a dataset based on an attribute and an attribute value</span>
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">test_split</span>(index, value, dataset):
+	left, right <span style="color: #666666">=</span> <span style="color: #008000">list</span>(), <span style="color: #008000">list</span>()
+	<span style="color: #008000; font-weight: bold">for</span> row <span style="color: #AA22FF; font-weight: bold">in</span> dataset:
+		<span style="color: #008000; font-weight: bold">if</span> row[index] <span style="color: #666666">&lt;</span> value:
+			left<span style="color: #666666">.</span>append(row)
+		<span style="color: #008000; font-weight: bold">else</span>:
+			right<span style="color: #666666">.</span>append(row)
+	<span style="color: #008000; font-weight: bold">return</span> left, right
+ 
+<span style="color: #408080; font-style: italic"># Calculate the Gini index for a split dataset</span>
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">gini_index</span>(groups, classes):
+	<span style="color: #408080; font-style: italic"># count all samples at split point</span>
+	n_instances <span style="color: #666666">=</span> <span style="color: #008000">float</span>(<span style="color: #008000">sum</span>([<span style="color: #008000">len</span>(group) <span style="color: #008000; font-weight: bold">for</span> group <span style="color: #AA22FF; font-weight: bold">in</span> groups]))
+	<span style="color: #408080; font-style: italic"># sum weighted Gini index for each group</span>
+	gini <span style="color: #666666">=</span> <span style="color: #666666">0.0</span>
+	<span style="color: #008000; font-weight: bold">for</span> group <span style="color: #AA22FF; font-weight: bold">in</span> groups:
+		size <span style="color: #666666">=</span> <span style="color: #008000">float</span>(<span style="color: #008000">len</span>(group))
+		<span style="color: #408080; font-style: italic"># avoid divide by zero</span>
+		<span style="color: #008000; font-weight: bold">if</span> size <span style="color: #666666">==</span> <span style="color: #666666">0</span>:
+			<span style="color: #008000; font-weight: bold">continue</span>
+		score <span style="color: #666666">=</span> <span style="color: #666666">0.0</span>
+		<span style="color: #408080; font-style: italic"># score the group based on the score for each class</span>
+		<span style="color: #008000; font-weight: bold">for</span> class_val <span style="color: #AA22FF; font-weight: bold">in</span> classes:
+			p <span style="color: #666666">=</span> [row[<span style="color: #666666">-1</span>] <span style="color: #008000; font-weight: bold">for</span> row <span style="color: #AA22FF; font-weight: bold">in</span> group]<span style="color: #666666">.</span>count(class_val) <span style="color: #666666">/</span> size
+			score <span style="color: #666666">+=</span> p <span style="color: #666666">*</span> p
+		<span style="color: #408080; font-style: italic"># weight the group score by its relative size</span>
+		gini <span style="color: #666666">+=</span> (<span style="color: #666666">1.0</span> <span style="color: #666666">-</span> score) <span style="color: #666666">*</span> (size <span style="color: #666666">/</span> n_instances)
+	<span style="color: #008000; font-weight: bold">return</span> gini
+
+<span style="color: #408080; font-style: italic"># Select the best split point for a dataset</span>
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">get_split</span>(dataset):
+	class_values <span style="color: #666666">=</span> <span style="color: #008000">list</span>(<span style="color: #008000">set</span>(row[<span style="color: #666666">-1</span>] <span style="color: #008000; font-weight: bold">for</span> row <span style="color: #AA22FF; font-weight: bold">in</span> dataset))
+	b_index, b_value, b_score, b_groups <span style="color: #666666">=</span> <span style="color: #666666">999</span>, <span style="color: #666666">999</span>, <span style="color: #666666">999</span>, <span style="color: #008000; font-weight: bold">None</span>
+	<span style="color: #008000; font-weight: bold">for</span> index <span style="color: #AA22FF; font-weight: bold">in</span> <span style="color: #008000">range</span>(<span style="color: #008000">len</span>(dataset[<span style="color: #666666">0</span>])<span style="color: #666666">-1</span>):
+		<span style="color: #008000; font-weight: bold">for</span> row <span style="color: #AA22FF; font-weight: bold">in</span> dataset:
+			groups <span style="color: #666666">=</span> test_split(index, row[index], dataset)
+			gini <span style="color: #666666">=</span> gini_index(groups, class_values)
+			<span style="color: #008000">print</span>(<span style="color: #BA2121">&#39;X</span><span style="color: #BB6688; font-weight: bold">%d</span><span style="color: #BA2121"> &lt; </span><span style="color: #BB6688; font-weight: bold">%.3f</span><span style="color: #BA2121"> Gini=</span><span style="color: #BB6688; font-weight: bold">%.3f</span><span style="color: #BA2121">&#39;</span> <span style="color: #666666">%</span> ((index<span style="color: #666666">+1</span>), row[index], gini))
+			<span style="color: #008000; font-weight: bold">if</span> gini <span style="color: #666666">&lt;</span> b_score:
+				b_index, b_value, b_score, b_groups <span style="color: #666666">=</span> index, row[index], gini, groups
+	<span style="color: #008000; font-weight: bold">return</span> {<span style="color: #BA2121">&#39;index&#39;</span>:b_index, <span style="color: #BA2121">&#39;value&#39;</span>:b_value, <span style="color: #BA2121">&#39;groups&#39;</span>:b_groups}
+ 
+dataset <span style="color: #666666">=</span> [[<span style="color: #666666">0</span>,<span style="color: #666666">0</span>,<span style="color: #666666">0</span>,<span style="color: #666666">0</span>,<span style="color: #666666">0</span>],
+            [<span style="color: #666666">0</span>,<span style="color: #666666">0</span>,<span style="color: #666666">0</span>,<span style="color: #666666">1</span>,<span style="color: #666666">1</span>],
+            [<span style="color: #666666">1</span>,<span style="color: #666666">0</span>,<span style="color: #666666">0</span>,<span style="color: #666666">0</span>,<span style="color: #666666">1</span>],
+            [<span style="color: #666666">2</span>,<span style="color: #666666">1</span>,<span style="color: #666666">0</span>,<span style="color: #666666">0</span>,<span style="color: #666666">1</span>],
+            [<span style="color: #666666">2</span>,<span style="color: #666666">2</span>,<span style="color: #666666">1</span>,<span style="color: #666666">0</span>,<span style="color: #666666">1</span>],
+            [<span style="color: #666666">2</span>,<span style="color: #666666">2</span>,<span style="color: #666666">1</span>,<span style="color: #666666">1</span>,<span style="color: #666666">0</span>],
+            [<span style="color: #666666">1</span>,<span style="color: #666666">2</span>,<span style="color: #666666">1</span>,<span style="color: #666666">1</span>,<span style="color: #666666">1</span>],
+            [<span style="color: #666666">0</span>,<span style="color: #666666">1</span>,<span style="color: #666666">0</span>,<span style="color: #666666">0</span>,<span style="color: #666666">0</span>],
+            [<span style="color: #666666">0</span>,<span style="color: #666666">2</span>,<span style="color: #666666">1</span>,<span style="color: #666666">0</span>,<span style="color: #666666">1</span>],
+            [<span style="color: #666666">2</span>,<span style="color: #666666">1</span>,<span style="color: #666666">1</span>,<span style="color: #666666">0</span>,<span style="color: #666666">1</span>],
+            [<span style="color: #666666">0</span>,<span style="color: #666666">1</span>,<span style="color: #666666">1</span>,<span style="color: #666666">1</span>,<span style="color: #666666">1</span>],
+            [<span style="color: #666666">1</span>,<span style="color: #666666">1</span>,<span style="color: #666666">0</span>,<span style="color: #666666">1</span>,<span style="color: #666666">1</span>],
+            [<span style="color: #666666">1</span>,<span style="color: #666666">0</span>,<span style="color: #666666">1</span>,<span style="color: #666666">0</span>,<span style="color: #666666">1</span>],
+            [<span style="color: #666666">2</span>,<span style="color: #666666">1</span>,<span style="color: #666666">0</span>,<span style="color: #666666">1</span>,<span style="color: #666666">0</span>]]
+
+split <span style="color: #666666">=</span> get_split(dataset)
+<span style="color: #008000">print</span>(<span style="color: #BA2121">&#39;Split: [X</span><span style="color: #BB6688; font-weight: bold">%d</span><span style="color: #BA2121"> &lt; </span><span style="color: #BB6688; font-weight: bold">%.3f</span><span style="color: #BA2121">]&#39;</span> <span style="color: #666666">%</span> ((split[<span style="color: #BA2121">&#39;index&#39;</span>]<span style="color: #666666">+1</span>), split[<span style="color: #BA2121">&#39;value&#39;</span>]))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="regression-trees">Regression trees </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #408080; font-style: italic"># Quadratic training set + noise</span>
+np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>seed(<span style="color: #666666">42</span>)
+m <span style="color: #666666">=</span> <span style="color: #666666">200</span>
+X <span style="color: #666666">=</span> np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>rand(m, <span style="color: #666666">1</span>)
+y <span style="color: #666666">=</span> <span style="color: #666666">4</span> <span style="color: #666666">*</span> (X <span style="color: #666666">-</span> <span style="color: #666666">0.5</span>) <span style="color: #666666">**</span> <span style="color: #666666">2</span>
+y <span style="color: #666666">=</span> y <span style="color: #666666">+</span> np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>randn(m, <span style="color: #666666">1</span>) <span style="color: #666666">/</span> <span style="color: #666666">10</span>
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeRegressor
+
+tree_reg <span style="color: #666666">=</span> DecisionTreeRegressor(max_depth<span style="color: #666666">=2</span>, random_state<span style="color: #666666">=42</span>)
+tree_reg<span style="color: #666666">.</span>fit(X, y)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="final-regressor-code">Final regressor code </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeRegressor
+
+tree_reg1 <span style="color: #666666">=</span> DecisionTreeRegressor(random_state<span style="color: #666666">=42</span>, max_depth<span style="color: #666666">=2</span>)
+tree_reg2 <span style="color: #666666">=</span> DecisionTreeRegressor(random_state<span style="color: #666666">=42</span>, max_depth<span style="color: #666666">=3</span>)
+tree_reg1<span style="color: #666666">.</span>fit(X, y)
+tree_reg2<span style="color: #666666">.</span>fit(X, y)
+
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">plot_regression_predictions</span>(tree_reg, X, y, axes<span style="color: #666666">=</span>[<span style="color: #666666">0</span>, <span style="color: #666666">1</span>, <span style="color: #666666">-0.2</span>, <span style="color: #666666">1</span>], ylabel<span style="color: #666666">=</span><span style="color: #BA2121">&quot;$y$&quot;</span>):
+    x1 <span style="color: #666666">=</span> np<span style="color: #666666">.</span>linspace(axes[<span style="color: #666666">0</span>], axes[<span style="color: #666666">1</span>], <span style="color: #666666">500</span>)<span style="color: #666666">.</span>reshape(<span style="color: #666666">-1</span>, <span style="color: #666666">1</span>)
+    y_pred <span style="color: #666666">=</span> tree_reg<span style="color: #666666">.</span>predict(x1)
+    plt<span style="color: #666666">.</span>axis(axes)
+    plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">&quot;$x_1$&quot;</span>, fontsize<span style="color: #666666">=18</span>)
+    <span style="color: #008000; font-weight: bold">if</span> ylabel:
+        plt<span style="color: #666666">.</span>ylabel(ylabel, fontsize<span style="color: #666666">=18</span>, rotation<span style="color: #666666">=0</span>)
+    plt<span style="color: #666666">.</span>plot(X, y, <span style="color: #BA2121">&quot;b.&quot;</span>)
+    plt<span style="color: #666666">.</span>plot(x1, y_pred, <span style="color: #BA2121">&quot;r.-&quot;</span>, linewidth<span style="color: #666666">=2</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">r&quot;$\hat</span><span style="color: #BB6688; font-weight: bold">{y}</span><span style="color: #BA2121">$&quot;</span>)
+
+plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">4</span>))
+plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">121</span>)
+plot_regression_predictions(tree_reg1, X, y)
+<span style="color: #008000; font-weight: bold">for</span> split, style <span style="color: #AA22FF; font-weight: bold">in</span> ((<span style="color: #666666">0.1973</span>, <span style="color: #BA2121">&quot;k-&quot;</span>), (<span style="color: #666666">0.0917</span>, <span style="color: #BA2121">&quot;k--&quot;</span>), (<span style="color: #666666">0.7718</span>, <span style="color: #BA2121">&quot;k--&quot;</span>)):
+    plt<span style="color: #666666">.</span>plot([split, split], [<span style="color: #666666">-0.2</span>, <span style="color: #666666">1</span>], style, linewidth<span style="color: #666666">=2</span>)
+plt<span style="color: #666666">.</span>text(<span style="color: #666666">0.21</span>, <span style="color: #666666">0.65</span>, <span style="color: #BA2121">&quot;Depth=0&quot;</span>, fontsize<span style="color: #666666">=15</span>)
+plt<span style="color: #666666">.</span>text(<span style="color: #666666">0.01</span>, <span style="color: #666666">0.2</span>, <span style="color: #BA2121">&quot;Depth=1&quot;</span>, fontsize<span style="color: #666666">=13</span>)
+plt<span style="color: #666666">.</span>text(<span style="color: #666666">0.65</span>, <span style="color: #666666">0.8</span>, <span style="color: #BA2121">&quot;Depth=1&quot;</span>, fontsize<span style="color: #666666">=13</span>)
+plt<span style="color: #666666">.</span>legend(loc<span style="color: #666666">=</span><span style="color: #BA2121">&quot;upper center&quot;</span>, fontsize<span style="color: #666666">=18</span>)
+plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">&quot;max_depth=2&quot;</span>, fontsize<span style="color: #666666">=14</span>)
+
+plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">122</span>)
+plot_regression_predictions(tree_reg2, X, y, ylabel<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">None</span>)
+<span style="color: #008000; font-weight: bold">for</span> split, style <span style="color: #AA22FF; font-weight: bold">in</span> ((<span style="color: #666666">0.1973</span>, <span style="color: #BA2121">&quot;k-&quot;</span>), (<span style="color: #666666">0.0917</span>, <span style="color: #BA2121">&quot;k--&quot;</span>), (<span style="color: #666666">0.7718</span>, <span style="color: #BA2121">&quot;k--&quot;</span>)):
+    plt<span style="color: #666666">.</span>plot([split, split], [<span style="color: #666666">-0.2</span>, <span style="color: #666666">1</span>], style, linewidth<span style="color: #666666">=2</span>)
+<span style="color: #008000; font-weight: bold">for</span> split <span style="color: #AA22FF; font-weight: bold">in</span> (<span style="color: #666666">0.0458</span>, <span style="color: #666666">0.1298</span>, <span style="color: #666666">0.2873</span>, <span style="color: #666666">0.9040</span>):
+    plt<span style="color: #666666">.</span>plot([split, split], [<span style="color: #666666">-0.2</span>, <span style="color: #666666">1</span>], <span style="color: #BA2121">&quot;k:&quot;</span>, linewidth<span style="color: #666666">=1</span>)
+plt<span style="color: #666666">.</span>text(<span style="color: #666666">0.3</span>, <span style="color: #666666">0.5</span>, <span style="color: #BA2121">&quot;Depth=2&quot;</span>, fontsize<span style="color: #666666">=13</span>)
+plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">&quot;max_depth=3&quot;</span>, fontsize<span style="color: #666666">=14</span>)
+
+plt<span style="color: #666666">.</span>show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;">tree_reg1 <span style="color: #666666">=</span> DecisionTreeRegressor(random_state<span style="color: #666666">=42</span>)
+tree_reg2 <span style="color: #666666">=</span> DecisionTreeRegressor(random_state<span style="color: #666666">=42</span>, min_samples_leaf<span style="color: #666666">=10</span>)
+tree_reg1<span style="color: #666666">.</span>fit(X, y)
+tree_reg2<span style="color: #666666">.</span>fit(X, y)
+
+x1 <span style="color: #666666">=</span> np<span style="color: #666666">.</span>linspace(<span style="color: #666666">0</span>, <span style="color: #666666">1</span>, <span style="color: #666666">500</span>)<span style="color: #666666">.</span>reshape(<span style="color: #666666">-1</span>, <span style="color: #666666">1</span>)
+y_pred1 <span style="color: #666666">=</span> tree_reg1<span style="color: #666666">.</span>predict(x1)
+y_pred2 <span style="color: #666666">=</span> tree_reg2<span style="color: #666666">.</span>predict(x1)
+
+plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">4</span>))
+
+plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">121</span>)
+plt<span style="color: #666666">.</span>plot(X, y, <span style="color: #BA2121">&quot;b.&quot;</span>)
+plt<span style="color: #666666">.</span>plot(x1, y_pred1, <span style="color: #BA2121">&quot;r.-&quot;</span>, linewidth<span style="color: #666666">=2</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">r&quot;$\hat</span><span style="color: #BB6688; font-weight: bold">{y}</span><span style="color: #BA2121">$&quot;</span>)
+plt<span style="color: #666666">.</span>axis([<span style="color: #666666">0</span>, <span style="color: #666666">1</span>, <span style="color: #666666">-0.2</span>, <span style="color: #666666">1.1</span>])
+plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">&quot;$x_1$&quot;</span>, fontsize<span style="color: #666666">=18</span>)
+plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">&quot;$y$&quot;</span>, fontsize<span style="color: #666666">=18</span>, rotation<span style="color: #666666">=0</span>)
+plt<span style="color: #666666">.</span>legend(loc<span style="color: #666666">=</span><span style="color: #BA2121">&quot;upper center&quot;</span>, fontsize<span style="color: #666666">=18</span>)
+plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">&quot;No restrictions&quot;</span>, fontsize<span style="color: #666666">=14</span>)
+
+plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">122</span>)
+plt<span style="color: #666666">.</span>plot(X, y, <span style="color: #BA2121">&quot;b.&quot;</span>)
+plt<span style="color: #666666">.</span>plot(x1, y_pred2, <span style="color: #BA2121">&quot;r.-&quot;</span>, linewidth<span style="color: #666666">=2</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">r&quot;$\hat</span><span style="color: #BB6688; font-weight: bold">{y}</span><span style="color: #BA2121">$&quot;</span>)
+plt<span style="color: #666666">.</span>axis([<span style="color: #666666">0</span>, <span style="color: #666666">1</span>, <span style="color: #666666">-0.2</span>, <span style="color: #666666">1.1</span>])
+plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">&quot;$x_1$&quot;</span>, fontsize<span style="color: #666666">=18</span>)
+plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">&quot;min_samples_leaf=</span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">&quot;</span><span style="color: #666666">.</span>format(tree_reg2<span style="color: #666666">.</span>min_samples_leaf), fontsize<span style="color: #666666">=14</span>)
+
+plt<span style="color: #666666">.</span>show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="pros-and-cons-of-trees-pros">Pros and cons of trees, pros </h2>
+
+<ul>
+<li> White box, easy to interpret model. Some people believe that decision trees more closely mirror human decision-making than do the regression and classification approaches discussed earlier (think of support vector machines)</li>
+<li> Trees are very easy to explain to people. In fact, they are even easier to explain than linear regression!</li>
+<li> No feature normalization needed</li>
+<li> Tree models can handle both continuous and categorical data (Classification and Regression Trees)</li>
+<li> Can model nonlinear relationships</li>
+<li> Can model interactions between the different descriptive features</li>
+<li> Trees can be displayed graphically, and are easily interpreted even by a non-expert (especially if they are small)</li>
+</ul>
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="disadvantages">Disadvantages </h2>
+
+<ul>
+<li> Unfortunately, trees generally do not have the same level of predictive accuracy as some of the other regression and classification approaches</li>
+<li> If continuous features are used the tree may become quite large and hence less interpretable</li>
+<li> Decision trees are prone to overfit the training data and hence do not well generalize the data if no stopping criteria or improvements like pruning, boosting or bagging are implemented</li>
+<li> Small changes in the data may lead to a completely different tree. This issue can be addressed by using ensemble methods like bagging, boosting or random forests</li>
+<li> Unbalanced datasets where some target feature values occur much more frequently than others may lead to biased trees since the frequently occurring feature values are preferred over the less frequently occurring ones.</li> 
+<li> If the number of features is relatively large (high dimensional) and the number of instances is relatively low, the tree might overfit the data</li>
+<li> Features with many levels may be preferred over features with less levels since for them it is <em>more easy</em> to split the dataset such that the sub datasets only contain pure target feature values. This issue can be addressed by preferring for instance the information gain ratio as splitting criteria over information gain</li>
+</ul>
+<p>However, by aggregating many decision trees, using methods like
+bagging, random forests, and boosting, the predictive performance of
+trees can be substantially improved.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="ensemble-methods-from-a-single-tree-to-many-trees-and-extreme-boosting-meet-the-jungle-of-methods">Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods </h2>
+
+<p>As stated above and seen in many of the examples discussed here about
+a single decision tree, we often end up overfitting our training
+data. This normally means that we have a high variance. Can we reduce
+the variance of a statistical learning method?
+</p>
+
+<p>This leads us to a set of different methods that can combine different
+machine learning algorithms or just use one of them to construct
+forests and jungles of trees, homogeneous ones or heterogenous
+ones. These methods are recognized by different names which we will
+try to explain here. These are
+</p>
+
+<ol>
+<li> Voting classifiers</li>
+<li> Bagging and Pasting</li>
+<li> Random forests</li>
+<li> Boosting methods, from adaptive to Extreme Gradient Boosting (XGBoost)</li>
+</ol>
+<p>We discuss these methods here.</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="an-overview-of-ensemble-methods">An Overview of Ensemble Methods  </h2>
+
+<br/><br/>
+<center>
+<p><img src="DataFiles/ensembleoverview.png" width="600" align="bottom"></p>
+</center>
+<br/><br/>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="why-voting">Why Voting? </h2>
+
+<p>The idea behind boosting, and voting as well can be phrased as follows:
+<b>Can a group of people somehow arrive at highly
+reasoned decisions, despite the weak judgement of the individual
+members?</b>
+</p>
+
+<p>The aim is to create a good classifier by combining several weak classifiers.
+<b>A weak classifier is a classifier which is able to produce results that are only slightly better than guessing at random.</b>
+</p>
+
+<p>The basic approach is to apply repeatedly (in boosting this is done in an iterative way) a weak classifier to modifications of the data.
+In voting we simply apply the law of large numbers while in boosting we give more weight to misclassified data in
+each iteration.  
+</p>
+
+<p>Decision trees play an important role as our weak classifier. They serve as the basic method. </p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="tossing-coins">Tossing coins </h2>
+
+<p>The simplest case is a so-called voting ensemble. To illustrate this,
+think of yourself tossing coins with a biased outcome of 51 per cent
+for heads and 49% for tails. With only few tosses,
+you may not clearly see this distribution for heads and tails. However, after some
+thousands of tosses, there will be a clear majority of heads.  With 2000 tosses
+you should see approximately 1020 heads and 980 tails.
+</p>
+
+<p>We can then state that the outcome is a clear majority of heads. If
+you do this ten thousand times, it is easy to see that there is a 97%
+likelihood of a majority of heads.
+</p>
+
+<p>Another example would be to collect all polls before an
+election. Different polls may show different likelihoods for a
+candidate winning with say a majority  of the popular vote. The majority vote
+would then consist in many polls indicating that this candidate will
+actually win.
+</p>
+
+<p>The example here shows how we can implement the coin tossing case,
+clealry demostrating that after some tosses we see the <a href="https://en.wikipedia.org/wiki/Law_of_large_numbers" target="_blank">law of large</a>
+numbers kicking in.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="standard-imports-first">Standard imports first </h2>
+
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #408080; font-style: italic"># Common imports</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">IPython.display</span> <span style="color: #008000; font-weight: bold">import</span> Image 
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">pydot</span> <span style="color: #008000; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">pandas</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">pd</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span> train_test_split
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> export_graphviz
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> StandardScaler, OneHotEncoder
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.compose</span> <span style="color: #008000; font-weight: bold">import</span> ColumnTransformer
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">IPython.display</span> <span style="color: #008000; font-weight: bold">import</span> Image 
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">pydot</span> <span style="color: #008000; font-weight: bold">import</span> graph_from_dot_data
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">os</span>
+
+<span style="color: #408080; font-style: italic"># Where to save the figures and data files</span>
+PROJECT_ROOT_DIR <span style="color: #666666">=</span> <span style="color: #BA2121">&quot;Results&quot;</span>
+FIGURE_ID <span style="color: #666666">=</span> <span style="color: #BA2121">&quot;Results/FigureFiles&quot;</span>
+DATA_ID <span style="color: #666666">=</span> <span style="color: #BA2121">&quot;DataFiles/&quot;</span>
+
+<span style="color: #008000; font-weight: bold">if</span> <span style="color: #AA22FF; font-weight: bold">not</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>exists(PROJECT_ROOT_DIR):
+    os<span style="color: #666666">.</span>mkdir(PROJECT_ROOT_DIR)
+
+<span style="color: #008000; font-weight: bold">if</span> <span style="color: #AA22FF; font-weight: bold">not</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>exists(FIGURE_ID):
+    os<span style="color: #666666">.</span>makedirs(FIGURE_ID)
+
+<span style="color: #008000; font-weight: bold">if</span> <span style="color: #AA22FF; font-weight: bold">not</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>exists(DATA_ID):
+    os<span style="color: #666666">.</span>makedirs(DATA_ID)
+
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">image_path</span>(fig_id):
+    <span style="color: #008000; font-weight: bold">return</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>join(FIGURE_ID, fig_id)
+
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">data_path</span>(dat_id):
+    <span style="color: #008000; font-weight: bold">return</span> os<span style="color: #666666">.</span>path<span style="color: #666666">.</span>join(DATA_ID, dat_id)
+
+<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">save_fig</span>(fig_id):
+    plt<span style="color: #666666">.</span>savefig(image_path(fig_id) <span style="color: #666666">+</span> <span style="color: #BA2121">&quot;.png&quot;</span>, <span style="color: #008000">format</span><span style="color: #666666">=</span><span style="color: #BA2121">&#39;png&#39;</span>)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="simple-voting-example-head-or-tail">Simple Voting Example, head or tail </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #408080; font-style: italic"># Common imports</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">matplotlib.colors</span> <span style="color: #008000; font-weight: bold">import</span> ListedColormap
 plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;axes.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">14</span>
 plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;xtick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
 plt<span style="color: #666666">.</span>rcParams[<span style="color: #BA2121">&#39;ytick.labelsize&#39;</span>] <span style="color: #666666">=</span> <span style="color: #666666">12</span>
 
+heads_proba <span style="color: #666666">=</span> <span style="color: #666666">0.51</span>
+coin_tosses <span style="color: #666666">=</span> (np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>rand(<span style="color: #666666">10000</span>, <span style="color: #666666">10</span>) <span style="color: #666666">&lt;</span> heads_proba)<span style="color: #666666">.</span>astype(np<span style="color: #666666">.</span>int32)
+cumulative_heads_ratio <span style="color: #666666">=</span> np<span style="color: #666666">.</span>cumsum(coin_tosses, axis<span style="color: #666666">=0</span>) <span style="color: #666666">/</span> np<span style="color: #666666">.</span>arange(<span style="color: #666666">1</span>, <span style="color: #666666">10001</span>)<span style="color: #666666">.</span>reshape(<span style="color: #666666">-1</span>, <span style="color: #666666">1</span>)
+plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">8</span>,<span style="color: #666666">3.5</span>))
+plt<span style="color: #666666">.</span>plot(cumulative_heads_ratio)
+plt<span style="color: #666666">.</span>plot([<span style="color: #666666">0</span>, <span style="color: #666666">10000</span>], [<span style="color: #666666">0.51</span>, <span style="color: #666666">0.51</span>], <span style="color: #BA2121">&quot;k--&quot;</span>, linewidth<span style="color: #666666">=2</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;51%&quot;</span>)
+plt<span style="color: #666666">.</span>plot([<span style="color: #666666">0</span>, <span style="color: #666666">10000</span>], [<span style="color: #666666">0.5</span>, <span style="color: #666666">0.5</span>], <span style="color: #BA2121">&quot;k-&quot;</span>, label<span style="color: #666666">=</span><span style="color: #BA2121">&quot;50%&quot;</span>)
+plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">&quot;Number of coin tosses&quot;</span>)
+plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">&quot;Heads ratio&quot;</span>)
+plt<span style="color: #666666">.</span>legend(loc<span style="color: #666666">=</span><span style="color: #BA2121">&quot;lower right&quot;</span>)
+plt<span style="color: #666666">.</span>axis([<span style="color: #666666">0</span>, <span style="color: #666666">10000</span>, <span style="color: #666666">0.42</span>, <span style="color: #666666">0.58</span>])
+save_fig(<span style="color: #BA2121">&quot;votingsimple&quot;</span>)
+plt<span style="color: #666666">.</span>show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
 
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="using-the-voting-classifier">Using the Voting Classifier </h2>
+
+<p>We can use the voting classifier on other data sets, here the exciting binary case of two distinct objects using the make moons functionality of <b>Scikit-Learn</b>.</p>
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span> train_test_split
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> make_moons
+
+X, y <span style="color: #666666">=</span> make_moons(n_samples<span style="color: #666666">=500</span>, noise<span style="color: #666666">=0.30</span>, random_state<span style="color: #666666">=42</span>)
+X_train, X_test, y_train, y_test <span style="color: #666666">=</span> train_test_split(X, y, random_state<span style="color: #666666">=42</span>)
+
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.ensemble</span> <span style="color: #008000; font-weight: bold">import</span> RandomForestClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.ensemble</span> <span style="color: #008000; font-weight: bold">import</span> VotingClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.linear_model</span> <span style="color: #008000; font-weight: bold">import</span> LogisticRegression
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn</span> <span style="color: #008000; font-weight: bold">import</span> datasets
+
+log_clf <span style="color: #666666">=</span> LogisticRegression(solver<span style="color: #666666">=</span><span style="color: #BA2121">&quot;liblinear&quot;</span>, random_state<span style="color: #666666">=42</span>)
+rnd_clf <span style="color: #666666">=</span> RandomForestClassifier(n_estimators<span style="color: #666666">=10</span>, random_state<span style="color: #666666">=42</span>)
+svm_clf <span style="color: #666666">=</span> SVC(gamma<span style="color: #666666">=</span><span style="color: #BA2121">&quot;auto&quot;</span>, random_state<span style="color: #666666">=42</span>)
+
+voting_clf <span style="color: #666666">=</span> VotingClassifier(
+    estimators<span style="color: #666666">=</span>[(<span style="color: #BA2121">&#39;lr&#39;</span>, log_clf), (<span style="color: #BA2121">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #BA2121">&#39;svc&#39;</span>, svm_clf)],
+    voting<span style="color: #666666">=</span><span style="color: #BA2121">&#39;hard&#39;</span>)
+
+voting_clf<span style="color: #666666">.</span>fit(X_train, y_train)
+
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.metrics</span> <span style="color: #008000; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #008000; font-weight: bold">for</span> clf <span style="color: #AA22FF; font-weight: bold">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf<span style="color: #666666">.</span>fit(X_train, y_train)
+    y_pred <span style="color: #666666">=</span> clf<span style="color: #666666">.</span>predict(X_test)
+    <span style="color: #008000">print</span>(clf<span style="color: #666666">.</span><span style="color: #19177C">__class__</span><span style="color: #666666">.</span><span style="color: #19177C">__name__</span>, accuracy_score(y_test, y_pred))
+
+log_clf <span style="color: #666666">=</span> LogisticRegression(solver<span style="color: #666666">=</span><span style="color: #BA2121">&quot;liblinear&quot;</span>, random_state<span style="color: #666666">=42</span>)
+rnd_clf <span style="color: #666666">=</span> RandomForestClassifier(n_estimators<span style="color: #666666">=10</span>, random_state<span style="color: #666666">=42</span>)
+svm_clf <span style="color: #666666">=</span> SVC(gamma<span style="color: #666666">=</span><span style="color: #BA2121">&quot;auto&quot;</span>, probability<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>, random_state<span style="color: #666666">=42</span>)
+voting_clf <span style="color: #666666">=</span> VotingClassifier(
+    estimators<span style="color: #666666">=</span>[(<span style="color: #BA2121">&#39;lr&#39;</span>, log_clf), (<span style="color: #BA2121">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #BA2121">&#39;svc&#39;</span>, svm_clf)],
+    voting<span style="color: #666666">=</span><span style="color: #BA2121">&#39;soft&#39;</span>)
+voting_clf<span style="color: #666666">.</span>fit(X_train, y_train)
+
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.metrics</span> <span style="color: #008000; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #008000; font-weight: bold">for</span> clf <span style="color: #AA22FF; font-weight: bold">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf<span style="color: #666666">.</span>fit(X_train, y_train)
+    y_pred <span style="color: #666666">=</span> clf<span style="color: #666666">.</span>predict(X_test)
+    <span style="color: #008000">print</span>(clf<span style="color: #666666">.</span><span style="color: #19177C">__class__</span><span style="color: #666666">.</span><span style="color: #19177C">__name__</span>, accuracy_score(y_test, y_pred))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
 
 
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="voting-and-bagging">Voting and Bagging </h2>
+
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span> train_test_split
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> make_moons
+
+X, y <span style="color: #666666">=</span> make_moons(n_samples<span style="color: #666666">=500</span>, noise<span style="color: #666666">=0.30</span>, random_state<span style="color: #666666">=42</span>)
+X_train, X_test, y_train, y_test <span style="color: #666666">=</span> train_test_split(X, y, random_state<span style="color: #666666">=42</span>)
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.ensemble</span> <span style="color: #008000; font-weight: bold">import</span> RandomForestClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.ensemble</span> <span style="color: #008000; font-weight: bold">import</span> VotingClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.linear_model</span> <span style="color: #008000; font-weight: bold">import</span> LogisticRegression
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC
+
+log_clf <span style="color: #666666">=</span> LogisticRegression(random_state<span style="color: #666666">=42</span>)
+rnd_clf <span style="color: #666666">=</span> RandomForestClassifier(random_state<span style="color: #666666">=42</span>)
+svm_clf <span style="color: #666666">=</span> SVC(random_state<span style="color: #666666">=42</span>)
+
+voting_clf <span style="color: #666666">=</span> VotingClassifier(
+    estimators<span style="color: #666666">=</span>[(<span style="color: #BA2121">&#39;lr&#39;</span>, log_clf), (<span style="color: #BA2121">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #BA2121">&#39;svc&#39;</span>, svm_clf)],
+    voting<span style="color: #666666">=</span><span style="color: #BA2121">&#39;hard&#39;</span>)
+voting_clf<span style="color: #666666">.</span>fit(X_train, y_train)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.metrics</span> <span style="color: #008000; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #008000; font-weight: bold">for</span> clf <span style="color: #AA22FF; font-weight: bold">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf<span style="color: #666666">.</span>fit(X_train, y_train)
+    y_pred <span style="color: #666666">=</span> clf<span style="color: #666666">.</span>predict(X_test)
+    <span style="color: #008000">print</span>(clf<span style="color: #666666">.</span><span style="color: #19177C">__class__</span><span style="color: #666666">.</span><span style="color: #19177C">__name__</span>, accuracy_score(y_test, y_pred))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;">log_clf <span style="color: #666666">=</span> LogisticRegression(random_state<span style="color: #666666">=42</span>)
+rnd_clf <span style="color: #666666">=</span> RandomForestClassifier(random_state<span style="color: #666666">=42</span>)
+svm_clf <span style="color: #666666">=</span> SVC(probability<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>, random_state<span style="color: #666666">=42</span>)
+
+voting_clf <span style="color: #666666">=</span> VotingClassifier(
+    estimators<span style="color: #666666">=</span>[(<span style="color: #BA2121">&#39;lr&#39;</span>, log_clf), (<span style="color: #BA2121">&#39;rf&#39;</span>, rnd_clf), (<span style="color: #BA2121">&#39;svc&#39;</span>, svm_clf)],
+    voting<span style="color: #666666">=</span><span style="color: #BA2121">&#39;soft&#39;</span>)
+voting_clf<span style="color: #666666">.</span>fit(X_train, y_train)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.metrics</span> <span style="color: #008000; font-weight: bold">import</span> accuracy_score
+
+<span style="color: #008000; font-weight: bold">for</span> clf <span style="color: #AA22FF; font-weight: bold">in</span> (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf<span style="color: #666666">.</span>fit(X_train, y_train)
+    y_pred <span style="color: #666666">=</span> clf<span style="color: #666666">.</span>predict(X_test)
+    <span style="color: #008000">print</span>(clf<span style="color: #666666">.</span><span style="color: #19177C">__class__</span><span style="color: #666666">.</span><span style="color: #19177C">__name__</span>, accuracy_score(y_test, y_pred))
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="bagging">Bagging </h2>
+
+<p>The <b>plain</b> decision trees suffer from high
+variance. This means that if we split the training data into two parts
+at random, and fit a decision tree to both halves, the results that we
+get could be quite different. In contrast, a procedure with low
+variance will yield similar results if applied repeatedly to distinct
+data sets; linear regression tends to have low variance, if the ratio
+of \( n \) to \( p \) is moderately large. 
+</p>
+
+<p><b>Bootstrap aggregation</b>, or just <b>bagging</b>, is a
+general-purpose procedure for reducing the variance of a statistical
+learning method. 
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="more-bagging">More bagging </h2>
+
+<p>Bagging typically results in improved accuracy
+over prediction using a single tree. Unfortunately, however, it can be
+difficult to interpret the resulting model. Recall that one of the
+advantages of decision trees is the attractive and easily interpreted
+diagram that results.
+</p>
+
+<p>However, when we bag a large number of trees, it is no longer
+possible to represent the resulting statistical learning procedure
+using a single tree, and it is no longer clear which variables are
+most important to the procedure. Thus, bagging improves prediction
+accuracy at the expense of interpretability.  Although the collection
+of bagged trees is much more difficult to interpret than a single
+tree, one can obtain an overall summary of the importance of each
+predictor using the MSE (for bagging regression trees) or the Gini
+index (for bagging classification trees). In the case of bagging
+regression trees, we can record the total amount that the MSE is
+decreased due to splits over a given predictor, averaged over all \( B \) possible
+trees. A large value indicates an important predictor. Similarly, in
+the context of bagging classification trees, we can add up the total
+amount that the Gini index  is decreased by splits over a given
+predictor, averaged over all \( B \) trees.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="making-your-own-bootstrap-changing-the-level-of-the-decision-tree">Making your own Bootstrap: Changing the Level of the Decision Tree </h2>
+
+<p>Let us bring up our good old boostrap example from the linear regression lectures. We change the linerar regression algorithm with
+a decision tree wth different depths and perform a bootstrap aggregate (in this case we perform as many bootstraps as data points \( n \)).
+</p>
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span> train_test_split
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.pipeline</span> <span style="color: #008000; font-weight: bold">import</span> make_pipeline
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.utils</span> <span style="color: #008000; font-weight: bold">import</span> resample
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeRegressor
+
+n <span style="color: #666666">=</span> <span style="color: #666666">100</span>
+n_boostraps <span style="color: #666666">=</span> <span style="color: #666666">100</span>
+maxdepth <span style="color: #666666">=</span> <span style="color: #666666">8</span>
+
+<span style="color: #408080; font-style: italic"># Make data set.</span>
+x <span style="color: #666666">=</span> np<span style="color: #666666">.</span>linspace(<span style="color: #666666">-3</span>, <span style="color: #666666">3</span>, n)<span style="color: #666666">.</span>reshape(<span style="color: #666666">-1</span>, <span style="color: #666666">1</span>)
+y <span style="color: #666666">=</span> np<span style="color: #666666">.</span>exp(<span style="color: #666666">-</span>x<span style="color: #666666">**2</span>) <span style="color: #666666">+</span> <span style="color: #666666">1.5</span> <span style="color: #666666">*</span> np<span style="color: #666666">.</span>exp(<span style="color: #666666">-</span>(x<span style="color: #666666">-2</span>)<span style="color: #666666">**2</span>)<span style="color: #666666">+</span> np<span style="color: #666666">.</span>random<span style="color: #666666">.</span>normal(<span style="color: #666666">0</span>, <span style="color: #666666">0.1</span>, x<span style="color: #666666">.</span>shape)
+error <span style="color: #666666">=</span> np<span style="color: #666666">.</span>zeros(maxdepth)
+bias <span style="color: #666666">=</span> np<span style="color: #666666">.</span>zeros(maxdepth)
+variance <span style="color: #666666">=</span> np<span style="color: #666666">.</span>zeros(maxdepth)
+polydegree <span style="color: #666666">=</span> np<span style="color: #666666">.</span>zeros(maxdepth)
+X_train, X_test, y_train, y_test <span style="color: #666666">=</span> train_test_split(x, y, test_size<span style="color: #666666">=0.2</span>)
 
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.pipeline</span> <span style="color: #008000; font-weight: bold">import</span> Pipeline
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> StandardScaler
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> LinearSVC
+scaler <span style="color: #666666">=</span> StandardScaler()
+scaler<span style="color: #666666">.</span>fit(X_train)
+X_train_scaled <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>transform(X_train)
+X_test_scaled <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>transform(X_test)
 
+<span style="color: #408080; font-style: italic"># we produce a simple tree first as benchmark</span>
+simpletree <span style="color: #666666">=</span> DecisionTreeRegressor(max_depth<span style="color: #666666">=3</span>) 
+simpletree<span style="color: #666666">.</span>fit(X_train_scaled, y_train)
+simpleprediction <span style="color: #666666">=</span> simpletree<span style="color: #666666">.</span>predict(X_test_scaled)
+<span style="color: #008000; font-weight: bold">for</span> degree <span style="color: #AA22FF; font-weight: bold">in</span> <span style="color: #008000">range</span>(<span style="color: #666666">1</span>,maxdepth):
+    model <span style="color: #666666">=</span> DecisionTreeRegressor(max_depth<span style="color: #666666">=</span>degree) 
+    y_pred <span style="color: #666666">=</span> np<span style="color: #666666">.</span>empty((y_test<span style="color: #666666">.</span>shape[<span style="color: #666666">0</span>], n_boostraps))
+    <span style="color: #008000; font-weight: bold">for</span> i <span style="color: #AA22FF; font-weight: bold">in</span> <span style="color: #008000">range</span>(n_boostraps):
+        x_, y_ <span style="color: #666666">=</span> resample(X_train_scaled, y_train)
+        model<span style="color: #666666">.</span>fit(x_, y_)
+        y_pred[:, i] <span style="color: #666666">=</span> model<span style="color: #666666">.</span>predict(X_test_scaled)<span style="color: #408080; font-style: italic">#.ravel()</span>
 
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> make_moons
-X, y <span style="color: #666666">=</span> make_moons(n_samples<span style="color: #666666">=100</span>, noise<span style="color: #666666">=0.15</span>, random_state<span style="color: #666666">=42</span>)
-
-<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">plot_dataset</span>(X, y, axes):
-    plt<span style="color: #666666">.</span>plot(X[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==0</span>], X[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==0</span>], <span style="color: #BA2121">&quot;bs&quot;</span>)
-    plt<span style="color: #666666">.</span>plot(X[:, <span style="color: #666666">0</span>][y<span style="color: #666666">==1</span>], X[:, <span style="color: #666666">1</span>][y<span style="color: #666666">==1</span>], <span style="color: #BA2121">&quot;g^&quot;</span>)
-    plt<span style="color: #666666">.</span>axis(axes)
-    plt<span style="color: #666666">.</span>grid(<span style="color: #008000; font-weight: bold">True</span>, which<span style="color: #666666">=</span><span style="color: #BA2121">&#39;both&#39;</span>)
-    plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">r&quot;$x_1$&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-    plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">r&quot;$x_2$&quot;</span>, fontsize<span style="color: #666666">=20</span>, rotation<span style="color: #666666">=0</span>)
-
-plot_dataset(X, y, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plt<span style="color: #666666">.</span>show()
-
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> make_moons
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.pipeline</span> <span style="color: #008000; font-weight: bold">import</span> Pipeline
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> PolynomialFeatures
-
-polynomial_svm_clf <span style="color: #666666">=</span> Pipeline([
-        (<span style="color: #BA2121">&quot;poly_features&quot;</span>, PolynomialFeatures(degree<span style="color: #666666">=3</span>)),
-        (<span style="color: #BA2121">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #BA2121">&quot;svm_clf&quot;</span>, LinearSVC(C<span style="color: #666666">=10</span>, loss<span style="color: #666666">=</span><span style="color: #BA2121">&quot;hinge&quot;</span>, random_state<span style="color: #666666">=42</span>))
-    ])
-
-polynomial_svm_clf<span style="color: #666666">.</span>fit(X, y)
-
-<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">plot_predictions</span>(clf, axes):
-    x0s <span style="color: #666666">=</span> np<span style="color: #666666">.</span>linspace(axes[<span style="color: #666666">0</span>], axes[<span style="color: #666666">1</span>], <span style="color: #666666">100</span>)
-    x1s <span style="color: #666666">=</span> np<span style="color: #666666">.</span>linspace(axes[<span style="color: #666666">2</span>], axes[<span style="color: #666666">3</span>], <span style="color: #666666">100</span>)
-    x0, x1 <span style="color: #666666">=</span> np<span style="color: #666666">.</span>meshgrid(x0s, x1s)
-    X <span style="color: #666666">=</span> np<span style="color: #666666">.</span>c_[x0<span style="color: #666666">.</span>ravel(), x1<span style="color: #666666">.</span>ravel()]
-    y_pred <span style="color: #666666">=</span> clf<span style="color: #666666">.</span>predict(X)<span style="color: #666666">.</span>reshape(x0<span style="color: #666666">.</span>shape)
-    y_decision <span style="color: #666666">=</span> clf<span style="color: #666666">.</span>decision_function(X)<span style="color: #666666">.</span>reshape(x0<span style="color: #666666">.</span>shape)
-    plt<span style="color: #666666">.</span>contourf(x0, x1, y_pred, cmap<span style="color: #666666">=</span>plt<span style="color: #666666">.</span>cm<span style="color: #666666">.</span>brg, alpha<span style="color: #666666">=0.2</span>)
-    plt<span style="color: #666666">.</span>contourf(x0, x1, y_decision, cmap<span style="color: #666666">=</span>plt<span style="color: #666666">.</span>cm<span style="color: #666666">.</span>brg, alpha<span style="color: #666666">=0.1</span>)
-
-plot_predictions(polynomial_svm_clf, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plot_dataset(X, y, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-
-plt<span style="color: #666666">.</span>show()
-
-
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC
-
-poly_kernel_svm_clf <span style="color: #666666">=</span> Pipeline([
-        (<span style="color: #BA2121">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #BA2121">&quot;svm_clf&quot;</span>, SVC(kernel<span style="color: #666666">=</span><span style="color: #BA2121">&quot;poly&quot;</span>, degree<span style="color: #666666">=3</span>, coef0<span style="color: #666666">=1</span>, C<span style="color: #666666">=5</span>))
-    ])
-poly_kernel_svm_clf<span style="color: #666666">.</span>fit(X, y)
-
-poly100_kernel_svm_clf <span style="color: #666666">=</span> Pipeline([
-        (<span style="color: #BA2121">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #BA2121">&quot;svm_clf&quot;</span>, SVC(kernel<span style="color: #666666">=</span><span style="color: #BA2121">&quot;poly&quot;</span>, degree<span style="color: #666666">=10</span>, coef0<span style="color: #666666">=100</span>, C<span style="color: #666666">=5</span>))
-    ])
-poly100_kernel_svm_clf<span style="color: #666666">.</span>fit(X, y)
-
-plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">4</span>))
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">121</span>)
-plot_predictions(poly_kernel_svm_clf, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plot_dataset(X, y, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">r&quot;$d=3, r=1, C=5$&quot;</span>, fontsize<span style="color: #666666">=18</span>)
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">122</span>)
-plot_predictions(poly100_kernel_svm_clf, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plot_dataset(X, y, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">r&quot;$d=10, r=100, C=5$&quot;</span>, fontsize<span style="color: #666666">=18</span>)
-
-plt<span style="color: #666666">.</span>show()
-
-<span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">gaussian_rbf</span>(x, landmark, gamma):
-    <span style="color: #008000; font-weight: bold">return</span> np<span style="color: #666666">.</span>exp(<span style="color: #666666">-</span>gamma <span style="color: #666666">*</span> np<span style="color: #666666">.</span>linalg<span style="color: #666666">.</span>norm(x <span style="color: #666666">-</span> landmark, axis<span style="color: #666666">=1</span>)<span style="color: #666666">**2</span>)
-
-gamma <span style="color: #666666">=</span> <span style="color: #666666">0.3</span>
-
-x1s <span style="color: #666666">=</span> np<span style="color: #666666">.</span>linspace(<span style="color: #666666">-4.5</span>, <span style="color: #666666">4.5</span>, <span style="color: #666666">200</span>)<span style="color: #666666">.</span>reshape(<span style="color: #666666">-1</span>, <span style="color: #666666">1</span>)
-x2s <span style="color: #666666">=</span> gaussian_rbf(x1s, <span style="color: #666666">-2</span>, gamma)
-x3s <span style="color: #666666">=</span> gaussian_rbf(x1s, <span style="color: #666666">1</span>, gamma)
-
-XK <span style="color: #666666">=</span> np<span style="color: #666666">.</span>c_[gaussian_rbf(X1D, <span style="color: #666666">-2</span>, gamma), gaussian_rbf(X1D, <span style="color: #666666">1</span>, gamma)]
-yk <span style="color: #666666">=</span> np<span style="color: #666666">.</span>array([<span style="color: #666666">0</span>, <span style="color: #666666">0</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">1</span>, <span style="color: #666666">0</span>, <span style="color: #666666">0</span>])
-
-plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">4</span>))
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">121</span>)
-plt<span style="color: #666666">.</span>grid(<span style="color: #008000; font-weight: bold">True</span>, which<span style="color: #666666">=</span><span style="color: #BA2121">&#39;both&#39;</span>)
-plt<span style="color: #666666">.</span>axhline(y<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>scatter(x<span style="color: #666666">=</span>[<span style="color: #666666">-2</span>, <span style="color: #666666">1</span>], y<span style="color: #666666">=</span>[<span style="color: #666666">0</span>, <span style="color: #666666">0</span>], s<span style="color: #666666">=150</span>, alpha<span style="color: #666666">=0.5</span>, c<span style="color: #666666">=</span><span style="color: #BA2121">&quot;red&quot;</span>)
-plt<span style="color: #666666">.</span>plot(X1D[:, <span style="color: #666666">0</span>][yk<span style="color: #666666">==0</span>], np<span style="color: #666666">.</span>zeros(<span style="color: #666666">4</span>), <span style="color: #BA2121">&quot;bs&quot;</span>)
-plt<span style="color: #666666">.</span>plot(X1D[:, <span style="color: #666666">0</span>][yk<span style="color: #666666">==1</span>], np<span style="color: #666666">.</span>zeros(<span style="color: #666666">5</span>), <span style="color: #BA2121">&quot;g^&quot;</span>)
-plt<span style="color: #666666">.</span>plot(x1s, x2s, <span style="color: #BA2121">&quot;g--&quot;</span>)
-plt<span style="color: #666666">.</span>plot(x1s, x3s, <span style="color: #BA2121">&quot;b:&quot;</span>)
-plt<span style="color: #666666">.</span>gca()<span style="color: #666666">.</span>get_yaxis()<span style="color: #666666">.</span>set_ticks([<span style="color: #666666">0</span>, <span style="color: #666666">0.25</span>, <span style="color: #666666">0.5</span>, <span style="color: #666666">0.75</span>, <span style="color: #666666">1</span>])
-plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">r&quot;$x_1$&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">r&quot;Similarity&quot;</span>, fontsize<span style="color: #666666">=14</span>)
-plt<span style="color: #666666">.</span>annotate(<span style="color: #BA2121">r&#39;$\mathbf</span><span style="color: #BB6688; font-weight: bold">{x}</span><span style="color: #BA2121">$&#39;</span>,
-             xy<span style="color: #666666">=</span>(X1D[<span style="color: #666666">3</span>, <span style="color: #666666">0</span>], <span style="color: #666666">0</span>),
-             xytext<span style="color: #666666">=</span>(<span style="color: #666666">-0.5</span>, <span style="color: #666666">0.20</span>),
-             ha<span style="color: #666666">=</span><span style="color: #BA2121">&quot;center&quot;</span>,
-             arrowprops<span style="color: #666666">=</span><span style="color: #008000">dict</span>(facecolor<span style="color: #666666">=</span><span style="color: #BA2121">&#39;black&#39;</span>, shrink<span style="color: #666666">=0.1</span>),
-             fontsize<span style="color: #666666">=18</span>,
-            )
-plt<span style="color: #666666">.</span>text(<span style="color: #666666">-2</span>, <span style="color: #666666">0.9</span>, <span style="color: #BA2121">&quot;$x_2$&quot;</span>, ha<span style="color: #666666">=</span><span style="color: #BA2121">&quot;center&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>text(<span style="color: #666666">1</span>, <span style="color: #666666">0.9</span>, <span style="color: #BA2121">&quot;$x_3$&quot;</span>, ha<span style="color: #666666">=</span><span style="color: #BA2121">&quot;center&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>axis([<span style="color: #666666">-4.5</span>, <span style="color: #666666">4.5</span>, <span style="color: #666666">-0.1</span>, <span style="color: #666666">1.1</span>])
-
-plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">122</span>)
-plt<span style="color: #666666">.</span>grid(<span style="color: #008000; font-weight: bold">True</span>, which<span style="color: #666666">=</span><span style="color: #BA2121">&#39;both&#39;</span>)
-plt<span style="color: #666666">.</span>axhline(y<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>axvline(x<span style="color: #666666">=0</span>, color<span style="color: #666666">=</span><span style="color: #BA2121">&#39;k&#39;</span>)
-plt<span style="color: #666666">.</span>plot(XK[:, <span style="color: #666666">0</span>][yk<span style="color: #666666">==0</span>], XK[:, <span style="color: #666666">1</span>][yk<span style="color: #666666">==0</span>], <span style="color: #BA2121">&quot;bs&quot;</span>)
-plt<span style="color: #666666">.</span>plot(XK[:, <span style="color: #666666">0</span>][yk<span style="color: #666666">==1</span>], XK[:, <span style="color: #666666">1</span>][yk<span style="color: #666666">==1</span>], <span style="color: #BA2121">&quot;g^&quot;</span>)
-plt<span style="color: #666666">.</span>xlabel(<span style="color: #BA2121">r&quot;$x_2$&quot;</span>, fontsize<span style="color: #666666">=20</span>)
-plt<span style="color: #666666">.</span>ylabel(<span style="color: #BA2121">r&quot;$x_3$  &quot;</span>, fontsize<span style="color: #666666">=20</span>, rotation<span style="color: #666666">=0</span>)
-plt<span style="color: #666666">.</span>annotate(<span style="color: #BA2121">r&#39;$\phi\left(\mathbf</span><span style="color: #BB6688; font-weight: bold">{x}</span><span style="color: #BA2121">\right)$&#39;</span>,
-             xy<span style="color: #666666">=</span>(XK[<span style="color: #666666">3</span>, <span style="color: #666666">0</span>], XK[<span style="color: #666666">3</span>, <span style="color: #666666">1</span>]),
-             xytext<span style="color: #666666">=</span>(<span style="color: #666666">0.65</span>, <span style="color: #666666">0.50</span>),
-             ha<span style="color: #666666">=</span><span style="color: #BA2121">&quot;center&quot;</span>,
-             arrowprops<span style="color: #666666">=</span><span style="color: #008000">dict</span>(facecolor<span style="color: #666666">=</span><span style="color: #BA2121">&#39;black&#39;</span>, shrink<span style="color: #666666">=0.1</span>),
-             fontsize<span style="color: #666666">=18</span>,
-            )
-plt<span style="color: #666666">.</span>plot([<span style="color: #666666">-0.1</span>, <span style="color: #666666">1.1</span>], [<span style="color: #666666">0.57</span>, <span style="color: #666666">-0.1</span>], <span style="color: #BA2121">&quot;r--&quot;</span>, linewidth<span style="color: #666666">=3</span>)
-plt<span style="color: #666666">.</span>axis([<span style="color: #666666">-0.1</span>, <span style="color: #666666">1.1</span>, <span style="color: #666666">-0.1</span>, <span style="color: #666666">1.1</span>])
-    
-plt<span style="color: #666666">.</span>subplots_adjust(right<span style="color: #666666">=1</span>)
-
-plt<span style="color: #666666">.</span>show()
-
-
-x1_example <span style="color: #666666">=</span> X1D[<span style="color: #666666">3</span>, <span style="color: #666666">0</span>]
-<span style="color: #008000; font-weight: bold">for</span> landmark <span style="color: #AA22FF; font-weight: bold">in</span> (<span style="color: #666666">-2</span>, <span style="color: #666666">1</span>):
-    k <span style="color: #666666">=</span> gaussian_rbf(np<span style="color: #666666">.</span>array([[x1_example]]), np<span style="color: #666666">.</span>array([[landmark]]), gamma)
-    <span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;Phi(</span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">, </span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">) = </span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">&quot;</span><span style="color: #666666">.</span>format(x1_example, landmark, k))
-
-rbf_kernel_svm_clf <span style="color: #666666">=</span> Pipeline([
-        (<span style="color: #BA2121">&quot;scaler&quot;</span>, StandardScaler()),
-        (<span style="color: #BA2121">&quot;svm_clf&quot;</span>, SVC(kernel<span style="color: #666666">=</span><span style="color: #BA2121">&quot;rbf&quot;</span>, gamma<span style="color: #666666">=5</span>, C<span style="color: #666666">=0.001</span>))
-    ])
-rbf_kernel_svm_clf<span style="color: #666666">.</span>fit(X, y)
-
-
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC
-
-gamma1, gamma2 <span style="color: #666666">=</span> <span style="color: #666666">0.1</span>, <span style="color: #666666">5</span>
-C1, C2 <span style="color: #666666">=</span> <span style="color: #666666">0.001</span>, <span style="color: #666666">1000</span>
-hyperparams <span style="color: #666666">=</span> (gamma1, C1), (gamma1, C2), (gamma2, C1), (gamma2, C2)
-
-svm_clfs <span style="color: #666666">=</span> []
-<span style="color: #008000; font-weight: bold">for</span> gamma, C <span style="color: #AA22FF; font-weight: bold">in</span> hyperparams:
-    rbf_kernel_svm_clf <span style="color: #666666">=</span> Pipeline([
-            (<span style="color: #BA2121">&quot;scaler&quot;</span>, StandardScaler()),
-            (<span style="color: #BA2121">&quot;svm_clf&quot;</span>, SVC(kernel<span style="color: #666666">=</span><span style="color: #BA2121">&quot;rbf&quot;</span>, gamma<span style="color: #666666">=</span>gamma, C<span style="color: #666666">=</span>C))
-        ])
-    rbf_kernel_svm_clf<span style="color: #666666">.</span>fit(X, y)
-    svm_clfs<span style="color: #666666">.</span>append(rbf_kernel_svm_clf)
-
-plt<span style="color: #666666">.</span>figure(figsize<span style="color: #666666">=</span>(<span style="color: #666666">11</span>, <span style="color: #666666">7</span>))
-
-<span style="color: #008000; font-weight: bold">for</span> i, svm_clf <span style="color: #AA22FF; font-weight: bold">in</span> <span style="color: #008000">enumerate</span>(svm_clfs):
-    plt<span style="color: #666666">.</span>subplot(<span style="color: #666666">221</span> <span style="color: #666666">+</span> i)
-    plot_predictions(svm_clf, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-    plot_dataset(X, y, [<span style="color: #666666">-1.5</span>, <span style="color: #666666">2.5</span>, <span style="color: #666666">-1</span>, <span style="color: #666666">1.5</span>])
-    gamma, C <span style="color: #666666">=</span> hyperparams[i]
-    plt<span style="color: #666666">.</span>title(<span style="color: #BA2121">r&quot;$\gamma = </span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">, C = </span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">$&quot;</span><span style="color: #666666">.</span>format(gamma, C), fontsize<span style="color: #666666">=16</span>)
-
+    polydegree[degree] <span style="color: #666666">=</span> degree
+    error[degree] <span style="color: #666666">=</span> np<span style="color: #666666">.</span>mean( np<span style="color: #666666">.</span>mean((y_test <span style="color: #666666">-</span> y_pred)<span style="color: #666666">**2</span>, axis<span style="color: #666666">=1</span>, keepdims<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>) )
+    bias[degree] <span style="color: #666666">=</span> np<span style="color: #666666">.</span>mean( (y_test <span style="color: #666666">-</span> np<span style="color: #666666">.</span>mean(y_pred, axis<span style="color: #666666">=1</span>, keepdims<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>))<span style="color: #666666">**2</span> )
+    variance[degree] <span style="color: #666666">=</span> np<span style="color: #666666">.</span>mean( np<span style="color: #666666">.</span>var(y_pred, axis<span style="color: #666666">=1</span>, keepdims<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>) )
+    <span style="color: #008000">print</span>(<span style="color: #BA2121">&#39;Polynomial degree:&#39;</span>, degree)
+    <span style="color: #008000">print</span>(<span style="color: #BA2121">&#39;Error:&#39;</span>, error[degree])
+    <span style="color: #008000">print</span>(<span style="color: #BA2121">&#39;Bias^2:&#39;</span>, bias[degree])
+    <span style="color: #008000">print</span>(<span style="color: #BA2121">&#39;Var:&#39;</span>, variance[degree])
+    <span style="color: #008000">print</span>(<span style="color: #BA2121">&#39;</span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121"> &gt;= </span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121"> + </span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121"> = </span><span style="color: #BB6688; font-weight: bold">{}</span><span style="color: #BA2121">&#39;</span><span style="color: #666666">.</span>format(error[degree], bias[degree], variance[degree], bias[degree]<span style="color: #666666">+</span>variance[degree]))
+ 
+mse_simpletree<span style="color: #666666">=</span> np<span style="color: #666666">.</span>mean( np<span style="color: #666666">.</span>mean((y_test <span style="color: #666666">-</span> simpleprediction)<span style="color: #666666">**2</span>))
+<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;Simple tree:&quot;</span>,mse_simpletree)
+plt<span style="color: #666666">.</span>xlim(<span style="color: #666666">1</span>,maxdepth)
+plt<span style="color: #666666">.</span>plot(polydegree, error, label<span style="color: #666666">=</span><span style="color: #BA2121">&#39;MSE&#39;</span>)
+plt<span style="color: #666666">.</span>plot(polydegree, bias, label<span style="color: #666666">=</span><span style="color: #BA2121">&#39;bias&#39;</span>)
+plt<span style="color: #666666">.</span>plot(polydegree, variance, label<span style="color: #666666">=</span><span style="color: #BA2121">&#39;Variance&#39;</span>)
+plt<span style="color: #666666">.</span>legend()
+save_fig(<span style="color: #BA2121">&quot;baggingboot&quot;</span>)
 plt<span style="color: #666666">.</span>show()
 </pre>
 </div>
@@ -1330,40 +2194,494 @@ plt<span style="color: #666666">.</span>show()
 
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="mathematical-optimization-of-convex-functions">Mathematical optimization of convex functions </h2>
+<h2 id="random-forests">Random forests </h2>
 
-<p>A mathematical (quadratic) optimization problem, or just optimization problem, has the form</p>
-$$
-\begin{align*}
-    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\boldsymbol{\lambda}^T\boldsymbol{P}\boldsymbol{\lambda}+\boldsymbol{q}^T\boldsymbol{\lambda},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{\lambda} \preceq \boldsymbol{h} \wedge  \boldsymbol{A}\boldsymbol{\lambda}=f.
-\end{align*}
-$$
-
-<p>subject to some constraints for say a selected set \( i=1,2,\dots, n \).
-In our case we are optimizing with respect to the Lagrangian multipliers \( \lambda_i \), and the
-vector \( \boldsymbol{\lambda}=[\lambda_1, \lambda_2,\dots, \lambda_n] \) is the optimization variable we are dealing with.
+<p>Random forests provide an improvement over bagged trees by way of a
+small tweak that decorrelates the trees. 
 </p>
 
-<p>In our case we are particularly interested in a class of optimization problems called convex optmization problems. 
-In our discussion on gradient descent methods we discussed at length the definition of a convex function. 
+<p>As in bagging, we build a
+number of decision trees on bootstrapped training samples. But when
+building these decision trees, each time a split in a tree is
+considered, a random sample of \( m \) predictors is chosen as split
+candidates from the full set of \( p \) predictors. The split is allowed to
+use only one of those \( m \) predictors. 
 </p>
 
-<p>Convex optimization problems play a central role in applied mathematics and we recommend strongly <a href="http://web.stanford.edu/~boyd/cvxbook/" target="_blank">Boyd and Vandenberghe's text on the topics</a>.</p>
+<p>A fresh sample of \( m \) predictors is
+taken at each split, and typically we choose 
+</p>
+
+$$
+m\approx \sqrt{p}.
+$$
+
+<p>In building a random forest, at
+each split in the tree, the algorithm is not even allowed to consider
+a majority of the available predictors. 
+</p>
+
+<p>The reason for this is rather clever. Suppose that there is one very
+strong predictor in the data set, along with a number of other
+moderately strong predictors. Then in the collection of bagged
+variable importance random forest trees, most or all of the trees will
+use this strong predictor in the top split. Consequently, all of the
+bagged trees will look quite similar to each other. Hence the
+predictions from the bagged trees will be highly correlated.
+Unfortunately, averaging many highly correlated quantities does not
+lead to as large of a reduction in variance as averaging many
+uncorrelated quantities. In particular, this means that bagging will
+not lead to a substantial reduction in variance over a single tree in
+this setting.
+</p>
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="how-do-we-solve-these-problems">How do we solve these problems? </h2>
+<h2 id="random-forest-algorithm">Random Forest Algorithm </h2>
+<p>The algorithm described here can be applied to both classification and regression problems.</p>
 
-<p>If we use Python as programming language and wish to venture beyond
-<b>scikit-learn</b>, <b>tensorflow</b> and similar software which makes our
-lives so much easier, we need to dive into the wonderful world of
-quadratic programming. We can, if we wish, solve the minimization
-problem using say standard gradient methods or conjugate gradient
-methods. However, these methods tend to exhibit a rather slow
-converge. So, welcome to the promised land of quadratic programming.
+<p>We will grow of forest of say \( B \) trees.</p>
+<ol>
+<li> For \( b=1:B \)</li>
+<ul>
+  <li> Draw a bootstrap sample from the training data organized in our \( \boldsymbol{X} \) matrix.</li>
+  <li> We grow then a random forest tree \( T_b \) based on the bootstrapped data by repeating the steps outlined till we reach the maximum node size is reached</li>
+<ol>
+   <li> we select \( m \le p \) variables at random from the \( p \) predictors/features</li>
+   <li> pick the best split point among the \( m \) features using for example the CART algorithm and create a new node</li>
+   <li> split the node into daughter nodes</li>
+</ol>
+</ul>
+<li> Output then the ensemble of trees \( \{T_b\}_1^{B} \) and make predictions for either a regression type of problem or a classification type of problem.</li> 
+</ol>
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="random-forests-compared-with-other-methods-on-the-cancer-data">Random Forests Compared with other Methods on the Cancer Data </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span>  train_test_split 
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.datasets</span> <span style="color: #008000; font-weight: bold">import</span> load_breast_cancer
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.svm</span> <span style="color: #008000; font-weight: bold">import</span> SVC
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.linear_model</span> <span style="color: #008000; font-weight: bold">import</span> LogisticRegression
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.tree</span> <span style="color: #008000; font-weight: bold">import</span> DecisionTreeClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.ensemble</span> <span style="color: #008000; font-weight: bold">import</span> BaggingClassifier
+
+<span style="color: #408080; font-style: italic"># Load the data</span>
+cancer <span style="color: #666666">=</span> load_breast_cancer()
+
+X_train, X_test, y_train, y_test <span style="color: #666666">=</span> train_test_split(cancer<span style="color: #666666">.</span>data,cancer<span style="color: #666666">.</span>target,random_state<span style="color: #666666">=0</span>)
+<span style="color: #008000">print</span>(X_train<span style="color: #666666">.</span>shape)
+<span style="color: #008000">print</span>(X_test<span style="color: #666666">.</span>shape)
+<span style="color: #408080; font-style: italic">#define methods</span>
+<span style="color: #408080; font-style: italic"># Logistic Regression</span>
+logreg <span style="color: #666666">=</span> LogisticRegression(solver<span style="color: #666666">=</span><span style="color: #BA2121">&#39;lbfgs&#39;</span>)
+<span style="color: #408080; font-style: italic"># Support vector machine</span>
+svm <span style="color: #666666">=</span> SVC(gamma<span style="color: #666666">=</span><span style="color: #BA2121">&#39;auto&#39;</span>, C<span style="color: #666666">=100</span>)
+<span style="color: #408080; font-style: italic"># Decision Trees</span>
+deep_tree_clf <span style="color: #666666">=</span> DecisionTreeClassifier(max_depth<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">None</span>)
+<span style="color: #408080; font-style: italic">#Scale the data</span>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> StandardScaler
+scaler <span style="color: #666666">=</span> StandardScaler()
+scaler<span style="color: #666666">.</span>fit(X_train)
+X_train_scaled <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>transform(X_train)
+X_test_scaled <span style="color: #666666">=</span> scaler<span style="color: #666666">.</span>transform(X_test)
+<span style="color: #408080; font-style: italic"># Logistic Regression</span>
+logreg<span style="color: #666666">.</span>fit(X_train_scaled, y_train)
+<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;Test set accuracy Logistic Regression with scaled data: </span><span style="color: #BB6688; font-weight: bold">{:.2f}</span><span style="color: #BA2121">&quot;</span><span style="color: #666666">.</span>format(logreg<span style="color: #666666">.</span>score(X_test_scaled,y_test)))
+<span style="color: #408080; font-style: italic"># Support Vector Machine</span>
+svm<span style="color: #666666">.</span>fit(X_train_scaled, y_train)
+<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;Test set accuracy SVM with scaled data: </span><span style="color: #BB6688; font-weight: bold">{:.2f}</span><span style="color: #BA2121">&quot;</span><span style="color: #666666">.</span>format(logreg<span style="color: #666666">.</span>score(X_test_scaled,y_test)))
+<span style="color: #408080; font-style: italic"># Decision Trees</span>
+deep_tree_clf<span style="color: #666666">.</span>fit(X_train_scaled, y_train)
+<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;Test set accuracy with Decision Trees and scaled data: </span><span style="color: #BB6688; font-weight: bold">{:.2f}</span><span style="color: #BA2121">&quot;</span><span style="color: #666666">.</span>format(deep_tree_clf<span style="color: #666666">.</span>score(X_test_scaled,y_test)))
+
+
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.ensemble</span> <span style="color: #008000; font-weight: bold">import</span> RandomForestClassifier
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.preprocessing</span> <span style="color: #008000; font-weight: bold">import</span> LabelEncoder
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span> cross_validate
+<span style="color: #408080; font-style: italic"># Data set not specificied</span>
+<span style="color: #408080; font-style: italic">#Instantiate the model with 500 trees and entropy as splitting criteria</span>
+Random_Forest_model <span style="color: #666666">=</span> RandomForestClassifier(n_estimators<span style="color: #666666">=500</span>,criterion<span style="color: #666666">=</span><span style="color: #BA2121">&quot;entropy&quot;</span>)
+Random_Forest_model<span style="color: #666666">.</span>fit(X_train_scaled, y_train)
+<span style="color: #408080; font-style: italic">#Cross validation</span>
+accuracy <span style="color: #666666">=</span> cross_validate(Random_Forest_model,X_test_scaled,y_test,cv<span style="color: #666666">=10</span>)[<span style="color: #BA2121">&#39;test_score&#39;</span>]
+<span style="color: #008000">print</span>(accuracy)
+<span style="color: #008000">print</span>(<span style="color: #BA2121">&quot;Test set accuracy with Random Forests and scaled data: </span><span style="color: #BB6688; font-weight: bold">{:.2f}</span><span style="color: #BA2121">&quot;</span><span style="color: #666666">.</span>format(Random_Forest_model<span style="color: #666666">.</span>score(X_test_scaled,y_test)))
+
+
+<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">scikitplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">skplt</span>
+y_pred <span style="color: #666666">=</span> Random_Forest_model<span style="color: #666666">.</span>predict(X_test_scaled)
+skplt<span style="color: #666666">.</span>metrics<span style="color: #666666">.</span>plot_confusion_matrix(y_test, y_pred, normalize<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>)
+plt<span style="color: #666666">.</span>show()
+y_probas <span style="color: #666666">=</span> Random_Forest_model<span style="color: #666666">.</span>predict_proba(X_test_scaled)
+skplt<span style="color: #666666">.</span>metrics<span style="color: #666666">.</span>plot_roc(y_test, y_probas)
+plt<span style="color: #666666">.</span>show()
+skplt<span style="color: #666666">.</span>metrics<span style="color: #666666">.</span>plot_cumulative_gain(y_test, y_probas)
+plt<span style="color: #666666">.</span>show()
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+<p>Recall that the cumulative gains curve shows the percentage of the
+overall number of cases in a given category <em>gained</em> by targeting a
+percentage of the total number of cases.
 </p>
 
-<p>The functions we need are contained in the <a href="https://cvxopt.org/userguide/coneprog.html" target="_blank">quadratic programming library CVXOPT</a> and we need to import it together with <b>numpy</b> as</p>
+<p>Similarly, the receiver operating characteristic curve, or ROC curve,
+displays the diagnostic ability of a binary classifier system as its
+discrimination threshold is varied. It plots the true positive rate against the false positive rate.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="compare-bagging-on-trees-with-random-forests">Compare  Bagging on Trees with Random Forests  </h2>
+
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;">bag_clf <span style="color: #666666">=</span> BaggingClassifier(
+    DecisionTreeClassifier(splitter<span style="color: #666666">=</span><span style="color: #BA2121">&quot;random&quot;</span>, max_leaf_nodes<span style="color: #666666">=16</span>, random_state<span style="color: #666666">=42</span>),
+    n_estimators<span style="color: #666666">=500</span>, max_samples<span style="color: #666666">=1.0</span>, bootstrap<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>, n_jobs<span style="color: #666666">=-1</span>, random_state<span style="color: #666666">=42</span>)
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+<!-- code=python (!bc pycod) typeset with pygments style "default" -->
+<div class="cell border-box-sizing code_cell rendered">
+  <div class="input">
+    <div class="inner_cell">
+      <div class="input_area">
+        <div class="highlight" style="background: #f8f8f8">
+  <pre style="line-height: 125%;">bag_clf<span style="color: #666666">.</span>fit(X_train, y_train)
+y_pred <span style="color: #666666">=</span> bag_clf<span style="color: #666666">.</span>predict(X_test)
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.ensemble</span> <span style="color: #008000; font-weight: bold">import</span> RandomForestClassifier
+rnd_clf <span style="color: #666666">=</span> RandomForestClassifier(n_estimators<span style="color: #666666">=500</span>, max_leaf_nodes<span style="color: #666666">=16</span>, n_jobs<span style="color: #666666">=-1</span>, random_state<span style="color: #666666">=42</span>)
+rnd_clf<span style="color: #666666">.</span>fit(X_train, y_train)
+y_pred_rf <span style="color: #666666">=</span> rnd_clf<span style="color: #666666">.</span>predict(X_test)
+np<span style="color: #666666">.</span>sum(y_pred <span style="color: #666666">==</span> y_pred_rf) <span style="color: #666666">/</span> <span style="color: #008000">len</span>(y_pred) 
+</pre>
+</div>
+      </div>
+    </div>
+  </div>
+  <div class="output_wrapper">
+    <div class="output">
+      <div class="output_area">
+        <div class="output_subarea output_stream output_stdout output_text">          
+        </div>
+      </div>
+    </div>
+  </div>
+</div>
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="boosting-a-bird-s-eye-view">Boosting, a Bird's Eye View </h2>
+
+<p>The basic idea is to combine weak classifiers in order to create a good
+classifier. With a weak classifier we often intend a classifier which
+produces results which are only slightly better than we would get by
+random guesses.
+</p>
+
+<p>This is done by applying in an iterative way a weak (or a standard
+classifier like decision trees) to modify the data. In each iteration
+we emphasize those observations which are misclassified by weighting
+them with a factor.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="what-is-boosting-additive-modelling-iterative-fitting">What is boosting? Additive Modelling/Iterative Fitting </h2>
+
+<p>Boosting is a way of fitting an additive expansion in a set of
+elementary basis functions like for example some simple polynomials.
+Assume for example that we have a function
+</p>
+$$
+f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m),
+$$
+
+<p>where \( \beta_m \) are the expansion parameters to be determined in a
+minimization process and \( b(x;\gamma_m) \) are some simple functions of
+the multivariable parameter \( x \) which is characterized by the
+parameters \( \gamma_m \).
+</p>
+
+<p>As an example, consider the Sigmoid function we used in logistic
+regression. In that case, we can translate the function
+\( b(x;\gamma_m) \) into the Sigmoid function
+</p>
+
+$$
+\sigma(t) = \frac{1}{1+\exp{(-t)}},
+$$
+
+<p>where \( t=\gamma_0+\gamma_1 x \) and the parameters \( \gamma_0 \) and
+\( \gamma_1 \) were determined by the Logistic Regression fitting
+algorithm.
+</p>
+
+<p>As another example, consider the cost function we defined for linear regression</p>
+$$
+C(\boldsymbol{y},\boldsymbol{f}) = \frac{1}{n} \sum_{i=0}^{n-1}(y_i-f(x_i))^2.
+$$
+
+<p>In this case the function \( f(x) \) was replaced by the design matrix
+\( \boldsymbol{X} \) and the unknown linear regression parameters \( \boldsymbol{\beta} \),
+that is \( \boldsymbol{f}=\boldsymbol{X}\boldsymbol{\beta} \). In linear regression we can 
+simply invert a matrix and obtain the parameters \( \beta \) by
+</p>
+
+$$
+\boldsymbol{\beta}=\left(\boldsymbol{X}^T\boldsymbol{X}\right)^{-1}\boldsymbol{X}^T\boldsymbol{y}.
+$$
+
+<p>In iterative fitting or additive modeling, we minimize the cost function with respect to the parameters \( \beta_m \) and \( \gamma_m \).</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="iterative-fitting-regression-and-squared-error-cost-function">Iterative Fitting, Regression and Squared-error Cost Function </h2>
+
+<p>The way we proceed is as follows (here we specialize to the squared-error cost function)</p>
+
+<ol>
+<li> Establish a cost function, here \( C(\boldsymbol{y},\boldsymbol{f}) = \frac{1}{n} \sum_{i=0}^{n-1}(y_i-f_M(x_i))^2 \) with \( f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m) \).</li>
+<li> Initialize with a guess \( f_0(x) \). It could be one or even zero or some random numbers.</li>
+<li> For \( m=1:M \)
+<ol type="a"></li>
+ <li> minimize \( \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2 \) wrt \( \gamma \) and \( \beta \)</li>
+ <li> This gives the optimal values \( \beta_m \) and \( \gamma_m \)</li>
+ <li> Determine then the new values \( f_m(x)=f_{m-1}(x) +\beta_m b(x;\gamma_m) \)</li>
+</ol>
+</ol>
+<p>We could use any of the algorithms we have discussed till now. If we
+use trees, \( \gamma \) parameterizes the split variables and split points
+at the internal nodes, and the predictions at the terminal nodes.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="squared-error-example-and-iterative-fitting">Squared-Error Example and Iterative Fitting </h2>
+
+<p>To better understand what happens, let us develop the steps for the iterative fitting using the above squared error function.</p>
+
+<p>For simplicity we assume also that our functions \( b(x;\gamma)=1+\gamma x \). </p>
+
+<p>This means that for every iteration \( m \), we need to optimize</p>
+
+$$
+(\beta_m,\gamma_m) = \mathrm{argmin}_{\beta,\lambda}\hspace{0.1cm} \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2=\sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta(1+\gamma x_i))^2.
+$$
+
+<p>We start our iteration by simply setting \( f_0(x)=0 \). 
+Taking the derivatives  with respect to \( \beta \) and \( \gamma \) we obtain
+</p>
+$$
+\frac{\partial {\cal C}}{\partial \beta} = -2\sum_{i}(1+\gamma x_i)(y_i-\beta(1+\gamma x_i))=0,
+$$
+
+<p>and</p>
+$$
+\frac{\partial {\cal C}}{\partial \gamma} =-2\sum_{i}\beta x_i(y_i-\beta(1+\gamma x_i))=0.
+$$
+
+<p>We can then rewrite these equations as (defining \( \boldsymbol{w}=\boldsymbol{e}+\gamma \boldsymbol{x}) \) with \( \boldsymbol{e} \) being the unit vector)</p>
+$$
+\gamma \boldsymbol{w}^T(\boldsymbol{y}-\beta\gamma \boldsymbol{w})=0,
+$$
+
+<p>which gives us \( \beta = \boldsymbol{w}^T\boldsymbol{y}/(\boldsymbol{w}^T\boldsymbol{w}) \). Similarly we have </p>
+$$
+\beta\gamma \boldsymbol{x}^T(\boldsymbol{y}-\beta(1+\gamma \boldsymbol{x}))=0,
+$$
+
+<p>which leads to \( \gamma =(\boldsymbol{x}^T\boldsymbol{y}-\beta\boldsymbol{x}^T\boldsymbol{e})/(\beta\boldsymbol{x}^T\boldsymbol{x}) \).  Inserting
+for \( \beta \) gives us an equation for \( \gamma \). This is a non-linear equation in the unknown \( \gamma \) and has to be solved numerically. 
+</p>
+
+<p>The solution to these two equations gives us in turn \( \beta_1 \) and \( \gamma_1 \) leading to the new expression for \( f_1(x) \) as
+\( f_1(x) = \beta_1(1+\gamma_1x) \). Doing this \( M \) times results in our final estimate for the function \( f \). 
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="iterative-fitting-classification-and-adaboost">Iterative Fitting, Classification and AdaBoost </h2>
+
+<p>Let us consider a binary classification problem with two outcomes \( y_i \in \{-1,1\} \) and \( i=0,1,2,\dots,n-1 \) as our set of
+observations. We define a classification function \( G(x) \) which produces a prediction taking one or the other of the two values 
+\( \{-1,1\} \).
+</p>
+
+<p>The error rate of the training sample is then</p>
+
+$$
+\mathrm{\overline{err}}=\frac{1}{n} \sum_{i=0}^{n-1} I(y_i\ne G(x_i)). 
+$$
+
+<p>The iterative procedure starts with defining a weak classifier whose
+error rate is barely better than random guessing.  The iterative
+procedure in boosting is to sequentially apply a  weak
+classification algorithm to repeatedly modified versions of the data
+producing a sequence of weak classifiers \( G_m(x) \).
+</p>
+
+<p>Here we will express our  function \( f(x) \) in terms of \( G(x) \). That is</p>
+$$
+f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m),
+$$
+
+<p>will be a function of </p>
+$$
+G_M(x) = \mathrm{sign} \sum_{i=1}^M \alpha_m G_m(x).
+$$
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="adaptive-boosting-adaboost">Adaptive Boosting, AdaBoost </h2>
+
+<p>In our iterative procedure we define thus</p>
+$$
+f_m(x) = f_{m-1}(x)+\beta_mG_m(x).
+$$
+
+<p>The simplest possible cost function which leads (also simple from a computational point of view) to the AdaBoost algorithm is the
+exponential cost/loss function defined as
+</p>
+$$
+C(\boldsymbol{y},\boldsymbol{f}) = \sum_{i=0}^{n-1}\exp{(-y_i(f_{m-1}(x_i)+\beta G(x_i))}.
+$$
+
+<p>We optimize \( \beta \) and \( G \) for each value of \( m=1:M \) as we did in the regression case.
+This is normally done in two steps. Let us however first rewrite the cost function as 
+</p>
+
+$$
+C(\boldsymbol{y},\boldsymbol{f}) = \sum_{i=0}^{n-1}w_i^{m}\exp{(-y_i\beta G(x_i))},
+$$
+
+<p>where we have defined \( w_i^m= \exp{(-y_if_{m-1}(x_i))} \).</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="building-up-adaboost">Building up AdaBoost </h2>
+
+<p>First, for any \( \beta > 0 \), we optimize \( G \) by setting</p>
+$$
+G_m(x) = \mathrm{sign} \sum_{i=0}^{n-1} w_i^m I(y_i \ne G_(x_i)),
+$$
+
+<p>which is the classifier that minimizes the weighted error rate in predicting \( y \).</p>
+
+<p>We can do this by rewriting</p>
+$$
+\exp{-(\beta)}\sum_{y_i=G(x_i)}w_i^m+\exp{(\beta)}\sum_{y_i\ne G(x_i)}w_i^m,
+$$
+
+<p>which can be rewritten as</p>
+$$
+(\exp{(\beta)}-\exp{-(\beta)})\sum_{i=0}^{n-1}w_i^mI(y_i\ne G(x_i))+\exp{(-\beta)}\sum_{i=0}^{n-1}w_i^m=0,
+$$
+
+<p>which leads to</p>
+$$
+\beta_m = \frac{1}{2}\log{\frac{1-\mathrm{\overline{err}}}{\mathrm{\overline{err}}}},
+$$
+
+<p>where we have redefined the error as </p>
+$$
+\mathrm{\overline{err}}_m=\frac{1}{n}\frac{\sum_{i=0}^{n-1}w_i^mI(y_i\ne G(x_i)}{\sum_{i=0}^{n-1}w_i^m},
+$$
+
+<p>which leads to an update of</p>
+$$
+f_m(x) = f_{m-1}(x) +\beta_m G_m(x).
+$$
+
+<p>This leads to the new weights</p>
+$$
+w_i^{m+1} = w_i^m \exp{(-y_i\beta_m G_m(x_i))}
+$$
+
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="adaptive-boosting-adaboost-basic-algorithm">Adaptive boosting: AdaBoost, Basic Algorithm </h2>
+
+<p>The algorithm here is rather straightforward. Assume that our weak
+classifier is a decision tree and we consider a binary set of outputs
+with \( y_i \in \{-1,1\} \) and \( i=0,1,2,\dots,n-1 \) as our set of
+observations. Our design matrix is given in terms of the
+feature/predictor vectors
+\( \boldsymbol{X}=[\boldsymbol{x}_0\boldsymbol{x}_1\dots\boldsymbol{x}_{p-1}] \). Finally, we define also a
+classifier determined by our data via a function \( G(x) \). This function tells us how well we are able to classify our outputs/targets \( \boldsymbol{y} \). 
+</p>
+
+<p>We have already defined the misclassification error \( \mathrm{err} \) as</p>
+$$
+\mathrm{err}=\frac{1}{n}\sum_{i=0}^{n-1}I(y_i\ne G(x_i)),
+$$
+
+<p>where the function \( I() \) is one if we misclassify and zero if we classify correctly. </p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="basic-steps-of-adaboost">Basic Steps of AdaBoost </h2>
+
+<p>With the above definitions we are now ready to set up the algorithm for AdaBoost.
+The basic idea is to set up weights which will be used to scale the correctly classified and the misclassified cases.
+</p>
+<ol>
+<li> We start by initializing all weights to \( w_i = 1/n \), with \( i=0,1,2,\dots n-1 \). It is easy to see that we must have \( \sum_{i=0}^{n-1}w_i = 1 \).</li>
+<li> We rewrite the misclassification error as</li> 
+</ol>
+$$
+\mathrm{\overline{err}}_m=\frac{\sum_{i=0}^{n-1}w_i^m I(y_i\ne G(x_i))}{\sum_{i=0}^{n-1}w_i},
+$$
+
+<ol>
+<li> Then we start looping over all attempts at classifying, namely we start an iterative process for \( m=1:M \), where \( M \) is the final number of classifications. Our given classifier could for example be a plain decision tree.
+<ol type="a"></li>
+ <li> Fit then a given classifier to the training set using the weights \( w_i \).</li>
+ <li> Compute then \( \mathrm{err} \) and figure out which events are classified properly and which are classified wrongly.</li>
+ <li> Define a quantity \( \alpha_{m} = \log{(1-\mathrm{\overline{err}}_m)/\mathrm{\overline{err}}_m} \)</li>
+ <li> Set the new weights to \( w_i = w_i\times \exp{(\alpha_m I(y_i\ne G(x_i)} \).</li>
+</ol>
+<li> Compute the new classifier \( G(x)= \sum_{i=0}^{n-1}\alpha_m I(y_i\ne G(x_i) \).</li>
+</ol>
+<p>For the iterations with \( m \le 2 \) the weights are modified
+individually at each steps. The observations which were misclassified
+at iteration \( m-1 \) have a weight which is larger than those which were
+classified properly. As this proceeds, the observations which were
+difficult to classifiy correctly are given a larger influence. Each
+new classification step \( m \) is then forced to concentrate on those
+observations that are missed in the previous iterations.
+</p>
+
+<!-- !split --><br><br><br><br><br><br><br><br><br><br>
+<h2 id="adaboost-examples">AdaBoost Examples </h2>
+
+<p>Using <b>Scikit-Learn</b> it is easy to apply the adaptive boosting algorithm, as done here.</p>
 
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
@@ -1372,94 +2690,27 @@ converge. So, welcome to the promised land of quadratic programming.
     <div class="inner_cell">
       <div class="input_area">
         <div class="highlight" style="background: #f8f8f8">
-  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span>
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">cvxopt</span>
-</pre>
-</div>
-      </div>
-    </div>
-  </div>
-  <div class="output_wrapper">
-    <div class="output">
-      <div class="output_area">
-        <div class="output_subarea output_stream output_stdout output_text">          
-        </div>
-      </div>
-    </div>
-  </div>
-</div>
+  <pre style="line-height: 125%;"><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.ensemble</span> <span style="color: #008000; font-weight: bold">import</span> AdaBoostClassifier
 
-<p>This will make our life much easier. You don't need t0 write your own optimizer.</p>
+ada_clf <span style="color: #666666">=</span> AdaBoostClassifier(
+    DecisionTreeClassifier(max_depth<span style="color: #666666">=1</span>), n_estimators<span style="color: #666666">=200</span>,
+    algorithm<span style="color: #666666">=</span><span style="color: #BA2121">&quot;SAMME.R&quot;</span>, learning_rate<span style="color: #666666">=0.5</span>, random_state<span style="color: #666666">=42</span>)
+ada_clf<span style="color: #666666">.</span>fit(X_train, y_train)
 
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="a-simple-example">A simple example </h2>
+<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.ensemble</span> <span style="color: #008000; font-weight: bold">import</span> AdaBoostClassifier
 
-<p>We remind ourselves about the general problem we want to solve</p>
-$$
-\begin{align*}
-    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}\boldsymbol{x}^T\boldsymbol{P}\boldsymbol{x}+\boldsymbol{q}^T\boldsymbol{x},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm} to} \hspace{0.2cm} \boldsymbol{G}\boldsymbol{x} \preceq \boldsymbol{h} \wedge  \boldsymbol{A}\boldsymbol{x}=f.
-\end{align*}
-$$
-
-<p>Let us show how to perform the optmization using a simple case. Assume we want to optimize the following problem</p>
-$$
-\begin{align*}
-    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}x^2+5x+3y \\ \nonumber
-    &\mathrm{subject}\hspace{0.5cm} \mathrm{to} \\ \nonumber
-    &x, y \geq 0 \\ \nonumber
-    &x+3y  \geq 15 \\ \nonumber
-    &2x+5y  \leq  100 \\ \nonumber
-    &3x+4y  \leq  80.  \\ \nonumber
-\end{align*}
-$$
-
-<p>The minimization problem can be rewritten in terms of vectors and matrices as (with \( x \) and \( y \) being the unknowns)</p>
-$$
-\frac{1}{2}\begin{bmatrix} x\\ y \end{bmatrix}^T   \begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}  \begin{bmatrix} x \\ y \end{bmatrix}  + \begin{bmatrix}3\\ 4  \end{bmatrix}^T \begin{bmatrix}x \\ y  \end{bmatrix}.  
-$$
-
-<p>Similarly, we can now set up the inequalities (we need to change \( \geq \) to \( \leq \) by multiplying with \( -1 \) on bot sides) as the following matrix-vector equation</p>
-$$
-\begin{bmatrix} -1 & 0 \\ 0 & -1 \\ -1 & -3 \\ 2 & 5 \\ 3 & 4\end{bmatrix}\begin{bmatrix} x \\ y\end{bmatrix} \preceq \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
-$$
-
-<p>We have collapsed all the inequalities into a single matrix \( \boldsymbol{G} \). We see also that our matrix </p>
-$$
-\boldsymbol{P} =\begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}
-$$
-
-<p>is clearly positive semi-definite (all eigenvalues larger or equal zero). 
-Finally, the vector \( \boldsymbol{h} \) is defined as 
-</p>
-$$
-\boldsymbol{h} = \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
-$$
-
-<p>Since we don't have any equalities the matrix \( \boldsymbol{A} \) is set to zero
-The following code solves the equations for us
-</p>
-
-<!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="cell border-box-sizing code_cell rendered">
-  <div class="input">
-    <div class="inner_cell">
-      <div class="input_area">
-        <div class="highlight" style="background: #f8f8f8">
-  <pre style="line-height: 125%;"><span style="color: #408080; font-style: italic"># Import the necessary packages</span>
-<span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span>
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">cvxopt</span> <span style="color: #008000; font-weight: bold">import</span> matrix
-<span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">cvxopt</span> <span style="color: #008000; font-weight: bold">import</span> solvers
-P <span style="color: #666666">=</span> matrix(numpy<span style="color: #666666">.</span>diag([<span style="color: #666666">1</span>,<span style="color: #666666">0</span>]), tc<span style="color: #666666">=</span><span style="color: #BA2121">&#39;d&#39;</span>)
-q <span style="color: #666666">=</span> matrix(numpy<span style="color: #666666">.</span>array([<span style="color: #666666">3</span>,<span style="color: #666666">4</span>]), tc<span style="color: #666666">=</span><span style="color: #BA2121">&#39;d&#39;</span>)
-G <span style="color: #666666">=</span> matrix(numpy<span style="color: #666666">.</span>array([[<span style="color: #666666">-1</span>,<span style="color: #666666">0</span>],[<span style="color: #666666">0</span>,<span style="color: #666666">-1</span>],[<span style="color: #666666">-1</span>,<span style="color: #666666">-3</span>],[<span style="color: #666666">2</span>,<span style="color: #666666">5</span>],[<span style="color: #666666">3</span>,<span style="color: #666666">4</span>]]), tc<span style="color: #666666">=</span><span style="color: #BA2121">&#39;d&#39;</span>)
-h <span style="color: #666666">=</span> matrix(numpy<span style="color: #666666">.</span>array([<span style="color: #666666">0</span>,<span style="color: #666666">0</span>,<span style="color: #666666">-15</span>,<span style="color: #666666">100</span>,<span style="color: #666666">80</span>]), tc<span style="color: #666666">=</span><span style="color: #BA2121">&#39;d&#39;</span>)
-<span style="color: #408080; font-style: italic"># Construct the QP, invoke solver</span>
-sol <span style="color: #666666">=</span> solvers<span style="color: #666666">.</span>qp(P,q,G,h)
-<span style="color: #408080; font-style: italic"># Extract optimal value and solution</span>
-sol[<span style="color: #BA2121">&#39;x&#39;</span>] 
-sol[<span style="color: #BA2121">&#39;primal objective&#39;</span>]
-<span style="color: #008000">print</span>(sol[<span style="color: #BA2121">&#39;x&#39;</span>] )
+ada_clf <span style="color: #666666">=</span> AdaBoostClassifier(
+    DecisionTreeClassifier(max_depth<span style="color: #666666">=1</span>), n_estimators<span style="color: #666666">=200</span>,
+    algorithm<span style="color: #666666">=</span><span style="color: #BA2121">&quot;SAMME.R&quot;</span>, learning_rate<span style="color: #666666">=0.5</span>, random_state<span style="color: #666666">=42</span>)
+ada_clf<span style="color: #666666">.</span>fit(X_train_scaled, y_train)
+y_pred <span style="color: #666666">=</span> ada_clf<span style="color: #666666">.</span>predict(X_test_scaled)
+skplt<span style="color: #666666">.</span>metrics<span style="color: #666666">.</span>plot_confusion_matrix(y_test, y_pred, normalize<span style="color: #666666">=</span><span style="color: #008000; font-weight: bold">True</span>)
+plt<span style="color: #666666">.</span>show()
+y_probas <span style="color: #666666">=</span> ada_clf<span style="color: #666666">.</span>predict_proba(X_test_scaled)
+skplt<span style="color: #666666">.</span>metrics<span style="color: #666666">.</span>plot_roc(y_test, y_probas)
+plt<span style="color: #666666">.</span>show()
+skplt<span style="color: #666666">.</span>metrics<span style="color: #666666">.</span>plot_cumulative_gain(y_test, y_probas)
+plt<span style="color: #666666">.</span>show()
 </pre>
 </div>
       </div>
@@ -1476,33 +2727,9 @@ sol[<span style="color: #BA2121">&#39;primal objective&#39;</span>]
 </div>
 
 
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="back-to-the-more-realistic-cases">Back to the more realistic cases </h2>
-
-<p>We are now ready to return to our setup of the optmization problem for a more realistic case. Introducing the <b>slack</b> parameter \( C \) we have</p>
-$$
-\frac{1}{2} \boldsymbol{\lambda}^T\begin{bmatrix} y_1y_1K(\boldsymbol{x}_1,\boldsymbol{x}_1) & y_1y_2K(\boldsymbol{x}_1,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_1,\boldsymbol{x}_n) \\
-y_2y_1K(\boldsymbol{x}_2,\boldsymbol{x}_1) & y_2y_2K(\boldsymbol{x}_2,\boldsymbol{x}_2) & \dots & \dots & y_1y_nK(\boldsymbol{x}_2,\boldsymbol{x}_n) \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1K(\boldsymbol{x}_n,\boldsymbol{x}_1) & y_ny_2K(\boldsymbol{x}_n\boldsymbol{x}_2) & \dots & \dots & y_ny_nK(\boldsymbol{x}_n,\boldsymbol{x}_n) \\
-\end{bmatrix}\boldsymbol{\lambda}-\mathbb{I}\boldsymbol{\lambda}, 
-$$
-
-<p>subject to \( \boldsymbol{y}^T\boldsymbol{\lambda}=0 \). Here we defined the vectors \( \boldsymbol{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n] \) and 
-\( \boldsymbol{y}=[y_1,y_2,\dots,y_n] \). 
-With  the slack constants this leads to the additional constraint \( 0\leq \lambda_i \leq C \).
-</p>
-
-<p>Using the <b>CVXOPT</b> library, the matrix \( P \) would then be defined by the above matrix while the KKT conditions would all be collected by the matrix \( G \).</p>
-
-<!-- !split --><br><br><br><br><br><br><br><br><br><br>
-<h2 id="support-vector-machines-for-regression">Support vector machines for regression </h2>
-
-<p>Material will be added here.</p>
 <!-- ------------------- end of main content --------------- -->
 <center style="font-size:80%">
-<!-- copyright --> &copy; 1999-2022, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
+<!-- copyright --> &copy; 1999-2023, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
 </center>
 </body>
 </html>
diff --git a/doc/pub/week46/ipynb/ipynb-week46-src.tar.gz b/doc/pub/week46/ipynb/ipynb-week46-src.tar.gz
index 142868d1230a4e3743d2ba1bb957bbfdbba502a5..233750ecfcfeabb775aacc5d0563124752e80ec4 100644
GIT binary patch
literal 294343
zcmV(zK<2+6iwFQc5>aIU1MIv7P+ZZrHi#3P5F|JR5)u+z8g~s4+}#P*xNEQkm*5@<
z9^Bmm1ShyQ?(WVU?tAy8bN~0hf2!t<)y%Hy>ci&9K6@|uzO~kF5q$@Jv5%I<_6+p8
zhWgfq#&+~J)}}E3t&Ne9k)4r|;_(~A&i0RQMrPK3`+jUJ%<L3QtZdAz%uGy->>vt8
zCU!;^HW&)V|K4Hz=graHLEjD-ij|qUv9-Oi_5bWV;4~AH|KdLX_9_(r@%{gnhKP?0
z0|x^mqi;wdt4N_n@$c=i4+aJJFaw`#;Pd$R#Q(lq3Mwfn{QKeK6$!wH?_W-V{nL+U
z2)3fvx3RG_rnhy}x6`+F_-JhmW6QwE$ic)7$c=#s2MY_gy_vq<M{83fZccU~MrKh#
z5jqYLAyzt8F%b}*kRS&W9itEnvxtxg6N{)A$A81L{C}m1{D=O_>AOf6>l+!{G5mMi
zf&ZocgP8tB|3PdZCKw8r|DMPH*X{q~`p?2hVde1A%9xjlos*M;iHU=gjh>T<g&D*I
zVrE5UVrHh0mJ||H5SEZs5v6z0cW|(y|IhT5S4KpTRYu94Sw=+YqlUSun~a--o4&Z1
ztG+6mwT6g2+Xo9tXBi6>Aq8PZRs~T>MrkDlaRm`W2YCxoRvAS`XL)lOR|V1cqB2S!
z#1za;Q8_^rioipq)&Hx9{vR+AsQ-6Q{QpDRAN~J7%lrQ?^q-xbk?nube>Qd&_W$4i
z|6kF*$bAsUKqY!q2MkFG5x{RFo&wJP(-S11Cq6|@68M6(R}dG1DIO%*0{##&R+BW5
zk%6HFj-SB527QD<fcZCnPV|o(!e_w3!2hEM=*xin=YJ7TGvNPu4EO2r#L!kRaTpi@
z7)cR9B^TJ;c|;dnNg||0EN85^3}5U|4bSFC@tte1pX~34xZ4CF;yY`Td}=^&R>z4(
z{Iqe+P`&YrEUw4sc$_z8s~Bi*?)%m?_17h@7c}6WA-)!XMV0e`!_9#CPYz$vVUK5P
zc6Z7D&rbNqVFm(6#t=E&e|0MgxF;eJFjhv--a!2WyBH1_4u}0ZIvKxLD2;4dYkPb9
zaM}T+h01)4{KV+g>M9)PB`S5ju>eHD_rS2rwVdwky>|17a#xFYXdeQNZZCqv2Q$(b
zx3lKn>I2+~J(N5;*s6}&Mq9!atESL<I?u5hPw$-#LT@S8y9=t><5}06b2^3G4`Hky
ze`g5z!j)E4#R^M94LC}T0@ZL+e7Rw$Sn<#B5neB@tp!}2Y#1)g)ww(@!3X2f1v_m0
zULMVrl}Mugvc%&#tiucpcLLestW67Yqo6Z1+7fnRq{%<&+`PN<t2Q)^1_HYl)Xh|v
zV@qxf;RtZlgg2KG*khyzYjwG3GGuv6J)BxKr5t$j>NdaJY9aTW?HGN=WY9TG_Ja54
zml&F-+$!%JcE?aixt&DjYHi0kZ-t>IvJO}Q^16MHyA)I>y6@g>O;)lW{PD(6kj{q>
zuNAUw7zYiWqWTxNZjI6>a9MGENaYJltT4$m8A!@>7o7O;=W0DWIFeydon|`{)Y|*!
zwswAb#w11jPP@zjWp__1&{WaI3ia}>&Fv)$tJR3Yz!*7)O-74xf+W<P(jM<LbknN?
zb*EcoU)z-~K_m%XUhi$H5BJ`A!MSo2N{DlFb8>Udth7pbK8Y-*f!kGToPT7=XX}ZL
zyw+$PfhEeHP5};7=`<cfsd)OIv!IeFM*VghOB4nf-wed)S2DpYxgRL>qm9f*=DhB%
z5j}CFpoCJy1r~Et{8F<A6RPvJpmm|ioyoBmSBt~>)o;voe)n_q65WTexjOqhTACp&
zD&Nbam6fv)(;m$XY`gXTF9phlfg`c5gn#7S_YmhZkq*JSBebyVp~C4ldk9X_YnQr=
zz{Q|e*k9~xqd#YT4qj+Pd6rW*cq0GZEljR%Vx3Nt(S2}EE}uL5?cR7UgGQzKUhd4i
z`+0`g53W+_WHLNj<+kGa2G6hR%?2-f<4Qi)qTu8oq&7HRaacE=Tpk%#9!#MSvY3t_
zt)Lln_Q+q}-!)`Pe5W0kt5vflPv7HqdLB*;rk09B?~A7o6rL=;Jl%>pqO^t*_=HRt
z<kNw}oN_5B-Do3FD>Fc3HXigxUF$=B$Yy-)`p^b1op8y$(CkT!qzcbfX%Vug$sW{O
z<+QipCn^=^|9+lQKA&go%5dI|<7_)@i^ZgUU|?Rnihy2AWFA69AsW1@R_{VtZc-&W
zFynYMxztgRg1%5=1%XvjiJPj`C;k0hDn!F^dueMlmv1|Yl+d9>vjz4;L#<v{<(thM
zj854kW@T;7@Q0Hj{;3*oep=N6|LPx_#C+_DEPeXkidzcbBYDN0@R)lcaQ)eT%n~Cp
zu=L3TKa9ko4Ap&z&q<@J&G!ED#r&$B`@p3cbaUSELI`#Xn_2dO!OSUP9;CEvCf;|+
z3m&O`rVC#(CRZfZoXzHBUB+?uR}9T-g+_$?x!R|AZ&gLpz3#>r_or-EqDu5Av&-Ly
zaGh-n3qex2oyjo6dB<GOh*&$z47)KVM?pr_wf%13soF1GHD+I;Ib4swdZ~F1C30h3
zq2Y<uhbAi3iARv-dpyjPKeP&llJX0Z>XzC(%}!@A9S|GjaaAU@n7D^c3`lzm<}>Os
z^7HizLfFxG=AG^@lU%dd6_)&3xjKcbdvx>}DVL*I&O7xeDrYiuGA?x87B7raD?WPM
z*1?UTbV@6&Mx&1~FX0C7FORa=^Y{?E!^y={IBaimQ#jq(jb7b;Lp+g}$IndVlW7`Y
za!#wDo%LR%7%S97ej7)p41aX8G4S&+Q~2-#;sq(@aM_HkZyB^-&k#G?HXE(S)2K23
zA{r9E*b#(f<Jgq7y*odOLON5!;$dlJbywNPk!2dOZ{u`7Lbl@ZmmtPJ3gV7eIS5J+
z`6&;N*Q9C}>iq1C`;c5z0DCwlP=@PrcU{HK(L_iLvigl=@u0F%q+NrFg43xw8_t_D
z2IH-173tJoTjr97kr7X;sWVW1aeVwKq>s=_uj!4^xu@^#Rhi9*LFfq5sV1vMU~ww<
z(r%sojDRyi_WM{I<G!ydjw++e-!wIHRA#%p!^yY<t>@iaXNfl<HCTYj2tti5T>KeA
zP%vd=HBsYd=$u{lK8s9EtKLJ@JwpzMIbmN{bAP@i;u(!>fj#+0O8xBgl`0TIflB_j
z@jJ2sjY_4n>G*eFZm*8@Z~eb;*c|FbP6B3-g4wuS1%CN#XMA*KC+xj)q5k|eMxJ6r
z+<AT&ghob%(kz?#VJ3$jr#1aXqrq5!SLQm@Y^D)0d(SS+hwolFhC&d|%6B?uJLawG
z2YylOhWhqqBNXy0cJ@Tuvl&On)CT8+kKIy2WP_|9C$#1j2}3Wyyl{l25rizSPzmUb
z@)$zG!if4?ydQk8dAow~q{ii|++^15j+S2vPy4X1UdvVmiQPBhVGncm=l)T!&pvQ7
z{F}_sxPd|eb$6z!uR$o_cmdMlfqGbBG$LfxL$bh-#&>>ie@DIm=)-w9K3!4`zQ)hl
z(Yvr2oeJ~sT=Z9-H>d1ljyu=Nj{B8wVw%qO+#Gk@yzf@O_bTOUO>=A4J1poN@QmO4
zdb&3@a62i^x;q~zZPYI`TIUVbrlLH1bxN;R>T^*uL_U7xUR=|dCGusN*JMoMAmZTR
z=)>37Q8ijp{NgS5%K>P&R(%=1@QsS!-+%Azt3JJ&Sqm}u?#tmH_iAbO9BU4mg)EF@
ziC?6#+6qNZM(@;sCnvaN)8@A4tTfs4>nDnJzMrpMZbnfld4hGC4r{F3$WIR#$oXBs
z3$<;UXi{0%miqcj^P~Ae7iVwdl~QN)>goee+b&Fd76k)$w$BK?3FtIBHQhfiWl0~e
z{%Rs^;JoY^H5qsBb4=%HEGedJL9>d|IM|)8$iExQS9G^Po4Do>FCMLuGWDzQnopgr
zDX_=fgKY2`OvzRK80fi}FgbL$@VnUtolpASd2Rn5nrtal`LW#OxHBPf1u=F?`r{4p
zeK%+L7yot1BwUE!O7#ic=#;fbOkH=&rVI;}rwClxtsPz@x7e<3GMV|+nlp_89$~3B
zw(#<NkNACjIcUr@YoIYsj;_WUA{N^ltK=2ah4GYaJA%O2-QfU99fqyR(L||4ML8Km
zSAAd?q=elm_Qu`a%<r;5DG;47lF3kvt>Xqzs>8#>*22jW3NXf|r2AsVj6$QDe5~;t
zK7j3h@M8O<rl!uwSVw7jsurusd-jE%*V?YMD~4Q<ja8jqpB|iSFysx~A!FuYgUU}9
z+l2A##7**cz}yI9?5^D3#D^9lO9hWj0?ru`t1QNqDi+Qx(21F+AT}U79_L~$w=z9%
zDG4%^T%(jqWFp$#!IKUCh0mp9Ug88pslM!aCGE75&vUNO@C^hWyt;DpLD;8<Lf5eQ
zQqku=Fs>)-zUOjCP3~tRq0gC=p1e}$aqc->;_x{1@d|E%#N{il0C4i2%<vTyym5VE
zBLLyqr*aqk0~-#>5f)m!!OyX=HXW>T*Z{zhiFER^U#8>wY-xQzyy@mVU(m=Zjt2lc
zs24T-_zG3hdnd4M7Y7herBqf)g^uQToeidwm_>Jnq)rFZh085cpGmFo8#1N!h<{bV
z>a9@Qt{RUhJ6}97WWUXHIUQ?MS6uCn5{3lWYN97TSet=iE(bq(yS6pO=_KB?gxvPR
z)bW*C=|`de(5PlaUzeKpnQ_{#3?b@|cVC|RQHY8{jYS(8^X@fq^c&;euiKe^wyA*d
zQ;MJON+dD|^jc44Gy~_R);n(JLHzx<k5aSgQ6{Ndnp}>S^E(vc(YH;18Ngcwww&AK
zCM~hL%E?K(M=sl~hf*?RV7S&&ok0s|LJ5rFX7WLk@eq_7|AZZ>wvJQ6lSd6-;C{aU
zGXRZ1x#*PnyHx(3)vR+~n+hkt@HKh?k-D0)taIh%?f%T7_qufA={s^rR(Gg!MX7Pw
zcsJI5{`3{M)BYH~*cmtW?n<Y(v-Lu~aZg<0F|!4)d$nb0^<dty6fOfMU1#`pe=>_^
z$Ur*k*64SZlxH}^{O)uXdvtJ;^9_zHl|M8mi-+51^3D;*bIP*B1<^eHmEI_3s-(J^
zw?Qa70xjV8vCwnlO4JCyL~C5$L~&g3C)=&{f3cV>DFiXDYPmiW-nSnlnWn_@FnZP0
zyjV)&UC3J*38L<tb&js52CUz-&E{bgRcY10x9zCJyYg2)w)GE-d__ebe6N}Liot!K
zny)Rrc_ScDE_D56E3D^9efloOANgyQePU+I?x$R@@DsJx>ki4Nl~uR7@wH0jYa^A*
zFk(<fP-Yb2ih6~~&lJ}a3ivAkxV8;1QH{a-b%haKyOXL%PUfqENyghRdiu9%WYb^R
zE_cqnsEzylj60lbqw|+yWF*7=_t-)e!@h|hv~#jG`yXh(MgVLznJo>+_<ej=`P$pV
z?)o#0*^(^BS*L`{8)}Kysgcrcb#b-4p4aF#Rl~wn<`Z(6!PgkipP;l*VKF9ESq1>W
z_an>cGlXeY!qT_^hqZw~?iy=>1umPk7MKb4YFdwxC@S%hY8eZ3a;{E@tR5N#vcnPn
zKym@BJE2*N`vv{=IeEr@Nh1nHi=UI#tXaqtaN91YJtf+dlkH{>(w*^BJ;^;Tht9ka
z!!f6@?eQ^Qr|o&o`I-UE>bx833mI9vbd?hILN9S%_lq_*tE2gxFyfsc_*=~V1I|`d
z7btVux{CPRu6q^ylvo0arY*oRk7^@*4<UG)gJkp4%Og_=aa^F#Ln_*L9#1MhL&q<$
z#~@yNO;RFO`E6W&;*8rSG4Fa4?&i|wXEB-NOr>eN0*{%>kK^w=IT^g~J%;i7Mf)bK
zD&;^ZmA;2b1))dj0CmaKU&)D-<9?%%-C9IAs_a7XrlQrg*j&o+hnJ+n#%aD{q;Rq_
zJr)TyVU1|@ndjBS%5CAN``f4$`6|0;xP~~A_%nE;VP4ZTkIT~~(l|QxBBVMrT>2M5
zQJnTBmY0S<X4tL$dY`ODR8_HcMUeVMbp!!V97q6p^Qn|*qq!1OQY#J+v1QQVq9@Mt
zyY?!b2?%L!KF6_^jMjSCpLgG!XbPH%C%maJ9hAfJrZ$A!!w1TVAJ0x#yFM!zFAA8u
zIxalMF_GMkWHp?@^|(9~<n<X@IkyNw?Z~tsa7^)rG5{}X1aQ^wKz7ecFQA#da_jUo
zApt{5cZxcTaHs&1K6tJ+-^9$c&xF88g8ZucXjDVXMO_aBxO2W4O2@jox_lefX5^`7
zG*aP|en{V>Ha^t4p5WW9_2!e&K3AnUx-nL{ovk5@tiB|Ght%GdSRy;pp$Jkjoft>y
zNiv&xo^7+OQGr31`w(xdkg>OZ>zPs6SO#=h3aegMA8?$+mzSj8dGPWzA244xt4U0y
zvVJTYBgS)UzTey0%Ae;}otiSjJlDjVPj!0{Vx_{t!Ld1hm<%ON{RkYwkbdje4bYJQ
z4_HEf%&jXrW^PTSNxx(xF}J7}0e5utj*=3|t<jB~ksvUccSsLKoj)vU2<}*wN6_^M
z#O}#fb-;vPbEV?7T&+zeO!7OhQV$J>q!98i&F8Q$T`v_fS|bQ=VPlU{8Cr}XA&fWl
z9iC8p_cLAV^XsQK_hn~L6W&6BoifbaBIlWxX0ETF*$xx1Evn|~j}sY0>(=w_$+!)B
z92C-+f=kIHx1_~~CTSqw&pcV``2KgX)M!6mz5gnQb`||<*xX1^2PMqVG@h-8k=a+J
z(=U@An+TlZ@0R5X6?5`U$5}+bOJ6q7hl)el<3vvPM>&dByj!@rS$!k8@IiHw`?XT#
zb9$}X=SI_uxRF7x%WI3?eNJYw#|Q8*$xs?E+F<_X+&HY{JB7>F?pMcn%HqdLai!H|
zKdFDn9%BpruCw2e2Y59<`IicC_#5Vjw1Fe9a@jwt#1A_-d?6d)3N4C40!Z;|&j?3n
zl;1{i*nXpENaM7BMZ$0Qyxhi0B74NJ%V75sCE@l)BwIEelS({7@QZ@6&0X(HZ-S)p
zk~*)BbilT?iYH`(zzZ2-dt5W$+VB&2?UB;>JcHYWgYw-->YduY_r)J=ju?hMXICl_
z^Xv{K9PMtd+Z9=dRMgE0pZooZZ2_L(F2mm?1xm6k5skzuI!Qs@r)zx#=p;M{nl)D9
zL~q|#T94xtP8R+2w>2DJysac;k;~H)bl93h2}@!X{e|<=X8+c8ch(!`SM<xDt@n2$
z8TVkotp^mCuAhGFiK4n;P?SkwB~0aZRGN(K#d#@>aKmg|=Gl#Tt5TBIat3MfCV!-;
z--Gc0U@0%0Sa*`Lw7(y25zp|#>-LhC-*I~^Fo?C-d7nade{OTQIdzG|enAi2)4sW3
zm-s)#=8{q`?cuz=K?+(L^q%yDntb854$rp$iA$$5J&S#-EK;EKW~AJacz2-!89cX;
zx1Ywf_erw`T(}n!PoJfEK~`&jLC@v9n-S3J8lE$txp;ST;c~Ihi%HJ;sRj(W?#muR
zy*QY?Kp9HsS9P3;f2Yr3WEi7vr8{nLeG|XhCSpkUPWHq^M!wjRkX+QLwfSokuhafa
z82jE#M_U;;bGB>*;5ZBHYQPQc9i#97I1#8Ks$w*nvT<>7xy<WGae6KcN!fpOBlI6{
za!au2)BHf+{~d~HA;7J3WnHcH#jBM&5F%lb_-rZW>Wv4lJ?C<f%W!6OKexQxY4t)U
z878+-E$IXXZgc0hh<5NHjTvvmpzHi1hF6u0e=^49Hac398bl(M$oxG?`{DY`LOAWH
zD}z$FDhP#adjM|5W4hV>fDDs_CvSl<@{kKBI#*57-EyH^6wk3~C0H@{Lt!bm^}h9U
z7N?YIG6CB&xi)9*^-}O4I3=OuPM`cwz)s=`<LU>A*X5W{>+|G|UNK7&c)tDz(rRD)
zThm$wyx#ck&+=IqH_agG)o?N|36nah=;>RPIF;rQqVqzPe1IohFuZ*Kby)ViTb3Wc
z@b3DoFv1XEXOqi|58jcw{I9{d(#d@A-Ixz+R8GxaS_{;Xh&EW{JU-DgMGxH(<b^J`
zIPon`KQzqdN+~UaOa$k-9LL_Aq-<n`2!q~fSCeQrI7}2WC4=p~Ryr-Mak&4igO|Jy
z{B_Hdo>8a>--}|Ox<RuzMwVzCY6-ns&T<<|y*Ji#wSgVit<98FJ{a8G+&wh4Z$yuA
z3pVeNvE&NAHqFF=V5w7NG50WFAs=T>6C>O17w)d}g(JxMedFWfgX<u<s<M0fLa2^f
z0P6oPlZ>xStJ>bHTsVI3bI(7h$UT#8B9ojqECd42|E}oYp4IXmGI5NV`V)*i41q(h
zI|=dONrf`D%cu!Eq3fqxUj3!~PPes7>$NGI7E_q{Vd@nd6!Hpa-uwZ<_$T&+U*gHR
zoh&6>NO-qG=GJ@mo!jOEF}6F(k{<)$1$>D`KgKl?%F}6$Y4I3z+tEC)PbEt*vB>!p
zz!+LhZWzXWs-4C24H(@<NG&`9k)odlm=bpmrb!O)Zbbe>-3nn$Hy>0v)5M`O2G8(%
z;s!9QtE&@u9ElXNCBoIp3^E%Ohp@pv(2KAhft*{+OR04lew{{_Jk=UZ>eH}yHpxuF
zVyFAFvQ^_<U1G7|$xN!>xl%!t8vPFwKQylM47>W(Ib64K0eq5m%<W_r!9VdsRo{Vx
zl*Hz6nDT6Ed=!O{IXkE?k!5o@AMN~LI$(oAM{nwLl)2ag_-y++FP<AZwek5i9LaFR
zYv<rSD4+Gj78XVR1-ZQxPL9c5L9E$)G3hqQvPkHlN1(nlJOV(%+r_pmVKr+R%_ZK%
z8~HI0cQFjQIz1bvzjqT^Y&Q0_#{i}#H~&7Ft_6JYK_moUM!5p~01HquDmCRedS%os
z@o+i#v+^gU)L-KHB{Zj0$;f(^o6GZxwZpZx6Z|POnjKJ>jeE1t2lgW77*^a(tnUn=
z(NU$h%A`);=_>^29L<ZYRsba0ZY?ABr}Fj{)iZ<)zON;1C{iD9twy&7gn;CGZ=7Gy
zEQHTsCrGvXV7?&;i@Y(CDhMLS%|H7J3kz$m(G_j$_jl3lvHZIfuZMe2Ja^p%xX(|i
z0@GL4qnD;SP}tQimo_)e=(P>_p`Pr4ViVZRfW5B(9d`JJ)`NZr>xWb*$DJh}Dpyh6
zaIu?jGD^(wl{oFLa$<QYNa_98*H2Li=yCNN30X|g51(WGm2G-~Vk1BI*`a1i1v(3b
zps3e1mLC~!+oxRQj~ZJ@Vf(Ffs{W%P<B$o{U)icwTtom`;U1vQia<fDS8(!KVr;Vs
zD~M?LtxJB$N_m7REL0We!d&yUw!#4ik}<DjBvc2k@BMr_hvFfuP`NdqE7ZgE^Ztv4
za#JV~5`5zEaIZaJfb~typO1v*Ly8<>Kf%{m1Dblboe-H`(+6Mp&v+m{r@aL$v-rZo
zx8`F7eo87Iqh3_u6^!Qv%+`R3!*!cI5M`3s{s~I5e<>awt%!FVeJ-a28})#bzCgW|
z2cnQyQuvid8T254!J(ivdR*aP5V5tzGw7nj&z5T!8NKO@ASZ{vJ`3oaCFgY_(Wtcc
zQ|yjd8b}7^$v0p|)6}#A@_Kl>rT7?0E;H;76rQAiHK)lQ2DJ*zs5XIKMvXhWEKunh
zfBdO>D`NAB;t-=Z3YmIQ-=->@OEUSHXB(B1t0j`#h_2A^0u@UZ0;kuV4;DnDmS2`S
zOxr0HC}0pWPiHhJ%4)TEk*>W}QDK=XR4MtQTB4JG{Yo|wwVUk2b7o79@+spWnF-6t
zMxWm)bthwNqsIOIfnmvufL{pw#!S@=nD%VRboud@j$SwC12(s+AR{NYlei!9**5y*
z%BkGWIv6uxtrpCA(uPdsKK#wKF?f1<Q=6mZm;I?5=+4g0w~l)JL}L#aA9(8xM?PIi
z#Jx+5aH|!DHeQO@fR}zVhgj_koeOQlb-6jGBLa=LcbKvCc83yuskL1xX!8V=$K&oA
z?=c~eKT5wjKZ@dLG1&z$P$us-A?2GOEFE)-2JN8m(n{$to;rKO^`hyXSP;KKSLhe^
zADT;xnZo8Ux&SrV!8aL5^OH#iWl80R_r!y+u1|L&L_^{)KmI^J<>qh_Cn{vM+1~#w
zHRBiX;-^&m=%`}s-f$h-V>YF;H1oA~FWCx>bZ5c1#qAgsAu|X0a5v;xWj~6>X1e>u
zf#L7YBfenj_g}FUX=H`MNv1Kb03epPm%wBwP%Sk*Yq(K7Z#G)ov@INXx;0WOFD&e8
z{g+B9ApqQ0X&+%$@NZnUJKZ2l=P+AuEcNxZycPxz763N^KsAYU6+;zM#U>U(;tPj(
zTsRX>#$Alj1NJopAq(#bBV)A>Q~Hf(xHkHKscQ3-+MR1wCT(#$OJaZSJuRn#uQ7kS
zfn+q8qCQ~beR|1cGL#xpZKXln>v4ap9#EFAT0=kaaAX+S>{eA!IuHj>46+hys&}-J
zwh)gWOdWHzE4GSzci(j$j|_i2+}iS$%|b=1#qQToW(<w2{$-CmM=iJn^@pKW>);^f
z@kk8MT%`Jyao<<LYRhSxiv`{kes5TAr+pEfJ>P)FB?>T+<y<LdmBsjP*%c79*_9Ro
zV8kn1;;AQpnJL)}u`_dtR4%q&z)l6@(Pq3~d;gr-j6&i2`|q@BRR)A{bh@w`h-fyN
zn&{^M#xJd5*vdLK_!(T_Nmx`<S}L&i9uK2iF-JNRsqQS)Ve5B(o5ytgvGMzOm5S@L
zm19p?V%&HTf=-jyccIhG?r)+Yftnh5sR$@IdD9Qfvky~8D|s5<QMb+f)~-%AMnj02
z0k{?V^^GFYc^Js3y6J#=$&SbZZRUL*$yU5=0{Ke$@2)}Z27So=o!LYsipfAiF($wJ
zxmEx*m&YN@`?z<$-81kZINy8qRyUbQhCxWH>`>N6@-ga3Z4FQSmtZWSNPJE;48qdS
z#6K4^*pNy{ZoWfaY|~a{ThTW8DOI4B$wT`Q{vO~A7mBv+2z<z0#;v>=M*2%Eaksg$
z=@jHEtNZs8r8oQY+#OB;2#;dvjyRW2=R&+XS?j3Y<7RZ6aR?;^kx3CS2!*T<psyKs
zT!gxjBi~I&fnbcVABJ_k+nS}-g=eoZo_~f%#?tT2hS)B(iz*dp$P6q%>RqV{R0~b=
zB+2wdBFMBc)}>4yV=%(ev4twcy*7T(il6hM71mR+?wKVr^xA)-q(EEdb2*gzITZi=
z8e}~uumBfOs<-r_GlT%k+S;09!%ib0B5K=dZ~9rN20Xpp#-R3@9EJH9c^~-B@shFJ
zbZC7A!D1)WyhY?_xj?&|O5cSV{ezoT;DaAB7UQHV-}Tv6SUBm-3P4-(R4VXu1`?J`
z1$0f0u!xw)#P6=de^}wyrtoAQ6dY)A+J7qnIN1Hw26+mny$afvVF<}|MIoYB-`A3i
zpASPJbjkNhV6;u+HO=KU-4-ul-Y1{&-o|mA9xi3UKyVoO;%A{<$`cD%%*~)GjuL20
z=c|hy>3g=|c|&Qt+?l7D%TArsgr3$%(cmFf?|wjERP=6qxvD!NT`tqgsBP{cn)>U9
zp#{(3jxu7EBCUWB0tP{YLt+|~D5^{s>ahs)(v30v4AQKD)(69RH`b415&`hh$e1`}
zro%q>k#amGI!$6x?E#^M^0PmK@gEbDx2sZO<wkrE{K&>nh-i5cmQvAG2-IR;$rE()
zKg^YjG+PAe`ZWQrcIMO?p}>FNtBHQ>BRDhtGm{vAh-2Sc%NcDLtmO4yLF;Nl5)AUG
z%wBjY?_cjP0;p_no>JKhjjE4@d^n$SZL<PtL;C^prPO$f6qb}nWVhD)MGziYZ|)Lh
zn^ItQ;z74(;t~gkSXZ{(s<P0e06blf!D2Ef)StlgNTZ(*x1`frdn~|h)a6ZPP!fUe
zo(hRC;72lYb5^!_N-|6EyKFMP)2dY4BbY!sap$}uWgUuVU|2oTjk$m2d+)h@`cZj2
zQorL)$#r#D@GHIYd%71qmYovs)a#{V^EGqB-3}Kw5KR68Q7MSuTrA$C%(nI^LU(H+
zq@<?W9<{snC$h0<6+gTsh!xc#=l6mmd>%o{T^&)dn$Q^#k7)I>$?fcYtYpHl>8=Mk
z%~i!{vmuREof{ra1nJKzi^-^0VX<y3t?A!rrIybi6ZmbOHxuyF<<o5!XVfI&00{3f
zG0nw4$H+@N1f1DZ>xIVeiQn1ksu3r%01#kQHh_w}Cj$??7$>-k=$<EyCA)Y~aDx%S
z2!`w1Ah@i$I?&@*s`*4QTVhGKOmE|tw1;Poryk9(qRS%n3M4fVd<dOtwWUvY$BDo<
zkuPH%QHPctB|4g3efM;qZVS(jJ)j;n`t-7_Ja(%xh%WQ5S|@1{0i?y4SCZy0e(1CO
zT>~{ZyaE2JHtMO80DfDO$5k-<+Xeiim2Rvf9qGhV_tG66r(KRlJ~v90el^<gmV-Mz
zRa{@x;D;qDaix$LO=|Hpo;oVIooOZL0Nm?V!t19;#bpCf51@Q;Xb3y?!0?3lMQot{
zOMsV&WPbZ;vp;vm*jVroZ@y5#mc$PJa_@(f57&Ib?|JP{#_QfjFHTP_k%H@rhTB<x
zMvNZo>gt*w;J)V4PW@}R%${Jb$z7E+UXFP*XGma>$2E#4_T?iM^}M+t9>m!1nk><c
zqBFH7pNzHtSz=SR?8QgU=g~U$arAC=$M`|5hHSo`J98h5K@qYY=p4M+vT=C6a`qo2
zT0jJQ5&iD7P5hJ?)XT{|jr3jqOeD;&B;Zx{+so4}OI4Oen6B#vGkm_x{EX-^Duve@
z-%-BycBJ+r3ofl13?Z|Tz@&BfRJA8^N<DY`1Us5o7-0ze9{1v58@xGVi)woF=GGYe
zODWv$Kz`6Sl~lLK7zr+|(y&dF4?7vJn}5^Id7<xZG6=l9(jCDUtW)nn7(vF{^;uuo
zbG18^@Y&VzdI0)Y;@mxMr{%G&QFB{J@?>#_YoE&D>1}tdM6Kn?1quCb<Lf^*LGjTh
zz5y|aK}$tCoO0o$f2y{M^u6b|B_DL1$8sYJl^;r!jyclFfq>(o6b)W2Wl+ai=?WFw
zohrjvGZscF(5$iw`iZtqrCDvcJykSqZ*w}ubX#UpHN)?H5BKP?EGA1j*ro7KHy1Ff
z%qNt(g>HuSdEKjaI1P3;<+l}nhTwI?{m?Aa8!@Ofo4}Qdr|Z7vpUBXeuivhfF(I~`
zug1WcFuCfDeVvDKE1Se2NLnb@%)<_z4pz5c*K^H$Dad{|cXeW;QwjqRkfWNSomB3g
z7;X?(C3s`QC12q@G9s;5t1j5rFG$b#glBaqU3O^wFX+=Q;zQ}&?w?2ir9h0Lu!*ju
zZ048uRAt|d%{hj2^%B_V^>x-chs3)=2`MQJw%N?`29TTu!B*?;^D|XU?t@zWx~0Qd
z8x76}OUrgh$3Jp|X8KP<;^_<L0eX~G>-mOfoDBDLcZ3vRXUdrd)dy3V1Q$m?hR)9q
zX2sj|b|v_;?_`KU>|dMhH^ezo6POGHkj7qqFMAK~nlK`S2C+Y%<9v*rF-tAX-_aGZ
zVe$Uz2z5DHUJ5+NrsBvCUw2v`liZ!F9)~RdGmcjUf}>6u0*vhv&~kS#4j?1ViZr<N
zN>5~xZ(A6=Py-bM-`u8k<nJ`T`x8xIg<HA-GeiX@l79bdMhJ`I1ZNJVnHk^tUAjqM
zrZzYUF8>0J9vjPX0`ORcUNd+<e4f#45#}qMB2qFt_-gA_1Svo4$Tv~HzF#Yw^My3*
zMVcW1yVoDoGUa#Vbv;HHG3dfPwAW5vdSH3ucDh+`H(l}IV!s%D9x1r+DpK%Dg-+9~
z?D!ymn^vj#o5{k#bOY8NmxD;(gq4z4aSCVIL#rT<=e5$_!BpKdz|S5Z9WI;BtTYtu
zkgBF|IcfyV+uA#KB$TaGR-ps~=;MK(B+8!MdM2AJN%}cx!Wr*X85p(2G{#4d>JJJe
zL}OeHCxmD@K@&(D*u>NWZa%j~Q~%o$h4b^$%1R~pX|@7Y)uW}8^#Q{t$ra{5ndug8
zQSq49s|(+Tlkp<C-$e7zW211#FzO58(Tnj<_w7(we`LAv%%zr$*1P)cBRnnP{vZ_{
zU@IO<6!uFo=Y{=xKgabXWwvZO*^q9d+_<F2-ABb&!o9z~)xK3JZU>;lSgz8vodh_Q
zFp=5#Q{g%`%?A>rF@u935%WKhLnX)KKoNo<Yg%-<N%~Zb_=N1L2J=|jw%}PX#~XAI
zhcG{Snd_TB(S#3hB*R!VE;`UjkXu9}r{;ch%pqzlJR&yB45U3{vCHt(21fHvoAGP_
zJ+!lHVpeIF3Arj2$SvhAymmQR56iT`seToRzpy^v;2hZaO%#<%{+mdkr>C&VAjo1y
z1cDI--qPHwZ^=)%hrhWHZfW*J-+{@NNYm2Jk||W#h7Iu_`#92=NHiQD^jw)%wd?zd
z>G-$m{UphcbUAlsR`^Xtyx2G&eJ03Os67uRnvnmI2j}y;^-oG=u)UXIvkh#{Um&aw
z`BNpD0$?KH!Du$9OuYyrtYYHGyRx-h>D0i-9EA;c7%`htbf0RGe{wSMU|mXJ|4e+P
z#Sq!q-c5vmRn+PJ-2?Kj@sTF4x~6I8$p%BI%Mu)~6t7664S#ctfAz08mYM_0j{B`=
z-H~iNLl`+A>lqSy(?K&z?++BJ#VLXHMpOV5HGf#j3Zg}i+~x7Olvw*N9fhNqE2CO*
zesi%?c&yXng`vKvN(OUlUhj5BMB>hoDSME%-R3)Xad~9uadR#cvIREIyN@{AiQ=#|
z=+VT@Ns?X)d!|^W-|$Cr>mK=w);{$Obk)WvTn=-VI?|4@kH>XrVxCq$jzW%f{`CvA
z_4oe{cFW9;aC)SVZm&-A5$(s@ydR8y8D!yKlCn7QEdLA(@(aRT>QAgFEi03MnZ#<I
z-QI&v%!KnpgSC6YYIX$ND$rx;3Xv2#k;qH-Q8@NC9Yc1Qb8hw@p002%34ZWCn_=q_
zm#tPW&(Ihz(0C}KunGb2I+65)=P~Zo%uGSsLx*^Ho}A#zU;^>uReYQ6apZi3??N3&
zXUc4o#oFCIf_@`e?jT2uru{{az77P};o(e0f?Y_3RKP=&QxAMJWP6-Vi8AexQ-+)p
zG0RwX;4W}5G2yTW{lUhmkF|UIapl6$Q4qXH!s{hcVO_ArXjj-{vV#>op)@4)BZ*D<
z5gL2>w?rlGl{4;hSozKZWT9;da=HII=rL5n$Zl6;_9Z$J(ae~=j*P0sv!S}yWeZ0v
zj4X@KE{dzlLOA`*x;C#mib4>M<80G9<xFapIr+XXp21vB7wJ4K#t?*z2l5{1#WNeL
zBrH%W%)~V9Pk8EGd?Soe6!Ht;mvYrYqGuFTRx_1qnLjTh$oZRTZ>|p}r<x6sx5w|E
z!b57k9;(9*>c+~P&fcRFG2=RM+T}0*_Er}bSU!`wzBy>Iy}QQP6O=%!0vC8oZ(Udx
zh*=W+vGJR5@|Eh6OT?_K-_XkMps}Ee-u_%&t-}%A`e2&zZ_irigEXS45ApP=<{O2o
zB}%uhiINB0u5NBZ1{-}owccB6eepBj%AF(WwKeumviKf)?<{q?`qjta19f;)$MTym
z{1hPo7PDRHlA)XMb~rN;tTY?(NO^VVc6*6(ym~~8nFnB%&9O}QU;>8qm6Y`9I`4;O
zR99*+TaO!jn~a^6?2vb<@o0J*?@Og#$xVvnchxoy4(PGh>TjUX^(C53pw}e?zWOiT
zfbehG!!si*tG%o*Zxpf*%fO9Pez*4+r#r^fDG_A9#fQ-H4-7RpDA3T*D#wJbC61fO
z#~i8qWd>9D@nO$PEX$34QIJEnPCpXNjUug};lEqF8J68cKE<Zac;DFA_!tuGX>ZC`
z$U)6hOITwG^h%IF5&AQPIRxnkh)#Q0AVCh4Gj)bl{E`XA927+*_>|m0%x<;tt$xLw
zpY=v9L%@rK`&KofyXN8Yh-#u(Gj}2G$*!P?+4+2}b=NRxV)0MJf>9R0^8Z@Tjv!1?
z-fsqA2+Vzcmkbe9ymy~PzThAVZ)^k?saO5(MrpqABtc&y<z*e3uGm<-pU@uHY1ZOQ
z;dWl=ZnoXMjWiRArk1q1T=b31qZQfy{2H9%FJY?wr_TLz_G5xv63o*e^T*7<^q}8G
zlC83dPF|rAWV#K^Y);Ih05Ig*%%FtNbL5+{1-UCTw$5XEIvs%ji@mXhUdLzUnBi@j
z63%ByRpv_z{*uvcGhO(<kx+EAa{J;-G9)Y3CDM8oNYhzeQH|IeslQ}F>aDo5`w}Ec
zW31E(1b$Z%vJ4l<a6d+5)A_y6D@st0u>Q&=l){*9iZ3R2=RncM+j<}LZSlabATU`9
zmzT8T8K9L@09gO`4G39w-{12+q1SE>6^|e@>@3aXJ2vbJ&Hq^{8{CCKeBI-He#>bz
z+@ujp%XfS?3)nb@q$;y7<;J~wBcHBj-~(qLL%&7u=tf5KgNRB?{)Y&)bU@YC%y$;^
z@gG4!Smc@D109JJ4yIQ!iJyAM3syQusMP8_2&=6i;Tn$H`BFCv7bU#waRlmKL+ehE
zDe^x~TMNMO$g06>P_Ntnfr7bz@Zuonm!&h|lXzSJ8}N&2RblhG+HHsho$uX}RE!}l
zxH~wru6o~H<NbU;lVuV{^qkA6l#)Lq44@gp_*e5S&NZ1m09a-O-@~zG@H?lm%+1f+
z?j4JEMv&{qN~UsExg{^}&o^+`E>ph>Cl9j5)E+MbY+{C2N7t11B8<6GiFF<ev$e<s
z#NMcJ@=w4$O4SGcI_hGfV&ji;{6>>di<3>~6Xgm*K*3q;&yOj9{M?he+G_D%J32fS
zm^~*wT>dGw+80mI<j!?P?Cw2tN=S=7l)iI+s2Jf*_<dLEo6(o$Kw(I;XUDwz1-jqG
zF=OVxC%5O_&nc|uYRx9HQG#*bgtC|p7Mw5}m$RpUqxyxcAhotEzp*x@re3(VwnODs
z^qS#9Gr`RlalBbjQ4}Eo6Cdw~d*06qQ3s{}3NGdH%xoLI9~O`Ub8~3`c2{W9mn)Mz
z(Ou@_Wc12=PorvV$@~5ac|fze9bj-vmj|Klgme5hhZIC?A5l^6R*DAZ;z2L~rGe9&
zeHU5EZ}Fg(Yp~tdq*-p6Sg?~?QjO@~X+Icm!+o^;lVyXx91J~yHUj*=9}u2E34Mu=
zzg^Vjfixc?wb*^2qeU5bh(3kK3D@3az`wBYZ5+Mke%4~UXvp|-#of*MQ51PfDxGG{
z3-@B4DL}o}2NEAcb%mp8yza$FJC8|fd-`14p)>qd{kEPF!$<(kq7$-W?|^Pm`YN~d
zW9j|4bU*YbQ2KH4Ge1N@%5QTLp=I)M)z{s@(=1RlngZc?eITU5bV#thnek3nPvF5T
zPct{1GE05*)n|M<qJQAJm|^-+)4@k?9Lf4oLD<WNJqG}6TxZ)<&c#}_9!n_^YIkoG
za|G7=6XnZ|P?TfQ*H!biBsF;^mG`iSIAGkJ_T;c&>K*q!g%MBW-&#63wbfNl$3@lP
z4VqJaIbQ2Ct@bqPh8iji=H`hi^q#PfhyYSTA+|bH6KZ6<!>$E@+2bA;?S_XIvmgGQ
zQ9)!Mi;B>tX=KwBGVL^2!6rRugI&R~g~JVCet=%!&sJM1psf<6$5VQ`HF{kA?w&yc
zOy8%uw_lz@nhj}smP-I8-PH<vXM3>V=~xd??-lbG^XE}n;`*_#Fp5I=<+FwLg8qag
zsIeD)yr^q`?Y@BWd#!M!+>Y%^R{E+r<Zw_)V}NKJKX7?GPp#*o0GQR~4;^jzQ2_2@
zQ;DH}eWN`SnG3Eo8z;)D#OAzziBc9rEs3M1p<w`+;=qW`o_Oj^XX<J%fJD?fY%%A2
zNXD#F0T9_!z@%t3c#&#1II0>`TMivYWew*WBIAX3x9o_<HNHtw><%ZTl5(Dwg`UN4
z;xlO#E6`ENVT2?QINa8-+W>NZmnYZkkqwFp_%vCp^F>fNu$^9<e(e!Cz0<8n0B8>L
zB?+G$Zk_#xg5Pdnzc<L~U>Xesa)wNsKa6dS+BiF_hAa^gL=s+>t66U;d}nl9eewuU
zrQ+_@4prJ8(0`k(ND;9cl_o8vzs?Vssi|T30-*_2osBw+;BlzSX;Lo})YSBW&Q3Ny
zh8ur1HuNm~Rf2Bg+vI-nO=f4J)@C35-8(OITx2h?POh1n2N(&7IK}RWJEe{TNJ~lk
z+|<-m-t=V~uj@{=)lAWK3!_!<>SEe*Zc0!PKw26D9JYR6Hw^DMTIF~xcYYa{_Fjpp
z`8Si_wv$R@@G}?@a?@ozpno%v40_z$m6Dn&UuCXrzln1LEuf!ag!d6XbwGpKDB%p?
za+~exXKxfV+eSx6-*Mj^uf6S0-3TRM&_QVMCS)-YskWT<GH2AKB?nxU2J21T)O9m)
z>}&ZKxb#{*kT#DuZ}*~0dAlTq4tJ;O`XbldAvUej?2j8uqQCdGx3&5DU76U)PGkHT
z$X5Vf`{69xLg}~!<#3@YgTujsHb!J`k*WRr#}W<ci>jT8k7$F<sAt8g_MBlJSI2?Z
zjvl&;t!=`!wkNNs(PaT}Sot&DbN_~(JnCY8ZZ48hl9|<f?8)8rSxAqz{w!3Pi<giO
zrLz}Yh$vKFzpfrIqX}#l_$p>hE5>>WU*K|H{c}$s-6w9FMFB+%vRUXl#o$N!E2^ae
zg=Vms5UVY0ry4%$OGE8g?A(2#X{7>3PjAZTx~-JLxA!S=MmRT#%_2$pVG4TFT?QgY
z26^SvzjpWu!7?x~^r!QapR5o5yul)utCV;O@ca*SILYrmD){v{gTW$UIHG$yv&W>O
zeJ7Z!H?uY1!^1<7-`SGr#@})7c&6jrzSB!~UD2*U_`8fh#V{DdanV$UOE$q)@mCfT
zs*yi#<J@>252810{1HFW5_VIxu?4wy?Sd%iJlyqE&pE0;=P>C{Z~<oHLmF3r+Kkf%
zz4maXFmh)Yv1%|AG)HEB+>cmJJP2J3aogCsH(O)7KgY@AcKT!}jaL|e#^|w|e&^Ad
z;6CTEJo%%$8|Myz>ZAJ@8rf3U6O*wYnyPfy1>`YYp*3f)dXTH~5}Flew=@srujFLi
zt4UpV(Vi=~L{?3XJNU)AE#X9L=8v|jD;}-%ADd_Qx{QGfnXhZK<RBA$R|}ITYnsvb
zHzPHGyAM&ySD=(hVtp-_0b^-prObH`6~6o2_*f9rc#*~nwGrrvA#CCqzr&_nky;tt
zXpW2`T2i3xvlruQzzPPcl^N)dd=rf-fb9N>ggq8Zh_FG5|E^MCB=zkG!lE)AOmVqB
zeN(K{)Z2E;Sj#pzILP63%Mr(*ONOMn@)A1aj(z=zn6JAUp*AK9P~%^2q<asS=<-|6
z*W;b64+x)ak5iS2O*w0~?@!In3X|m-0h7E--1_#fiWCGy4q<)-CUrb$q#KIpv^NuA
zIbAM@-x}*{jJZy{<3CV+Ddj??Ua~mf;H+;iuM+?r^3SAhCA`i==rf1Xolp2K_qTPZ
zg3siB0ccF5CqMww=;}wt>y~H3fYZ_>oyZL97<HvjJ>2Z+4&WU!fDU3F2bY^2g%B{@
z(&aKJ>>RXtv#ia<Gw8l>-k<ZuTluS&<+f7bG@Z|n=C90-KZ40XQb3tuH<FrWrci**
z_30K&MDnul+0JCg$jH+$VvtJN@k&>iMesnVgH;lz{o_`!qINRJ?Xfq*_p+&6c#1jF
zM>U~-e(>4Ft-qIlhKy^%mgqF07Cn!EHr|W5fKEP)8u36CLQrHQ1fmmuYWBP-*ai@7
z5MYh;k5_-0%~t>D4Z)|=kkF<b+lf-2)^2e62@K>>qS;(Xwh7h{Kha@xJ6T%JRHBH6
z;45Fe0xvZBrV+E4NCTSDBTLAaaU|T14rLG+arMye7(3U$Lzf$&_E1<nrgv!c+x@ro
z8eNXiU5{3}#rxvu#w|?=-^Gi{0Jr_}LNqup;ndBFd9+2pBk-DjFr7aF6_2*{_iu%O
zhUs$SkR>%U%XjMK2%ProO6kB#*fH|6Mha(qu&}V`Pva#9gvu`{NIZj03?P93UOGLA
zVw8(k`JoZe8~e6!C6-qC(=F1iR63uh0ywj$gBE2{G4}&0kNbJQ9BYwxVIZ14R0f$s
z2o=DaLJREdNpQFX)TXlk{4Icg+6^*afK6qY4yCCzdvNx~(G}ve$1?JCp{4M7Tw*gv
z(r=yb&A!%YtPLJu*mx<Eq(t|*-t*=hTRdi`*ALz`!)CD+7M+}rgt;R5mjj~L%^nD_
zD#E<4QHx^BBaeP))_8dAFEkZ?h=S+>3xxev<y|N-2vaheS`fe;cc~*SHcLXGge+pQ
z@j(i}AVh;ONtJl@uQV}=N_e1>9oV)A&+TZI=iXL-{bTYpU=#7~7Te(2Y-r)X`M&(}
ziy_lZOG}G_<+iD*DT?3Qt3QzidAv{+Yj;eK{O10yA!&;a{t|FY!hpDqvH5Akq6HSV
zW=qBZ_Jv8J$b*K4#(VboY)92-AZfh2Y%E`~aQck7qZv@tsZxD-qgO5cm%269q^p_o
z@$}jrDHZk`AE;&1NK(0-Rj^>*e|_x>yl(-s{*mU{$n$+@SMr~17C#}#Q~8lZ)Q|N+
zb|{q_83n=uI67flz~!)w&O{!e5i?sf$Rx9WCSs%gkjjNfOq{Hn7nU0sBoO+XX-u#2
zUH!YIWa{GP18QSncxzPLA5u8Xb=%jorQ$J#gD?a7XX1mTj~6BQniV6+c;(6rL;=^O
z!z)*yQAPVoIw2!C;2b>etXV4yHOEJfPVne}(jCe<>j<>3Bg8E<x*Brq?(bs~hO%M)
zqJpo+DSS|k?w`89yMRzuG>C^0eK*nS+qKO3{DuO47g%To3kI!P8`xjbFHc6Ey6Pzw
zYjKjhUJs^w-+#G2+vyk_#5_Mg-<)r#Nl(0TQ%Ct#>3pD5q}_nu;C!Gm`4uEuZ8aM>
zeEnm45<2nYbA@6f0J+_N<Nk;J^=knBhy3A*1&^Nh-VB>?y7AKJ#!$K;K)eDQjr-!3
z0K*@6mp&hPc?&F#3R*_5(@f<@?`qUIkRU9f8b4tF0bq-X683D4)oO;-sP`*Md9$Ct
zb2KEmbNGTGjEt8E$3&crtOwo8m13+0tXo@O|AN)4@V?3yx+;I2s#Vl>U_+;z=8xtn
zJTdweEsF82S17BTBmdod+N0yMv24j_b9MH}9f24fcQ+SIp4Sd+Q`gsL7EKbREst5$
zHb3OT$Tv6^+9qoDjQmp`PrqfL*Sfp82`x_}gVZ%NmH};bIbKBpoNuuZ<F7h}h<7;_
zDMpJrUfkjlWIt~%4#qjlKNpWZjc!q1v9N>Q^5Fq2@4wLGrwXmk&xs*7wkuuAXd`4?
z4iZ24b(`GS%qNQS@o8c)Yrqg4jKv5NPK5!VxHhFPZIBl4hxB6(*JEnCjlnPg@2mhe
z$!z4?PubL=_NT%ECR2T%ssY~Ksq20BooMd{w9!Dr>wd0OvBzT6<D*q)*VfyMs$8HX
zzr)^Bk80QKzZ*r(bh@^OjWR$L!VOqY^`FBL0POvIr{Y-?aRWWWZ?Q$)_eKwTW!7hP
z&qJZ!OB%t5y|Qy~D1j-fh1c!s_&ch_i;(zUnp#N{`Myu}069R$zsx<{?&p>OumRQP
zmTMv4bJ#4x0Qdn3CI2h4aW94YF#v(^?8G0dY@4WEsYJXijw4$+ezn;BnxzK3oLCx<
zYbaL!-ov%#nc`FEynMe7K$K0qL#73xOOH@@AuYNCN%YDExH%tEQBlI$aa|F_LZ8FB
zo~$gk`2`Z5N9A7ZLdQWx9!a7|t&}AdRA(M)qvs8NliS%R^vxjvg926`b$qAHZWK3C
z{Q~c;iU<IGn9Rpsk6Kb|ibCmdKH_s>eg%k^Q`(;y(gPe+PZ}>b0JHrpCW~zu`Eeeo
zsrT=IStxZmGT5E2puv$U-GQz;q6M%<j+^NeJ#;D^&*ko#jf8~6W~m*K%|{dPgmO(}
zZzDM1x+BQD!7X=z6{bT4h+wJ@ebWli8)pa>(LL>3I(EsSRGHjYxlkov{i-};f2}XR
zt<8g9B#JA`<!IT)&Q8JaTDQe3Yz%*w2kMHCkHx+p=dr2-p}ZR|BCtH400X%+k|nNI
z@9^e$t&bpB)o}c&dh6^gF`LDNNZdPhM%Uxj&vSUyP-X0`p^qD&Lakpl{3*<Tz1ZS|
zKFGz*O(mPgW8OA;vN2@O2=S&}zkiGU@;&Z5-DVPEc1xe4idLxT(?2RU(LEJw(JbMi
z*AHc`$(_Jzw)!V9Ib}w_aHgiG17@n!%C1mSIPH;umk$UDL8BIV_ScW-v@80soOURt
zPeMhWpRoCYA<g-6Zvs&9g!X2tY?nJxFxP{Us;^YEO7#TMNx77Mjy{}j#cXdrW{X?A
zy<J>g6T#ik@f+Bf$MTk#MmkH}KaXoId7{<egaMdEvD?d|GP98vgGKjv(rJYV+mj^`
ztG{9lS9_wiVMhRnGgA3cu}G^fS_n!;aG$U|79|j}F+kTjN5cWQOKhS@oyl^FQMgLC
z*`sZ$%#g``T{@JUp9~m23&{!+qL=6+#O;ZFOQ6%_768b%Fram<JslnJ@>$0}H+0L^
z>Y&p;1dIHxNRZ%{P{>zWL=T`@=9#|%)3pS+lmca(9Uj1ba5(Ri(Caq$w|P9PQ(gZZ
zs$OBTbak@f@^Ifo)F1EltG!M<oP;T!zMLF7gCYP-J5{TU0rW2n-80x6uS$NU(^R3e
zdkbFx05{>a-dN?$n2)15netg;4-EZusQx|zQ7=So&bF1=62Mxi{r&MZ0s!?7tEhwG
zrhw`^15`Y3q}%|iFB1Pzod6=Vw7+ki#|8LrT40?$OG`@`R{;b^`4SsIr=6dl!jtoP
z94Z<E%Dc3_o@26e!`TA`RAEv5OJRif{TW{uLPtMIeGeclSsGR5k!tH*zxw*nYOLo!
z<>uzvve&%9Hh?4N_Ywh^%2NOm^TS^}hGMrybA@*`iatX#4qw3C@-vc>LEV^<(AVqk
zdTFN0BJgsp3nNDKJ)jY0DL<`H{S$orCkWZhaRCGt$!Tu_Fx1Kajf%I7ife1SMuR&f
z1lI%!4grER9^8Vv26t<qaT*Bj!JXg^!QFxdcX!vu9U3m@yU#hv`-d^89;5f(wW?;#
zIoH~)P(QSw-wY}<$Rz$hh4Td_a0%s$?#$o%odT8@bR_gb&fZs_dC{N*P#Cxv2evUr
zTbY^dHY4BsbJs7R(38Y!^wIm#)${%mWo2bW@;e2BUYjSO;p|tj|8oV&7ifm*|Csq#
z!&AW@ueL|MG#?QjRlMsC_Pyw@A1_5N5wQMU3}~`Gu{WNRDw;Zve1RVwB=N$*yK*F=
z-mU+5!IprRiJc1Wn-}?O_3Uw9R<z;#V&h*r_qxGx7jq<oKP(7~eB-5MEDa`T6|)PI
zOD05;JL%suQ1y2tp0Ro@-qgoUUl?nTBs#(^ax$}Qhu=hu-gwDgN5r1F_1@is6#BgM
zWduiWc?;P-lO{HK*OGQNiYu7-S&Dq^^CD+&OK4*{>o?)7$$go+of8w-npMW1UnJOc
z9s1hCP4mZc?bT~2lk>sscRQ*36Xm+RmRv5om5!IG`%<=Ioqh-gMn*a3cQj52D=&)t
z<q<bt_<e)x8O1-ZeK6DB^xO(a<v-emdxIlRiu6Wq8}#HxChM>4`X@&dZTJbvXD=@8
zEH<-5g8t#maV+|JCP*qP?M+wiusnDQx<Ealz9$)7K@T~hZRjAEZ9!WZnD|Cybh7=b
z$@4^}ZAn?`zEzcXf&8v@Ycx|IdQ<a4ktAcNLcD|jykOqJ2>(J<@an&(|6Ad$KK>~h
z4PRkQ<uhTU>W^uY=l~NR!eJ`pGru!>GyYZ>j9OQ)iN0Lz<g~~WxXs4u6^-u%5T4##
zF_~QP&Ceh?!M>!LTmt~uX)fK`r~7H7^r3LG*Jk?g5=yF){39|*l-JKO;>Q5I$(uQc
z)xPiwlNM+<gx^p!Di8<7uxQoytkS10;&7(k3mlf^-R3DX0nV(tF=}}2`tZbV*c0wc
zlc8^a{j?gK`H{8B>ChJhSu9ZyJP>_CVfBSW!u&dL^49+776qq2nXpc+|FpiDrj#kx
z;zr^)?3x^+eOy4b2_=8lro2%+hD6z|-sr9-I$qg8wa+hK3PrxR?~3nkX0YwdwE`-h
zP3fJcabT~o9dS5Tf7Nv91?BzGwB;()2V2bW!?TFthtYcnj7^?P8|Y#KxjXTt5e!7e
zI}ZFzF8=(CRU|Rdx(lXsixF+Oy2OyvZw;deK+AGzhY1W~5|ln;`D4a3{A0#6+(a_{
z&y4X;?Rb&f<9Q;yb9uo5Bh-^Q3LnDkt<YFxeShL{J3eFVkhujjjIJ1sJO<Ej$!pYi
zl2p*x-8fmU;&D~ze$GV9Q><4Z>kjRFExqGf(N!#;+3wlXONr0w*7ZhX(Kz7ZSI+??
z<7bmV#UGnbSmdRLv$6(w{)fq?j2}(TsbrHsB2E(Th)XTvm@X-UB4B;d((Ju3f;jNU
zTV24H`-?26zxap>JK`DC6^eubPK8>?k<Ky}>+#<HLa3KiWVu2g5vVRFh9V?bT^^*Y
zms)Ve!tgM5k{SPqO<DPu3+KDY&r$r3dGw>orhLTny2ttz&&m-^VPG9=Lqja?`xD#s
z;`nXx7s1e`h!mrVseJ=y&FZGem$8nD&m0cmC(f3-^D~|10`=CwgEHG9O^p&dd$!Tf
zcZRLBB8;s5I%fKh?vka&`ybAVf_w=l3k>f473ZdNsIdDY2<HKFU{nHlzSg#m{1a2z
zj7whk>jRtLyW(T8>J@ro77y4O6ZNb|j%Jwm@CiRUSWrn?J&`!>N4xT55~CT^%k@hm
z3jW(q_5F$e0$p>d7oq?AADPB}Z|wc2B7LL>_XS5&Z~tZwZ{o+ZL9QjwW0=%9ls7DR
zMsW|g_jhUP8RUu%2X7elt9`@l5Mxgui}k<L*?D{Gi~Nhls*o-i@p3vU(5wL<!fy%3
zJ$VeX(Vr|P^edt{{`OT~E>2*0q4eM3(^JSI*H+D--)9ph#rG1HvHQP9WaQ6&b)nfD
z0ZMMn>;|sP$Y|MZS?aE7zy12mX#Ti%#aqhfMy@+x@_E9w()(KJz24K>6~2b&7A7-4
zNMdFRWxPlhFK&?R_T>Dq{)D#ds$PEjs4$+TEbB`mS!2#>(?y_{mr;%pU3UZ_Y1J(x
zsZ8Bo`9j+64VJ2-c9w@2Z);MfoyFqVGL{2mhV*3wca1uj-s1do9PX)Y-)|q&*^>-;
zw>6~XCQC!kD(-i3zr4(`;4|?cC^OKgbwV=bV@BqtU~DC{Ta^#I>3hu)`aERuK@|ke
zy0BkyN9Q7M6=fUFajzsHSjtA-=njRkv$K=E+k97--T=!rdP8fTNS3oWEp%WU@(B4h
zTfjn4ZOC=qpfjt(ccxv`vZvSni7o%Saj&V*gC!1>78ujWW*~lp7}#`xo`MB1O}pnT
z_ZbMNCNv|g^B7+CYuI-*>>Zj9<ookVjc@R~j_G`gU%H0G|7)&w+MlMwL{wpC)47Nj
z0V0U<>K1UrvVttp;}imQH;iE<7T8E`=UrSUtd6T+T)l6xjpxUh?Zt?@?OIRa8L_*6
z4VfZ**YxQ6ard~%eSAN#DkJFd;1f@CoYlW%Wbbu{C17FqfuO)jepPtj6J5%9A5DJU
zyF)L?=UzK@EK2t&+N3#J^5R{oZ*<LeUw=iG^s3&@F{d)&PEZAhzO?fF@7+j!-<Mi9
z82=SOu+Wmd`fWXp-R$B`gARDjj)APrzN_Xc=n<XdL7O{x2x?|$*po5&IgV|)3u$Co
z_+eF@Qtv1!X*vfcMV;P%J~i^|c;wsS@%D=Os1U2`EX^Lj+g2IUCGZnsyvJy7-{{k`
zozw|{xYGTZ@DsiEtLk*3jia)lJs#(Da-U;%&cdcPPpB>T^=MEOx6?Z3!0u=$UOXzX
zKys<40Q$$YA2#T=YA_556&+#tEa87ML0P?b3-(OOo(%Is|1thAKO7%8_f)tE+43aU
z?1;52;fh^qIwv+b>82bA?U&u3Dh)o=Tx2aKd7RcQfRC=pbQ2rDuHB<)<PKMPA3VW`
zv1z{x<n(DezJ!xq_JBJP^b+f%F8I!1V<>5!bhcQLb+h280XkxSdfI&(YAP2Ac(60L
zDnfESn?VCFdxrFivE5rNn%Vp`>)w&E*YZ-+l|eMG*8Cvs$IW^wmRiYgx`?ZKlAAtT
z1G7H+JDn3HrY8y^`TL6Nv10AA3&bsa#hoz7wnbs~3(jue0*p2bpLjuD0juqUs*dYk
zpy)yhEMd)OpcL^4-&DB=VHcO~uXSyfv#SS}E!kLjHiBi#PLy0(g1)bGUS;f?EF7^M
zi6H2p7GV%BGePY(>b0&{`82UeK+oAEfGKCb#mMzxoqdO-A&;+K-_wt==xryxl#VLj
z$3wGYJ%XBbSX5)#z#H;*#JN?Zid6=|o109BjS;Ikiz}L0#7Q%Yr*BU!!K)m+i5_69
zn~#8vKFp`^%L}X}m1dJyk7g5dHtJ_eXg1Z29)!k{WaQhm$#*i5D*=qM+UzIKW-Euz
zWUWxt3#`Tsa+{qKD@eUEM<Q+&F@N_Qa6N>cr&y0{OD?hdG^7|~WANmQw>F!0dvtAp
zd+d53r!u&8Py}Lwpe;H%aGMy@6A`GH8kwydB=<yWy3!8oxI6Tj%?@jdX&;P!;iesF
zxVJ%SuEQ+&qOkoC79(>`T83#8D05O!ydh7HpU4_TQ^o8e;2n&NwkFc|vM?^?G+OZ`
zSXRaKQfnj=_WIE8<zi>o#u;2n`Ry@9@NLxDH(WBB0^g(6iy(gIvRYT4$d`n*l0lmy
z<<n2Kc|t9Z(qAX9kf+TWhMk=ho29W&s>8H>_Ut%ga4|@HU(0i{kxsMs$pRzL0<ydb
zgbUf|`<zfiHzj)!sns^5HkxtPr%g4H@X|VK<1$>OOD^KWqKUqH`M9Xr^M!e?Im5=I
zcByi@=Q}z0=G>l@DIrhmlV-osFT1stq3K*A_xhZfOt4LLD?2~h1i^QqUnE%8mmqz=
z)BEWQHv}RFJD@KhZtt4^wklsfZ<*L?8@M$=>Qt%;ixJ2qd{`_tE|E%|IZxkc8l!}9
zyZf_FbjC^Gx%uRN3%BWuew14)QHf(!zD$VlJlk5yb)ZIP>HeE956{R_GA`-Adb{wy
zvsnmhSh7R6BVI{yk+muG&0yMn`?cWaW<60D)xz;wA?D%ywib!3?OL=y<{AdqNZ^kc
z0{nBB>#v_Vm6hmNUM5v9*OSoX&HTpF%M|{Worp8TE4RwHLsxkHXiR|ZH@(G0V_$GZ
zkHyXXBL>G)--=s@n6tIPV;F)$XsBc*MKE$`>&=Qs7{`o;(WkmS_TKJ7#hjNvm9vGc
zy(=Cktzp@7VsGAb`%#~yX`WOyGped|IIh^ud+d%3rQfYCd+iJ*pFB3fsrMTQ)O_}?
zVcHCA-1-D*zfCH32}-*-lv`25+$!CQtE1U<7i<D8J-<hcQIRH$lD$o*=y|3Na(l1t
z0sNs&2`AN&<hTV4`t}TnOl~g!A%>#<0^GW_haw|hb?<9_bVTY!4((JfWd{TXwnMs&
zfYE|^V*Z<3xi4~!MyhFwg61P{V9=wZQmqNiNn$k?MTwWzWyv;1<X1cSFrv~d@9^FM
zOt(Q5I@a8_C~2~xi%r^7zM7<tFw~>dS?k5SW^U;!=9Z*27eO3n8tXpYjj>jsZEk&?
z!uVe5=*oC8aVz;HS}q`?1FW)8qD5Bwdg3uk*OG*_5)%!aM{=1q#L;xev!Iz}ZM!V|
zg#(zun^&#WO!2|C$!Dj>@rOvG4!2Yl$77lyn_i2?&lCFM@OeM2&3ZaBfDBOdifuM?
z^_xN*aMakTe|S9b9LAp@-h#$ilEUnwqgZ39HBHU<a%4y(<a1_wqyo+}nHC=pn(43J
zFlx0T<zfVaGDSr94}4g9uR{g~Jb2b)Dp+9Q;Bty7wne{Qg@P5^Ri@6xG}%o$o~qRy
zM)dLdTisVJNSdACJbr5q^~%|UnwojcR3e|~7;XxY&&YYEI==u{X34k_I^~^4L(*5D
zQ`=HV?j;o4zA~$>SPThsM|V9xfNG;>CSaMo)5@}s4-@;b(C(kACL6h>x(QN@5VgHF
zRt|bv*M#M^8S2S?h_?&c2jx1tzYqoVI*W1BZa!;1Cq1#S_I!j^UuE@or)O^;#V|ef
zR0{QHB>G3UocblJ_$|hNmf9+A)~wp-&`7kuNU@+2KsEa_fJyr7kvClL)e3b78s7M_
z_911l%d-d8861<4Yb_Kdz3(YNyPL82gd#DunvF`T8=EJkH9tTp@|S!%8}%_lz?h*y
z1Mu8xcG;rSAM$I_tT_IndDUk5hB{1y2R*rxsO2@#4^QFVOAi6BNV%wZ7wSCfVA+oM
zQ_6_(**SheTdDO-rY>v$hW+fyPNtL3s0uOfV#3V9O?G;#>Gf+^VIgE>nD_7Cv2kIR
z+as~3qm4iQ8}b)2d~d(eje=wEC|q1W{(YxcH6#tn@Y~|~-6cTGN>D{kA<{=s6mu8V
zEiG=wO??8O?>?swDB;2?ET+h6WQlAsp*i>&$P+X}3?zZhx4{$P_D9l;>K6;}#5qbF
zP5`FETu|(DKb#<*j--*X(4w|02k3kc>>0W0!RcI$AldwOa6rKgp=Enk2a2v{>B5wv
zjLKcuS?-a9wrD6E6hwULJ*2&u@m;RS=R%z6FS*~*ID+8hjrN|i+v1Q4G&@?<R{M}H
z$4l6Cz6n(+Kq*4-nNOE4YOItUq+cle8IX2PDtXxMomF~!Gly|{3=^h8_cM^9n?D`0
zCV~mot9jj_xSAS_MN@1-qcvTrtqSjc<(WL)`sEpCz#xi{RJz#X?3c(H`Rrs%MNt5}
z9zk$qt%d&1&S8uI#vI+q_(7RDKKzOBocnS4LhD#tDKB75srdKR7qcN2o>PAAyJMZu
z?B^gnHrK4IKzY;0s0<Xa?ZFsqFKg%wms#NEc)5P_Gdw~yA<fS%vr52Pxf-?H$<itU
zk-pYo+<@X}CYHCJc6(W1i>4E==xqxjlisP}D9-{yXD=FsBB5ZHA0XZSXpVKNP?56O
zGSl-eWYSEbM&F-(4hNZcMLA#6_uFo{;-~tzkvc8qX$=p*`&TRFEROPMA43E9Xte~;
zFQg{h@Bxz#^@YVv1~yj|`jLv5kG2?2VSu`=R|4!V$P=+$>69&+6|K%!8?^CUG$|^K
zqHmsvU%$o)y{?ts2m%}`#ALSsU{}W2Vyhdy3CV}_z^`nI$`}A!%~AT<y^UUmb{qmv
zD{6aJ45mt#JXZDc#}@#4Pbh=C>@9H+%WV!K6MRjvmaKCsi<zmtzbAgEX;j%X!}d48
zx5ak|l_C2p`Ce6BUX$O^EFT#YjpP}%)R>C5EL<f%PuqFZBoi-hlv-%dWIyX`Z;n<Q
zP#7h&?6X@C+VUg!R_JrQM}oplBe0~HzohFZ7|W9+A9sJ%?Y&Rgdx-EFr+h}h<_V;4
z3U=fYrmQ23nABTOa{Ph16AdLF6mWdwr%}%9ss=&4^>}d55(0!y3%YP<>s#q}vlGi@
zrj~fy-91Nj`}_RYUeP$n!6JfmeAI+F|BLxXiLM<lH``ZD_YaA3_=j%Qm>HQp82~pD
zZ=6KOV(m#TJ$|G;9OatzJ(*9C)l+NYk;C=JZrBp4E;8WI4JQPWbUULOtM)OjPZY<M
z#0&yNN6w2^nCa>0WOBNV8DTE(hrzU8ls=|Z%A&Sk$F^YVPg42aHW4%7%qd@TM2R@=
zC4Qi3H=SqGnvDEr?>vB_1VCXJgs<GD3OcVndn3i~=F*wuyt_tzD=JpuNXpb!D^*F7
zB%yfVLAQ~a_iYZal8A>nr(qkkJ0zvSVbuR#FhxeC*UMb4@aU7cw=OkJEiNP_IFjSU
z`eSY3j+Isg8q<iUfE!mRy0(NHF59G|gIY6*>|te^(v4ftR5m&b{AHGLRpz8EO62`)
zDY&fVKAl7>`P!B?A}5|7d_}}Uw=qAa^xY^nQ`fUpaon0<Z7+hS=JZ!DuXr2uf%{2d
zs{#}~H0P!XY4~dKUali;qpXXC7cvq&a?9JbO3HcxFDxPw4>ui|3(RI~>wR;MG^(#W
zL06i%QtyUKU2+~M07ud)*rKeflTGUt>Bkk<>p<g%AL-dHJARSkAj;2sF+VqcEIWR6
zh4d^X_pnr`_6ZBIuC&-#r{=!W%dDSxu_i(KT*B%i{L@B7v_Cqr4CGX}Yp<P?GQAe}
z!?k6xYS>x2HxH%eq;x<WdnQY_lU-MDOP9fEjcNnwEdef?``7FGGQ|QF0k2EmS<<Pd
zrkF=o>jKh?4H1>P;||SQv~|Tjb`C*3oO$oZnBtKbqrl#<u0T0&h}J8y%j(L@vAuGP
zIL8cpWLJr<HO^^Opl+K$97rXp`jHC1yP3&dv1((grpqrDWB24^Xv~rAulq{=Ar3{8
z8MTCBZ+rWXY{tQN&)V5XlFa`<_W!@)U4z4Z9rT3c&s2~7ZzWG5`v(N9vpeche+%Kx
zlgX*spHCi@dqfPVrfZBz$A^uZY*qky#Pn7wycsENAGzO3X>~8V(C|Xk)zk`=-Tm!6
zEdGg3RO_DbXY%EiW*C^l;>eM0YG;cea)eCr6Gv(*VIPM~(a%cH@T~7qtG_3-L?7Ad
z47+*g5(-Ivml~~F%)o5r7d|K*V_L%al6I$@!6NcLa;3FJm~EBrsvE_?8KULqR<}rY
z#4xAeIu9eSXuggIL^FQRs+^3?dDf5K<oz%c2)X4hzDt?eL`ZA*NT8w1Bc^9Pqga4+
zG~?ib*dy6?bg^4xzTsQ&B)HIokf~n*J|3}@6Wu3Um)XZsMWQp&9tFg(voVH29PG4=
zKh~Jmto^zp<@RM34C7c;`L$YOwJ9wWtaZ6g>$PI@TDi?<)Ep(Ko!4*&rN#9OwnB=S
z|3isV_ghC=qKx!f1fYpx<gFqam%*@ijCgjieVg||sGM34ASYeX#ZI<uj4>cbQ%icI
z$w5bcsKCEwG3qWR@(FcfM!$gKmSfqg%vScKQ$B+}=;Pp^x!TeoZeSQLBSKeia~PwN
z*kif*RgOCr1;3XvdzysxaapaP`&l4xXP0{Uxg294+g*{RmAO@|fS$>^d-_SE4Q+K;
zCkI3pzQ3RV(OMHAnV2su*B~oAS`2nydEdb%saO7;dWmG>@Sr{`9jiss!sI&<Q!$Pz
zCUo*L9N$_Xyfk^Dgb#M<tmhC$+P!ebr(fe*+3ae5Er4iPM^e8X74nmN$I83US1oGQ
zP3JKg=Pm3XuA%h)^#fS(Lp)(Wcfh~M=?hO8&EsrIU4pF~gexUrLM)c*=>M>{6A+Ji
zm&=qG*VVMLxVLM4JtLhG<a|x~yK)qC9w=uUaCHi3EY4bf*T*CwdsCc(1d&_`AYCq_
zdvo+DG$pb>EAt!0^n9c~IQUm?{b=t!gxy8A*RCwA$WhPdt0pezOr4noc%(kWM%4fZ
zTs6U`VWiJK)A#H-V2f_2;)6%Z%%GW){;h=JF6~w50ohL~!#p2S)uRqGz`GgCT~c^0
z>_w99cWY;kkK$-}DGs^ZgHxg=iT&fHgeS;reDI|Yqoo>%bo4WwcFr5Mbq;QsKhIiX
z=948!H1&)JIQ7x>ol`)F{`3X)Gx&orHsmT1Tkhq>q-d1pyPtIzyvFlo<v8q2$DJma
z(_9}2#}+J39+w_k;zcJ0xQou{CXRMV@fdBoP97T#n4TMmXd1S=JNPYT5l<SvZt5Z|
znVH5A@kmPfX!`AY#C(&Mmx`L#lAX`Nvzn6n1lI5nR3#y)A9sppWI3y@mrDK6+gK?A
z>VAy4HS2Sr25*WqU8muIj<T8P_ZexbyhKqf=QUC!!YJC^X%qmF%WQ|BSIH~EZ#RHS
z3nz=yY85hEunNZ|C`t8a@9twvJ#@l39yM9K&$%)WQ<)arMWycTeOVyy*7iJCP&(K?
z!%516uYYjfHw7Xcoq~l!TsfL0H$z7Zo9Yb~PZnpDnxNIQq(Xz_yM?5vM|^hLIov?*
zeQ!$QvB-eUIFi}4D+@J3s1`x>dK&!E&~v2`>7}h!e_^r*9ujb%;MaV~8=wwjV)stE
z$c-1fb@Z|nKI!Nk)-9m=j5U>Aiu8(gKNBC7`_3Lki%;rI6SamIK%N%6hC~!*nil2r
zcDL8KZiW?%XLKM^j`kNsiu?nSLAgvf|3PFovS>W|xD!!)OWN;B4=IawM^S!>;OegF
z8$D$D?^I~S63328i~5{tUD56W18xH&E2wdzpV-C9Qi)`$8L4RaB*dzXxTpaza8d>B
z$YN#Hm6MlbfS-U$I;%rsB3|!Sor(2PNSnP>X2O;kttJH^+fk8!rsT%HN_F9~<@I-F
z8%rQ}W}avrU0gN)sSTTc*AseFD&8Qb9O@t&BjYq%!ov9m+|8k^;&bUZnC6x)qs=wK
zfI+S(KPwJP$WgdjBok0dWpf=;7Ix6Q1B~1umjUuKeyI*|x&c#2G+N*^w!x4x@&ZLH
zCU$=W1iw*9q&6sy73Y_rk?7;Ooj2;T&F0@h_Nb$bYne?ZSx%dBWuj&wo38nh!amAD
z3ZYm_C(7?d`(Px<QVOJclnII5orL|ex$06x0uQ0%f~BD@zn@9AU8tMW27KrO>iu5#
z9~=tcc-&Q}@l?>@u8{;kF+K_nHLDCo0I8rj>6LurC|XJT^;-l9UaBxK??-VSR4yC0
z#b^@{HAU}`t@o<CCWKlDYA!bTA1AAYotBHO-pFvXszHqzx_9aWa?Uk*b{VN_O^R^!
zkk`ylYN^g9pk<r_5^2>>4@0#N@spXPPn7u9Dkw^+YP7})EMWP$H|dcijaQOyGo4qv
z_PH43F_-OdC8jcS;7hUmM84Ctp4p*EL{^jXoM1)5+KJykc+s%*(Z+x*RLf{NN%*1+
zs%PnUxn$t39<S_Twkjz?QaO|AOj*438VDl>@P%3@1b$|dO1Y2)7H-uwR+%a_!n#&w
zDP=85_nM(I>4=(*FE7^U(o3rh-fSfb?oW>{+bcl43F?hYn)JcInHL%X%lO53WX{3X
zl23hSn8bBltol=&af9u?XPn);_KCv^A#6cF8iX}2$zzPKcB`TF12Y240J^nm1jS=(
z`~}o(!MF2KQ+)r%gjWSh(O6B_*E%EEPX3PW0~bCLCDM~<@hG;07+-bCXjj1DO?b>T
zQ5_dwq`})1162UtUyz>t?j0sn^o+px^}n$gPG4+WOlhq=?$&C`%x!IM;RkGL5+>Ob
z>pMx_=YkIiXOsXP%{u>KUKyTnD@{?-*r*7Si30_wK@iku!R?HldRVvNIfb4<c>H>~
zWJ|42^_<7(NkCAApq8-y180c2qy*z3dqB)#U1l2=so60Suf?EJmG%mM$7brI7>cSM
z@D_18vX8e`7bWk+b|h;wQk0Bbd$c$sZjS?glQT`df)D4-Gq2e61uQ(xIuv;Q8Hcn^
z4=}48=cGEaqi8h}126vphlY@hMay~(0SOfD*xh7mBh$RTRQ9#@v?C15x|Z%eWOh0`
zq4T%QW780z&={m)xVyU(pt8Ee$ur}bKK!b_d<~-Cb+=QHhQRpt5|<icvJ09GB|FH#
z@LnvkyukdHg>hceLy{whF2~(}f^V`k6_&o2X6!=GRG0QxRQ|av3Cn825P^oxbf4OG
zH1qqL%jAtwBI1*(uWtGG0aeM6@ZCJZC<`KLoTucBFUfC+Rar+*qF89>>9)nb9_kLu
zM8v!3<>s}(rxkML2f$7%JW(OT$UP2>K(?FZv#6*mOc>PPB?9-X&Oee&3$%h6YDfWC
z0lZOa&ZSaHX1yJ1?q42*3(A=phCLGw_XP|>-1DfBoRon>vz~`Tw!&yI%IXls)pbHR
z?L(qgsP-6T2~gN2$GG9B&`!OYN=_N)ey&G*Dy*b(@{u^yJhRR{Q+j7mN>{Tzb5Lze
z=l<!8Wv*X0!q{%7R4`bHzjwX!fK~b#dSOsKW;^Bivj&qzawtz1j!`P#3xu(Yz1$js
zXaoG!?+J)Wk6kLC2P2ZyK97#Pb(px>Wd<m+%IrDanS387<-TL$z3jWPMqgnX-{_k^
z;E@=tAemNWxvJjuXKAtc92k*)TW!88+&7LkX1s^4x+5L->LREnr-x%XJDWayV>Ia)
z!{QAG3BNhaHZ{}8&)NdB$vnMeS}n=Szg(0UW|)|s_$2?_hg;*Ve}gM5zVp>)Mtv>%
zAKQ}U{g#9COy|Z=%o77+8uF)ki?f?L--g<ng%Hb6<Z(s7?zVC7X9LPSkE5QGyMR2H
zxvjRvV$U&7v2I8(64li^t=o%|mxa9sS|dQtmpQKl2<(NX(*W=j-b~cYR3HVuXP{Qv
z?(t6wuibSUI-=d>RMaeoVW7(@PmD=5h;XqB4dj9dmY3!<;a9them^ap9}S}(x=I2^
z0}C&Nzjz)YY-0s8I6?&YSQ3agNM9XIfz>y6yHJ9cNo?XkZ3f^9^J6pfENU#st)z$?
zbZ)K#Jeut^D6lt?Cu3a<20D`ocn0&49XFiMj=`t$@;b0Aqbu+loEy6+VQwVx>`Go~
z`*zYwzggDWYMt3}m}Ad#$?RnVVxqO-%Rudc>UvIJo1bZMyti6R(_JEYqRP&^G`f~o
z<ZgJ+LxE^IIS!ScveqQ?CqTO{R)IW{S83X7UIdP&X3aY-QO9I_&oW%cB`xqI{iA4{
z4y+?(YZCp@%aZ3oE?>KQTNWaIolWkp9g%1*3UWg7BL7-Afln|YW6xvV+ur{z<!czw
zQk&FTEf`dapCkD+N5tDR7|R$1V0WHKomnqb<9Z7qE5oGWpZ;`NTh-9wNcM^e-^YL#
z<eJLU_$B=*@ajtxU?Nk$k^{t6y%YrqW3Y1cvL~!&BYEq{lrq^O;HKi0l+ZVuJPJK+
z`Ls_x>AmSlwtQi*PiliCWc{GFN~~p$%`Pw0M`I8WqEsMVG&{xxp8cIiQZ&Nih?mZu
zbkd`^7udU-ZxT40!^7E%#e2qICC-9xvY_sHVly}T0NlCY&o;Dn1w)c{N2>s59CG41
z!Bp-WhAqg0nNqC(F&H6aDqL6xG&f25>KF&q|CkMmun%c{DHl`1E%Mx=25oXqHy0b>
z8?fCiwP}XWy>3X4E$&5Go(eJ*BaEJZ{E5{sw)mC*EDzrs86k?xRTlky-G7Hc%!g9}
zip$tf(FBO#@45a*<`TY*RPBLZG*<!+-MQz*O8D*OdM|HJe=g*U0W&0#PjAF2ymPz>
z4qO7Vvtx4{DAahZ7>pf)^+FHaZe?^%+k=+oNeygFl1UV)k^xfw-Z5?}YzlvP854}d
z-k`*1<bUOCaEwprr+oi#!%E6AhaDJ|bKtMv`2pEB9ViY~k5w2Y4=Tba(s2*vy_gfL
zoci5O%Jou*PXRGqcUg)s+y(3v@4X(zmBR~i+nQMrqCF!JEVoBXvD<*T^oxy)ok=y@
z6d+x61-DoKSet$4jVctYZDElqq_&@>GX|xCdT8f=itb4~u#TOD)Jey*&Vjo5?0+q@
zE)$9KBDF927EbMQS>NW-NVk5uCARrSz<h4k;on|gLRJ6yy#eh`JRBv@>^C&(Yz{N;
zdh#_eyV_>M$(MME+>iZJk4t$lK$GFzVmp5KH|#UVUHyagYw=x#{KWARwT5~J3t@>B
zAFZ#ota(8wF0kZX5jDpToE72X+h{4xs@W{DCsx^DQ8uCf64&AHBYB+k>bvrPk#~Nt
zctdW%FNyouDk9}n7RTE=03fygS!cH`srA6h8J_*iKLY=w!KPac{eEa>DsG@4wOiTE
zA!hvEbPWnk4&<YeV5G8=ww~*r&U^fP>x-VTZ5q|-cUrH_d2Praj|CMm@8!0BlQ$oe
zJS~7DVz9J-D1NsVuTW}aTZgA++Uf2x<bgf#5cz*Lj+5t*zRW>hHB#+oETDIv@e%wg
z-xI@agGpqOWoh|yl`BIp^NJtmbl7?#S%iw=atf5@n+t3b5zSTEj9tva`>}dAT=Wwa
z{6xC%*0X0}s?*Be_2Z`9I>ypw--@7G9h9k*T94dF9&-AMvq-X<k&VQd*`6?1vWKgv
z^XY6qHj~=z`QQD%RFm4!X{(UhVC{eRpIU{1gT>CWdnJPTCt?Ik(nw2m|3+4?&HLN;
zlF_e|$-L5P1!5@bjDjFI5#p^NmmR01x-R?x$GIFhKxgw1AUFoc3w+Q^d6X@Ai@t?^
zqP&cTfoj$Tjou{?gN9qGA?r;Y>q4T+62~&K_wL%P(>~Qbb(78kubB^U2P~(cdaDoI
z&KKXc6lkwGD<SjRof=MSM1fN`olE@Q$lgSe{34cn+dgR%n&17=Rl3{P{dyQS6wQ8t
zrY`%#pK4nojxXpk3s1hFhNOlZu>aVM{p9lWlil-n=}H5<mliW^p<71qH17s02rA5p
zX_V7#M0z2DS&qwWdmr6lq<5XW`WfnfZZ@zSK_;?ScU4uWJUVh-ADF@Qg6xp(W^ZYH
zrwN|xtdV%8>B^1=O+Ygi3RWG))_k617=w6Y#YHY;@(RC|O9|$25tgb8*+7pssVyW0
z{mFt}S9)`cTANXJ0_wB$_}AUVtp%)B%P?_+nP}Qkdcrq>+pMLna@Qrsdg`UTr*Jzq
ze8DG+)ckZG$H%ztU4MuC@nYeTe}gTke1oV&aQ+uMtwgrp+8_@3G_y%wbIEOb)Dk|h
zGGd<_e5r?qW%O|8A-xmGK$=r{!*C;_>II6ydfJ6AM-ce@P<4##`be_8ODa4UrjCpT
z4uGnYz1z~<+S9Zx6$?-CP){d*L~(sYmoI9MIp!%sZ;NA(jVSKqf#&xD1@VfMejqdP
zneQB%%Xqc)Mg$qWRK_Q<WjlTs{8oH2m{CA#D+dw&aJlMiyF3jZ{xP2#+S_|!fWqj>
za4MHD1uE=;3w4O)-c7;YWl_oX56ka-_CsI8Y&R35zdca+33p9dN9nYbXD{Dvg|CvL
zCpE#UpTN_bJuC;k-{S+l`IXTwG*@kZ#J?HTXD~Xf#CEehJ0MHEgoc_W(51+enfesN
z3wia`DBWARTE;4G%eh5LT7o^{m0`7HI<~rGWI>I*sJ_k!)Iq;OuU)+21MYZ>l79={
zXw)!;>O?s1TMVzO+pVZSY4=B_8U{O4zwcmE`4?HFl7&!1*`8!ST~1SO72fwCLC5R!
z>jj9GhsO~&CBp>h2R99a;`P<xC4BH4hJ*9LM}8-*uQg&O8R=82AlTkIZVN{ySTx#4
z(DM9o4$xRNU`~{3#;K-SY5dApUXNgU`)zrqKv@7$EJ#uN6|y_F9FSk$e#i7<U9`w5
zjg@;EO%RkMbRl!H@slO~nAiodSwSnDK%Px7o!<4GOkj<c`neput2&xQR#-Akv<l_r
zi6CmvuVIOgKUGyd7mJ%xbjlf%Ki*di@6br+l?t}o%DUXQ<q3`R&5y5vC++veUqKOb
zBddmU!7he-rM0FNyC3dSyCxBdbdC5uIfI;L?tyyNKc$R9I{KOd5TOV>r$fiNdRl~o
zyF*&_<hZC?!Wl;+Uw6IVV{NAMAC29#x{(zr<>&4_0=1OW8r_kWX&AG*+wja)YYW@>
z=tw2~y_MXx*wp?5ZXc9j5q$xqKY}a{o_j#!)1|ZPBklPUhZwH<sngU&aZ(?hR**MJ
zXv!@1u<2l-qq`s~T9BjrMZaa^X0e-YG#x6xd%^@O$4s}GhY9vsS}5Kd^z=uq=3SM{
zT?2Ev)el{@^_2>|C=DaUE4Yd=6juOW=EJk<i<Pmg_Gl`1y*Br66(!rON79iYuTCu2
z%`IPGJjT-hX<Pedzxe_Y7fpPtzfgtNr3TYW&UJ9avyMcrNtcz;<*jO!R=Ik5%xIO!
zO4_`p#E5FAl;yl$@)4?3JX^r`*wq@r=Vq+4Ve)6gJ`{aVy~f~)m0cbdN$;Tc#lYUk
z$8bb6L{E9W-?v70%ngm)g<Ea&t0SX7b(OdikfFJ!amL~QWya_I3tJ$bh&K8&27`TG
z2KL(6mP2V7McM%&M)n4!r8r4mZE$dRlnt}r(g<ix-TI4+_#hCF$M00dd-M}5`{vS3
zhvM+Utg5zKBy`DapWI4jSrEbPyD$DL4fmjm>Tj9pch)3c`T<md38anX<k(ib)t3sx
zkD4mBub>D31$Zi9!Fw!WrHC;sx8m5o_1`YT=s6p;Y+W}sIF04=W_I`^H5Y{|Unuu-
ziVNa-70$Kl4sZMvX8LqV+w)J%Ph-taff3Quiufx}>)rJcL9*(Kk1~PIqQlrL_k+pi
ztm}bNe%c8*6IAXRY=u{+jdS4Z;sKPg0!?ZSP0>f@172wgB<!qAQL;F6<^PD&*xT1Z
zZDau?fA$6P|G7^;Ni`2$f2cvbQ2eXJ3X*m*9L4v(k*1I)KSit2BnT5eV2KN((9$Vl
z8(1p&Yj+9q)$0ysV%N;5pIrH3Bux)Dz2LrQ$%$vNbT>0>%Tq~Nl;v<+pF;Cp<l}6v
zuPq`!#{!4_<$?%qqu})gGO5kvgXh35b{08wQoi`j{X7YKVF)Q7_zQK%cC1kwS5Bri
zUpY`|A#=y(<06~7n#9p<S^YT24I}23f_y+V6*y;J09aS;wi}#gfAiU%_P~o57XfI}
znSH|Bk?`pXgGpZwKdr$Nt%Wu^#{2Cpw4>4LHU=A7<xB*{kJN4Qg?x(i1h5lxeu?RO
zNe~rzD!SJH64XKSMTq1+^Kvt>|Bo$LNxdeecyV-@Q3b^&p$$!cJLp}1WNe+<*QCc<
zA5NgaP%(dhJ@aO~O$#BiszUoC+79GO1P}u^kTFgqSjwH(dmxVX%6gTNC$tpDVznZH
ztN5Ye9jVj}x~@0sTP@rGL5l+x{;en$2KT}y{W#~YflaP6Ch9I|%;nk5?Mk1XyK@D}
zd-h)@M~^I<Z{7>$5)G2Sj}AxKdJl1nDoy^lzp6`RVkCm5S9A)=WO(b@+taGcTKL7e
z8oID3zPDwCTUy4b*e>h+&yFoFs{6a3YW~0xot)lc>OKgE1D)$<!>=Wla~=y=EsH*f
zwtU8EaJfJO`I<r{oRXXLPAtcnETeR#q(+dS#)m>nMR~0~j{uuKmMp$~h+~XSS8g*|
z&78uRsE0mrPRn0b>WdH|d1*r$5}AFee-W#SvSWfE21oryLo>?V{UPS|zJhY^)8Q-Y
z(%;|p-Z3OZv!_vR&7)uZAR;5yV`t<_<m3R?p|7dySUXk!Y7mdStDC$oGG)#kf0iFT
zm*xpVdtvE5>Za+Ba$2EVx^uORTF^%Pc<&kNi&F6>=<3Wk6UEP1Z~J9Wg65|*Y1kpi
zp#s~I07uKM0^!N4->KB369kf|7PDLFf=DqqTur0Ck@8L)*t;Av-!u6OncS7I=Ri3-
zT?x4^2Z_T`?R6wsCw_9k-n|E#bA(wrx2F^_5({dfP94o@hHR>x90z{l&4YBInp=mu
z;xvE}PbtZ?s@R+(%;A<t^Fj92%lOVZ2v2HPY$bl0Kb4W6Vot_~YOD0zw$~eATWmRA
zKJS%dC~&)gNZQuq@9!%rA#>DMY}SiqpXaIe#}ZFU)|KP_Alf^+*SFzyEz-T^xu>~*
zuulNiSF_sG;GsqT)av4qeEm4&Frkzfyl452FYPGqn(uIvoZ>DWu^r;S%7W3~`FzHa
z3x&&E)4a86y%6B><I8lcC;yiI90}@;Azq4fMAX4VR{Rc?+s~HUAYSGdn6`pTNdH`f
z*E(LNT`!OOx#j{B9o^rI*+ob9(W~YPd$C8oLZ4xuaMW7vtBt3XdnL{nh_AA}<9x8>
z1P8g%W7^*Jewcmh#glIDDfbZ}^J+_!r)_d<jkvcgI^SYscL@9L+6J~}25o*FP#5_R
zMv;Ge!EZ470Z{%=0{Y*qAwyqSc=VK<G;Zg}_gaxB{-W&tx8K+^eVcgpVvKqbnMVtR
zt_xjY=e{$_znKlM+?PH;_LcX0C9(I+ZPpsIa96I&w9pEVPSH;qS;#e=^{gSJql<ma
zn>xaS5xUcVR+&$6gZsw<eUY7hLZP$+c^ZG++Ka^Mc#)nRsl=(W3jwCFE(}cf2)fQ8
z6km!5ie5hNkM5W|#^_v3J;Zu?ixDI%!jt+*gRfMCEv-ma>hJG)SBV{s4DG%*HrDy?
z7!M5044$09kDD~)z)Qs-1==8fwX33BV;iM{M|6_yjIkXL7I@CMJx+G&N?e%~_^F!@
zV!yzgF&`p+=gge37LbuQ+8Qz#kltMxd92NcB}#q?sRURGW<e$*Ud`p!isA)bqAb^5
zngt4l;>2A1LdT=S{_knAS^w$YjUY;qfA{rHD^ct}-4jdIzujkT88!a7yhA4J@PdEK
zE%vz&-oO3uS@bsxRLy>FZtyR;U7`cn=YP0d@tG@drxNv&Ycm@{iWbF1=%umwDO2!N
znhEkxTEfMU%GbSrJDv(lCQ3IdvPQ)!Lbu$r+qmpjsAl(DFIt>DV?aa|pAt@3%*1yq
z!CC(G482$8k0YgTv*AWa6`J~7(QS1}BL?uNT>8U^DDkbk^SgE(ZVZr9)M))zhUqB^
z!r)S{QUeSd7r*}dTL2w>@|VsSSZdEp7@Tnm+?8{wT+c<TrQK{~gh79KNQM;eFri7v
z{my?Onx`K;;<F>{Gg+%I9<{Dw=Ol^@)<e$W(zgLXCVvjcwVt<6IA=oPdaR~A4I1^F
zNow6DTVNS?z;V1|j(;=UOh!upX!6FOXBJm+Zdy-YAM@plC0pQaSHS*|KCeo{QK?%c
z;L`4TML6OjcmQb4?S#ostR_@mntVFA!L)aGs8stc3)J6252%>&!FT9=ZpD~qGf{Ap
ztO2^=|DQVLeObJ|^yHeq%*^+`5oX706;O4D+XIoUcR^Qc?sp3PgPOyaFVp<7aT(vk
z^Q6l+Z7vUJ@+<mY;HPvt_7hgdkB1f!bM+b0!Kh^)eIZ)QozZZXfd-@BRc{+yZ&*F6
zOgc_%Lf^-h;pSU6`)hxe@02oACu^7bp7o>7c2BWql5rZ)T(QOy>IA;sLI~X%3O2w%
zdsosb7uk2>T?ovu%dWgA_YK717yJRz%cVjIajAviA7{ykVR6VGmo%ipai-}-1a$9#
zuRQJJkCzJq@wMUy2=~=wu=Ld<pUH`E6=vxbjGyk6T|S2GD9zQn=H4s4)GnOw71F4|
zQ?CfO>k|=o<Ixz4WtQ}t<(u=PTHd4<PcBe;OStU5jRH?5aiDV{21D3X1BmN&dLOfJ
z-$LF`>s=I`M(C@fv15g1s<!kd?A?l05g6$--rAGhsTsW3r)3mS!lL5D{C2Z^@t(cy
zVu_SkHGygQau7l)W=Zgkp<elIk+QC_;}msG8}JY&zjUdL<j{hVX)K9i4-g|@0{h!H
zy#<Ko$Vt*sqFzB6ad=+#aBDUAX6RuR-&}#Sz3Dh3z?(AI;C4{mT)!k&F%@NN$V~fO
zL_#As;TfL8kCNu^wWfr5a}!Sav55K4<Q0UC$_&g|O#4Q^kw$uVL}JHlLoS@r+Fdeh
z_m5pha)G5Jcf*5FZO948$3JQ=7UH9}NZFkIb~>D$V=tG8cvqF*H>9hsaJD-vZPaYi
zpEf6h24ggnp?p;l-IYnYcqmn#F9LAc8ciS*Ea&?YhB8axYqI<NE;Md5O@OH>#(21K
zF-p<uk<)*oVeXzuAaL^ui#0m!)`X93-6f{;?&kNEB-9>8;<xw}n@xw~4np9Eo~lC@
zU9_owv)m672Dc?#4N_s)NacL*I3LE{u3;GYw<;<I`P>vm*-KRD3~Z9I6X^tfl3q)O
z!~13Oi4?G+tw)a#t=5+W)NGO1oo!n4VS0$S!&Hoey7@>NU3CG=hw*2f=)kIfw%5LF
z?_VTxgZ2^Kp>;USgpjq(1j0$R6IC{jqJFpKe0PVV`~aWt7hFa6N9dVtrsOlng5O4*
zWf;t!$26^2!IN2%=5u$e-~m))kYo^9#K7j!G;iVk@CmZpz0NcQ`XG3)@@;WSsPXam
z)_EVqxa`2EB&hRx`X+O5ShICPVD3(v{E@xIG53W=!l`QF5zR&umW(?u8U90Ap#l}y
z{RX39m^sf2;m_j)x@`GjQ;9f#%?Tmo?g$@$iSCHRDZIbCn-m}RG@Rt($kQw&saLGZ
zXe)&s+4T{f(1vOK;78aTOOh@PzGF1W{899C48<JBq52>(biZeWb=I3<Ep7xgtT|CR
z5g?tJ^T@Hu*LBepr`U9<OaBzK7*%I<oxxzP%R{k{p1#v=+vL+!<MXjZnQpqqT>wAs
z-hy&wy!8-o$WO%E*HY1E@oaE;e8$?CnFakc+ZS+QU)8g6Vo9XCVcbrVUNPUO-FfTY
zX^>khnMdVn%ZIMqi0i5@qKPFzlBJ7bCAaj8u=qwf*6(yA2D6z3Va@iGy&)QXG{Mp3
z?A;~8L|*M!Q=H5D=t5*`L68)6lg*cr8znc#Y^fe$0s$8|SU=(TSi#-JV5o4TGl1$k
zvsD`Nu6l1q=!SH+|3%{ywHBAGzDgATwr-t?Ik>{p#d5paMHqTM8~zrK<-QGWiH-M%
z$G2zuFp6ulnAqnqBI+7?`U$V4m@;#P4}v=wcf#?rjK>ia`jMR}US<lJczBp&wtged
z(j}lJPW1#?(^%(=U$0n*lB$>HNgP}kgd7vI8Jj^jSxqzcznyw^`>X9ATcHHkX-$6n
za9V`5#H?Vrs)<NH5JyDLbp(BX$Y1lvU#Rp&@C(iq=4St~U35@XZXA4L({WQ|7xJ*{
z)y_cT?I|nvxomYs5Sg_Tky%#oiR#ax;m^R_>Z8akTuJ~5CunM3Tl>ekb|V?HQ4#5L
z$43E`1P&3c?V;Vb<@$kX8zj?q=CgDZ87|21fdYNcw9@+Q36qV&Mn5a1zW}+)=@*{I
zriE0lK%XJ7nhNil<w8YDMo&~PnhIG75-j={p>N|I43`{;M~vwEF&vbZI964JIX$d?
zinOcO@|lYOdCOIvQevd*(yEZskP`<ruM0Z6(`g@}JhcSCc;k?XXudaACt)bQ-I6~{
zQS5*ni&;-=f{KmywkixMNO0O}0dM0ghe_s$!s^IWu|KKP@}!gAn@EEC#-yW3Fp0qh
z><gMXmgC5TOnvB_RKq!revawZ$rB7_ZEEd0zB);AiqOCvFiS89nXeB{nwi)B02P`S
zN2lbI^Jp`vmBqv)WMA{#={T!UmW0n;-1b94%}W>ZgGbkK=xa$jQR|}Sq4#b$su4E|
z@&u&rAG=iC#l{Dps(R9Ff{k$=g#K+&MLoo!G3*fLv&I+RH&1($X&}bD@GSL4M)&|r
zK(xQV&Wye$hTRy8!k{jq`~%Wfuq1A6JE;S?s$g7_vQaj!F7@=@aq`1=xPUPktkuFD
z{qaGm{e#$No|yTVZ_8tI>tyjSOZn1D=8yJo(+}3_t9W!+kHkdT226W^9SqfSb%vs`
z!c_CIFsY*})J$IbvP+&Yv63b2ZYb(GPma8czWNTVZDtRITeP`7VJp)Yv=aQg=wFTc
z9$n%@gxK|r&DHpjK#D2?EUu@9eUY8hKksju;FY&*33)PwnlXU9JvQiC&*#kWD_p)c
zYIt+p<pnXV^WH=1{Y##LR-_|Cq}iZWdlf!n-iH#L>+V!8I|_^t(xX7r*xp+YTi4-?
zTrczp=iPZaTk}r&j05?ujG)6PZ52AdmU89;Pp8>HTFLuIV$5`Pt@|`rkmV|#xn^x)
zn*h11sf!Xz&My~khaY^f){HdWXas38Kg8m*2eL-7c>%g4Q!k|WBr~v_gHc;3{Y4!A
zfl6dw7n&Cy=VC_7{|7>)(q75*E#($E%3Tkz=ufy;u$hKQzZx8QlE_7AcmApxzm!nW
z1bJ+ANnt%4Z_iTo#j%T>J#O|S&tT@EX`QLe|2&E<#P@*{^!Qx$L{g>bFDTJJx_lU<
zr+QK7-*V(*2i?_^&7pNu4tin?eIl>D4xgzV_h63Mpa1luSdlQ(Db+(^8*y%~9rfH)
z%FZeM!H6*3qQ>!a?0BQxrh!oisoc9EBWMa9>>R9=?e5U9gt=IZxS{YnR_rzZ1}@r^
z#|KdD^x+GKk9)9$-Lc_A23-rHdWGu7wta@VVYIx}j?{B4u+V9KTZo6Cq)@}ZxH9Q|
zI<Rg%RZvoGJ1gJcQf?{h`JMDxFqVv;re}NaX8!OxN4|O>&Zi#dA0zqV#LPE+FCWt@
z^6A-P1u^HTHdaTIv~sh4Wx=9!rJ8pi1nU6QA0h1M?Xyls2f;F~dWGito-+g<WIw+7
zn})iYy;MEe%zwi74}?vj{M3gA;TA+kZkciKLm(3J@ggbdM=brU!@V(pk*!iar!^wX
z*zPYdq2aq*m1nF86L;QsbWTNEzF<uKKFx_)vk-C-qsGjtM$-7HD3_f}dU`O0SYrjB
z_uGgtQ+1#a8XYSO%@2rj7aAR%bi%e<Xv0C8PoNjB<6tmM?Ku(Nev~HJ?<oz-AQ$>m
zD9Z|lB1;)2%PVdLy_XSsybd0e`-;rOX`P1<*%$}+Be;ZL>S&L`1%L#@B5qt$**&&R
zGwn)@H@+Wb<<;N(T9Of7Y3ydVIoT@uMKr!VP#IFE<*Y0juqvp1LTp68nK*bT?h@3Q
z-^J~^qA#^*|7+=%KxY3jDyAP~r$Gjel^wQ_1`QHD2KHhDwxvSNSf!v_j>G6DX(I&n
z#>5FH0wJckXD6bwyEc6Aa(jdQ=E_RPXbHP?2{b_lTn7ye^$dvLaUucL^&BfqU&|vc
zG%+U5M`clL!FmlS330bK91x9LuX#*Y<6(bKfO9jFtK!yEjsoTW`$_LDo&TdhkKwzs
z$X!e)g#U|5n!*=HKY3D5w4BtQ2D|6gN)};ns?k91X5+RJsEv6>1%8+gqv*H4N4mF5
zWtpJpu&7~^G`0#3m_fc2ZxbfSsLp@tWv^#7nl;hLmG(spsQO+Skyt$H%t5?bh0HU;
z=TnZuG}_yYAkN8Dn7Cw6cAq($k(UlXJLL>2bg78ZCMA9{&?NKjnUPZlE+*D!bjo>n
z6D}m;@_qE{wNIfqP9(YPtD@wZGKt)H&arWDF6}@0^yEdOH!kPTSd;Vh9KXlFhVJ?D
zlQCUiNYE>*9d){mkmI)rh3Z-|XrHJH+=e&w=q2C}x9iDkg)ovxXa4bH{PPy4D1;Ud
z-h^0Vm~|_CN;#|SESjj+2d|Ml>C`5A!zpDeHp1>Lwp109jicT|4^7Q^fn+?9nuY%#
zRbLqv<@a^1fTT1?cT0D7w{({%9n#$*9g@-=(%p@8cgN5@^w7M+|M?m3r@5}18BXoJ
z*IN7RbDuJ;DH}g#^6Je;b$infhX>fqJ~{SB4uRR}3ukJZ@I_?+qJ(Y0T=NgW%|eKv
z<p(@R{MJ|U-$p;IH1RAMKSevo#c5ED)Plk@-gDM>oHA@i*Zt#6H`8#Q*F*~iNWwde
zf8yH@DkhI>tD_$vgkH|Daam3SR=2pjSd7s$$*~VV@Hdn7TR2$T5<<6_s}>8?#F43S
zS3~jb>L$qJXBKhWKzfcCSSNhFci!=8TOKB%cXOXC-yVW7eWbhTGK!{XE!Oc=9thgH
zm8FR2o=cs~n@IMGb|J(AGJMITei+`W;jGQ9zzQ{OE9s5l&%cUUy8Wko#6Csae|Q(r
z9n`-Gh@|yFl#hNwt3vmY(e<EZ(&X;d!de}_!IkW+UPfFBSpT4@Ou&nN?Agt+6WRV`
zO*4AC);$Z?`UP;u59ZiWmmEqyE?VUo=w|{*9{Dew!n(WUj=WtJBb*%A{nbV^=^E6P
zhX+#Y(g6)4g^mTa_=qeI6|Yd^3wj}qj`qe#!cEj)lKg`TAYq}d+*KI0{ok2!Ar_L3
z!zWvTtXiLhunBTwIif6cZpnrAFuPx63CkQ!-kPRsWXWIaQa$Ti3RE2RTIwzN>h?&V
zxo^wOA=W`0Kg6ah?^b_ECnpDgyHN_t-i+lk4%O9T@i*8RuHjR0*KwHnsMp*6T07Uv
zF+$T+(1oLqy7&vrYUmsO?3O11#xHE%tMzESzzYFjwXOb-IP7|mwC3Gkf{&K$%YdLF
z0($=l_pcQAL1Abd55GUMH=9yQ7x=;|cp*5<OQJhD&La~(aVjN6|4S)5JV=O9?|^TL
z9pC?@AsP@M(e%yNt>?We5+<&j{&3%Rq78`Cl^}FvCjp?WGyE%LWiL~T+Z0^|;5+_Q
z!vMCai2*!eSnoS!y`Ss*Qi~hkiU20KQr}T-f}y|+{EOfXoaZtNRX??a(@!FxH&SZz
zw_Qg4E>b6%xHG$;eYY0gr+!4m=uaqApwGpc{i5FJ0_PXeOMi36At}CGm*}92?-Rer
zdQ%9cN-CM-SVrW;6|%7thNRnlnU-kxw0pz5jb=g0zf06LbvoBCtt?EIg86GV)`iSN
z!m@@M5b9mqVi#WnYB%tPKN6x+gTok3e=W#P1M4E0R^84$gG^%WEqbUkg)nomkDs_S
zw$^>LqVXjZZGFeCQnX&CQ7I7ZO<%93U`8x7un|K@nnk7XSJ&DM$S)@iGauLx-r5F2
z6+$56q-6U>!A%t#`x6*Jr!00W<>Z;&lDfx0x9ka}qJHviErv+&ZocWM3f5?>8zA9+
zr6U!vcBzX;OqF|ldI6D|#0&MU@fDK6X<wAX>?WT%sF}x9tab8s9CYzsyIQlg+QxgI
zaFh85+BwI2TW>%6T{r_aJPw4gZ*`KjADh0eVc9Nsw;%>pxUZSZz*_}Df4<03NmDi)
zAl_--tsy#@*pTihnz&3k1w65;wUt-8A1VLMJW7c{2`snAp-tp|rv9rvj$kXwn4T5s
z&d_d(im2RG&r}sW2)9GaX^VcfG8(iCr@<|OML4{H2FRbl{8(|`k<eg7FdXr**f)0n
zqzT*Co4pkv7vje5FqG?Xok2q79h7uGQ7FYBXQA>oWpq`1Y5)FoOk-QAm&}6S_9RG<
z)8NC(OZI3}HnA5%rA|*IjK-s)%=WXKxduea=(3Vi%{^s-%T?>`S9`@@R}1?$ziQ*B
z=U^|*)!|x`Ri_n<Tt`vwk7_b1O^7Ity+qc;z1VeI(Sz>-p2gr8H6HHN&~Mf}f+u{e
zd#g_qu_x<H#CoVB|INFtwnXg$%7<rLo~bd8&6xq)#dCM~bgA?8BZ?QQldH@txBw&y
zLP7ysNPm$orWO^wVeENaFjo~FBHI0NeBt$O5On1=8Q&K8Z-~La0|pdb(){%`3kdpx
z6JZ-F-iP@HU+>gz>X^sXP17H<S>&TDNWB?LUX?-F{RCEYYL2%kp$zpbA><jN6NkJ5
zGy>(aI9`GTCP$K=ZB1+FH}eM-VvQdP)C%m*Y+&t`K6xg}5EjM@hP6H2^KdzBqWX7r
zRo0lg-FzC#w7E##a%<Cb$2YZob4Xy#U$g~HHOpR`)LcdkY5u5ar6;FXw6`73PGw_s
zLfDQ?aQEbA#ASa)@{N5sY82CPfM-fGSFAt_rXq#(+VX}x+%aVT$F?r4n0V$sh!r3m
z@>$Xwr?T3lKYXU)o(ZZNrnYmz(uK+G*5-)E4JT=GDbrDa^X=h>CF-WMdlN}#5S%R2
zZOTr6?26Wk&D}xkykg{KW580;MXZJmAZX$6i(mFSe-%%*N4I^U%KMC_>$YdQB{zsZ
zPg`H~*5t*|m4Y~QZ{>`g{*zDT3>oLm5kF}`hK55HQfcq+(pdP)x$%(JF}V+NSezqW
zaOz>jFSa^nbFi*vBERR3#xEhCiU#o%$~HsHyV@BZ`&AESm6ub){DZP65X@-Nxed&u
zG4Ox(RP-QdlyiwY)3yq1Pf{5WcYqg=ybTNK%DMY3t==+aJDB2iE~##>`lO=L$<_`r
zhKoL#zB_8h-4FN@Ksr7a?XWyP@8@$rp`x}W8N=yBBlBc)%Ia92szcGeT?iU61!8kd
z5LtfD-dVkh|LA6zNLd>*HxSF6*48rvlXbwwF1@bX`YdJ+d~Vu@aP`XpX6Fn}l&f9m
zH03R`nim|bj7&(-%VA?CbHjP#mdQ9oV?>)=&n6t4p0}cuQZKX>5H}JeoT0EP`Yz;3
zof+LW%r6;@-UV!p&q~YjQ!FGe93sDL32KVu9|yZ}*`GCqdX{VS!n_opEhEB^$rB$f
zAiSK{HX3uxuq*5MerQ^@s_s<2?Cy*E_|%aW6g5$8_a^?K4)Ln>Z2VTr%g(P6*raY!
zC;T>W(}SaTbnR}vm=cq8=B@X6+&gOJpVV&0*{!AQ>Eo$3D!)AC)XQt44(>AmX*nf4
zXBh;i?RZ~N(++WGXx7ukD-F%l0%~q5KP@R8@i^hcBkXb%ftVF;s;SrJqJ;HN{(9E0
zL|?JJzQoj;KbsYBD27R2x2PMZa?PA8-U-N#eyO>sa{#H@FXZ>w6MxM$(<LZjZqOH)
zOq~<2By2f<b*Q+~n~3Q&u~X@Fi9Vo}9+N7b0<Eex$$3;*%;QYE^C7M|U#0xYDd^p6
zc&~N^-iXV#dmr|7X`buK+L!u)M3W11VgH=Np>1c?#}a-w1V;VVPKlT{JF29#&yr=v
z10+5VOU3?+-@ioQKfS5m-zmQWBn&{OK%Q6>D79TEKq1IXH=oZVs%EXgzpL@BxLIdc
z6QqMrilpY1^#`*hhaFRTlk`{B%vi!UZO&bB9L1{G2pmMcNR!UxYNU)DM6JH2ak|$R
zgMp}M*ZACeEFbr<YCllBa5!~8nwEb&eyZM^&D2eWW_kizU#2Pg9iw}6Q%2<7;j6wJ
z*)3a&N~UOuxcf-uzRD<p9gmZa|4`lTjpb|P<{5RNq->BdpjEkxG143?BQoTEZS@TS
z<Lo#i&)XX;Gb|-sf@<S!UMn6B`Lmt5VRfwU7k_Z*_@p+>bIXsLFjwSR<MKoY4v;y3
zM7+N+fA(^qz*Q)g@{gloLJY8rv^$o`ty4cD7TD4TJ>lBh+<%fUcln`cZ2;jz)D(H^
zQSP5}gnakrJ(Y`Y2Cw8{b?Qj}jOnGkn})A}FN3naffUUCr1pALT@PYdcx2j?7EW5<
z^)JCwIOr!YU3&df|DHXQKvk6ttNk-U?V5ExSp3Q{dr|SA(!RY~)5;=63O*{Wh{T#q
zW3!+AO6H0ze9TGo$4Gm!ix10KjLDkOagsMWNk|HdDa-PtmpMLNL}ViEB~f^~EQ>}2
z-1`ez?G`}uFVhW_DXg(<%UIxWOa8fR83~H_awz&pK{w*Fxz&{R$D8G=yZjO*pY2(%
z$42MqRZ7JCE8CxG67DW_QS*A}`yct>aaACeUJAhu8+r!MM(srx6+(^_c(lq~9gCBp
zf&|11n0>M&bIz#&$3@Qc*w;Oi0&63>0br>y(P(47?2ZyXK<m6&93orkZH3%jMD(CZ
zdMU_d-<b7>51vMX_1<&jtGveFh_xk0+$gbE4H*>M`XrCKZ^Wo?gh1g*pvqb99v3Ce
zkbij^anO*Ri2nY&a|;gl0W}ft;jCrnLZ_m^dsFT;(dz6noKAbL{N4e}6wEOr?FWV3
z)cI=%8L=TR#y#y`GmTe36sBd04g~yovo@+Sn4cadRkd9sK}HfFWO$F)qHshxVU$#B
z)2AX?_}BBYZm?E+Pd&>l*=NA!aA(9HY|<PBFBKy>57?-8EnEw^A|&>UM8fkISo5AY
z5Z^qrWTEgLws1l{xg`*jDGZA+u5$^Xdc(9PMyP{R7m_PW5G^NyzUM}n^t9?4C32km
z<4aaL<URzk5YFY>85l9s>waPI)z!jRjeDY7<Ttc5@wCI4J9spWL-!do7o5(9eZXRL
zYIp0VPV6X))i5-*4xIZHhsCs9y_egfxf&jOUgaQYnSW{5sOLW$5Bmt!LzDKGsuTHw
zDXrkEPmO`UCp%Y0NPLn~h5k#zN|H>7Eq7>Y&JG^tKpElGWcS=zaO#~7jQR{!?27YK
z`hw&So^4YLVWDQ%Ft#hLD1&Nz@)h*c8*Xn)6**evFxas!GOGAbX^&N8P6%*!gU~Q|
z_iBDYeqSRyDXN2zMKW?wxyLV5$e%oA7A=!+640lsYR1Q*Z?BGRqaia0?yXjHTgO72
z45loghoH~y0mXiM9~ctbyj5ZN+DV2KGKB$FH<2-L)s??c@u<iYPV&Qq5>?fd_;;gF
zv&M{#7SaS0fL+0Ec%4cY@&lb?z3wMb{WR^Clbr7@GKT7+$)mVsrJJ?pHC^;s6}*d+
ziX9|g>JoDsGc3g|Q2n;O<c<~AjpO!sW@|Kk+9E!A*sGAPePvm7>Tn=ftz?*7%2!+A
zz~O;RUr1ZMeRpf$l&Lr+bLTA*GG-_9bx7K;RX5EpG6v~F#KIkK#j7!!b(`G~0UqwY
z5XmcKTq8>O>7UGBFnAea$Gc;)ewW7Ij6<oCm%g)k%Cdb*2kyqwL-M6sF?&%_n2@Kr
zNb>4ImF|pRBa`ns-?0s%jzk_osAiDKAV#nsaKENwgr8=J?vc_CI%Llm53rR!F*#~-
zxM}%H%aPU^kfEL?0X)$Ewo+)U1CG=mWb0*#C&l0Qn^~4fTbD%|Ej$e&qsZL88g%2C
z@QQE(((X)fpFYjI!>wvf<}is*jx9KdQ?xv0DZQSNcfbYm4B~ZcS{uyj?LS?dK$`S$
zB*wtn7A!+n^CT)aqo@Lw1kLG`Hzrq5-fFw!>zEmRLKfeD$xbdR3DWh&F3V&2**s`T
zpEyJgJqSpA6^M$WaGn<)tzuOR+wn7o)?v3YjoC12cIHG$+IRiGWEIgH{;Hy{jRzY`
zf{oNosRSJ*@wV2V{Z<OfsQYSp#ha_yi#1_i&cYSGNf|HFR<h!9CG0jS7jJ*3wt9_n
z);q};u&5eE&BSdll~{4Ex`5C6fzXdE_t-YjR!5`za>k?bv4K%N@+kXc7y(^yR4eKP
zaVSfqvtb*7CY6`>CJ%?v;>(s#%0bVEoCZ$K;^#3nj|rax`%4>TI1<GS<AXtd8E6_s
z2s&yrP{dd>82y8iYyM^*1t?$J?fJ;v+1vh6m8q{JC(oZ_irJ0i-Ryfgb6iF&cE(CB
z&w0lLGDBOcb0LTN9_nsMy*2G{laA<wpZGr0X?0hzv@4%4N$W5uLug`s+0YzWOmOaQ
z)ak@-KWlNf<K}&8FH4Xt&i2`by|amTI2>EMTL1AZI`d|1Y+5~5;;AtI4UNn#raL2_
zf1)kcd)Jn<!9&thOXVuVbkMI6tZ4=4;%OS@Y-wS3gA?FJTAzjA#?0Y!f0nH~Os0>u
z?S>Ql38Z>aSRS=)XT>+v=GUb4QUVkQ9<g#Rb>zQPCeNgU(>GQn9YQCGn0?5!v87g~
zX_6H8|Ci5l>Yv#8FP{~6(roOm(>m*^ZNC8mOD6QE)yN_*s!rM<R`YX@&GS;2BJ1X4
z3cI@lf^fYS@z#b4eRCqusgkNDo137|HtrlnhMIAPeT%?4KwmV$omYxR@6i<u)8}oQ
z+}fH<v0AmO(ff`*#fr02w`s=WB{lRZ!?*Vwr9*Dy^EIHm&*#QcOSqa+2=JI5imG#5
zB0jVArJZmNg82iWE91#TkAElu$tjL_{{vRWw@Hb=3iauKmv9_kcz2eM2qH?B0MNHT
zRi#346t^mn`cqhyNChnFe!V@GRQx_mipe7Oo$1=U-$B>o6iSOFcLrqYT7E-_MNo%C
zFz|XT#9DC`RwTlp4lp0(wBZ;vZpsJ%5y!nFwHmRw4;sH3dQH+V`~Bf)dhQ5L{CV!)
z3C8(tBv03mL;GCsrkbBu$^zQ=@}?*2Vf*)Aj75w!aW9MY6bFI-R7K_TMSwW2J3541
zckKISde++OJ$@A!KQ|+H#_$~7n0TPz!^?(En_L08yHfK$YJTiXUGc$yZ|C)sk&!qY
z(`{`?*tQi!Gnl0f(3gK-l%Ky_xeuH<oa2(0wrI6)Adbj8i%eo@r+X<>_`qU2I`4>C
z&SPrU8SsoAm*-7~n0ou4C?z(WG0Qo3T_Hz`_)}T6ffl5$toCnRRa8eGV$hB2MFQ_X
z<&MN1-@~-Oex@s;Dt=#vg!=uOk68f?FTOHIfXA4dNPd#4$^ERc*Io$6nhvm^9gC>t
zK2P+S4_~QEkLop=N=A#25>vBR+zgkWEvn3xrA7Nx<X7kIW%)3&G|NqvmI}b7wN(@;
zqn}Z0t~(uS`-C9(RdcqrLrB{JOzXkH#|Q2)vG2D^I8=@7iF@kBWA*~cxl!r~AAm4p
z2BH!%o91e2i7DNeeEIjeGD(wCgp$QRmm5xCc`r3rWiAffUg@KE$n~Bh;F5JiV32tK
zjPfQs>zRali2rZ)?<+a+LK9j{{w2ll+;<Y{2kp74B*qB=HRl>ZQoZ}DTqj%6?{yRd
zA?=4BUn#73Uat1>6)Fhil@ev(9XC1>wSh$9YEY?`<!yS`@D`x(u2j_8Y2!X=(Vakr
z?hFXSd1e?Ld#A-E?7J;{gqtb#li_hh+#lpft-r|PHclYzeh%_u3Nh{-YQ}5szhzf3
z(62+r6>-O;C4P&W^&FI&II_z%OfK0YT_5cHCO~dA{(ky6N2#cE4ui&b4sYFKq=7fL
z&|iy8R+FZF{%ZfWt8`4aUO0QRSZETsvBV}_GbNrWyDX=QOj{|jC&|7#pI|@{l6DC4
zSH0>=Ux*&%k+(|eZCX7Nwhli?PRG*=--H`NtXx5Q`wa;0GZ*7Fdcz|zcBf+c&x_DN
zzK$`*ENd%}1l-E!lKWWLN6nm=(7%}@Qj(Wr!so=Fu>Yhk$zD4zOZj>~qz^xMX)qh_
ze*QFO7BOIeA&CTI8L#RxmiH=`v=4_W<gu(xu_Vv6Y!$f&G^9vV>WO5-@Pi-HJDa@r
zB$7ufdMDN;;OYdi)K;GH%Z5HjLPJvPeQ4<FnCLm*%KL4J_!fqFms>xXgG0&3=ZU(8
zq9JMT+6$8&heDQHbPtH5lQ`JN`3o5a_FWsUU@)~N4vOW^M>aDjw`Bd)YXK#I)LrvL
zKce9B|7F+oi$;6!KuosXBqggZ$y**r<C#|8jzuc)sdAhQCMT;U&(ABv?Cbrm3F|bQ
zo8O3drj=NPJrwT&i9I@Y+t-lntEJ?D)j@+NZvs8Pc-=P)s~r42l_#>ULPDG+nlTug
z;P(4zFwAQ^nO?|bVf2iW{YoD*`EAc#sEX!+q_yfID(KnKFjn%gsgHw%d<Z_)<6RGc
zp*h<Rrd`guE%(?48lWx^;`t^#_j-G>k0$#%R4rYt%EAy%sm%6tJJ$p~-db_@@Z*<?
z+F!XGJ<&r0NY>_j_Y7+=yhm=~?|wQy)3Nbu5cFqop`tJ+{jj5$TT6I2(fZ=pTR)ie
zd_H4Ix8{TvP*AEI`6QX8IN}e0YC87Oe^vIpa}0lw!f89cQ{c7^Rb|}7E;969oXKWD
zecOz4DMZ@H5oB^rA2ilEik^7pG$n(*oX?}6*BC)5YWh-?c++B!$_|4d12M{BY-stg
zSR67y5Ht3ZPMa#jl!o;;qM!amiH3bt+?F2V`<-I@70_`XtsdOR6yu5h+`)w3c!$!B
zyCb;^PiP?kCPUd%C6I#y2HjNlF;d;q8BJ%f!nyJ7|0Ts3_<e41*BJOCtLy!*6gmji
z?7Ms4SJjLS>uE%HE<&K|9LAWbllMI*qt`RJL}kciTDmfFd%HluF-o1E7|U4kd&0NV
zk98YtfVO!t@7UgbZ%_2}OFPu-rH8NtIFPrP^m*DA!Ma5zm+SCkykGTR#<HYwJ33Wx
zs<lvRlXL-Ex)5!QyZOv0&fR@hBYG(fvzhih;@c0?rZ|^au(fYFCB*m|Lq?et`*-Iw
z_~o@zsQhnnLVw4xUlPQ8eE0wi_V?E|Q2rU;ijnNO=J$B3>g!bd5^1=t<dD2(*G%2g
z_;T&a7iM61R_>?C*Bk(&#+u4s`DM3L3yc#T&BrS}8Mbl=AgsJq>*G8qOMkk%IMieX
zKcnQZa7$xDw9KC$%oKm!r%Sp<!!;4m=I54A%(HWTnw6CQ>E14FQo!+F)`mjRk;#4!
zD^u$EZJjwYm>Y2AR)2eL4Vrg4VA0l?q905;9?oQH@_-SXs2tmmQCcG<TM0NuGvD>G
zKzbZ!((``B|CamT`WI3v;uzkLht~tAf9fygAUHTUTp>d^z2{7$!yHevCQY_0@eECY
zcxaslo~OwN9CUmpeYQ5HypBKjUOTT=R|*J9&Ls{V!zxa8o3`F#m9H9ZCh<z8k-kYF
zXpB5)x9ahuMOJ;Z1Z5ECMA*C*#vCWAm*a4~*$Q-i4}_Ci7rYzQ)H%e@g`2mo+8^xd
zg7~Q{tvdlgc)D=|Q0A$C3Gu^YjZ_*uq>g&`Q?-VYL&qmaymqlin@cmzr!+yB#wwmk
zppwjkti2Haz}UHfkF`Md9iqq6{v+iV7DHKC<0s%-EuQ~xi-w%|ImSEcOP=K)(URsH
z7KQ*J;xw@>?ZrUk@KA?gv{s+bDN)W}TywIJlz;5Gp4(W-{@g)@eFL(ak(~O-&D5kt
zGBZu3<Ni>tFK}77@iX-aL?|daVRSh2sDAbH>u$}^1_k%=TY60Hr(@}Fl5_yLQ<rc5
zd#%y@;-Wu=i%7aakRCpIOEQpoKOQ<Z-TQC8Av6U<gd@KZw`Lt*Q&(u}>{zCQTTcjO
zd>*=TW3i!JLbTLfXrrt6<F-S6hUw#z#N-6%4I|;#_Gx-LdyzH80oT_77l^rv<x3-%
zG$VRD|7=C+L2QgGHjODr{4Ql&Z$+n8Ekse!L+qfH8B9K~cD(UuH)m<><YaGosfND3
zTMdf%{2s$$D7d#B2qC%F*UMFTyc;Jrm%PZZrwu)GA-;EVQ3+^dl|FenBOG!lTy!W`
zr`~YOG`z<lY{VE&?itQ7DBS|rlD~pQRsJSZ>O1M#$)<{#CUrY<q<%}}U?9I>;J!ri
z=Kr4wuq8MEYn|@jiGUhd&Qc%Q2Gfaw&u75Xy><@TUGG^r;EwOfTK4<Ut<kbio>o(u
zSq&~o(|DIu=bv0c*=DzNzZ^}<Z;c(6zWWh#!S~tbujQYUf-QgG&~4D_&(u)+y~Sba
zpvvH^GiXT6eDt<R<45p`XbX#cf0hkAxX(o929>L1>((tRF!Np#eHD--vwzwpaZ4iz
zD8k_eHX<vO-A@%wc^1~=ZFQLT38ZkE0}V)9J)JqK|0noQ?L)@5G_bbDc+uy*XkMi^
zI&^n-0@ChfMhV!=mNAO2GZaTv!kumYQ{U9Pn~`YxWz<{W<!3;|-1v+drWK=Jp%@t?
zt@7Glf}__>YoVDan?2lKftrMT9%pC<&vkLB^RK8@j_0!B%`Xq%J=^<yM0%Wo8O`y|
z>8&KEn)&wydt5J(&W;ZT=TEjoOg7V_T;N9?$RwM<Kp9&TH!N;2OTcX3j?e804NqaM
zm+^2cOHa$C0v~y4xz+pT<Dx)IF14=8lF#jHk)L|pid0rS9$3$Wtyw<iJU2d8t+|6`
z{I34M#0*t9+mvWJWStANvWBa@pZ0h6Q&#|5MKWqS{P@;1d0UEIpp<Vr?MmChV`MA&
z^Nw}<4Mg#EdT)0ha3@uvC>LJE@d26OaL)izTU|}XT%{FzQJpbz@^(4S#QXW@klpz?
z{gHsv8x@TcKUoQ0FAT3ayD(<wi3Bb(DdQwutTLF)wYY6qiiI2}n9+}H=C>{-dIzyF
zLhfQ!l0(kvLP3#f8^a%P|9y>{F&fH1T9mZG3H@*O3Ws90P(>kuNc4?<OR6kE(Xn&$
zzW;9Sx)+&Nq0_5FW=8y{ZW2?8k++u%7Lbq9^pPNj^DfHq^s49JMXg?q;y&cpC9|DN
zveO5f#_86By2$R{_|@gDjQ{jxP;1RsG^yAc1Itw=r)I{IJ!rCI3{^Ct{Ko;e2RJK1
zALM$GwM?hO*)!l4y7~UzUa&$Muknrl$jx!FL`Hu}PPz#^QX<vV>S5wQ6p6K^Aqhos
z>AXjFQ!naWb(?i58?e#va3aiA4EO%5Qn<93HsZ-jzFlQxXdI$TG3~IopqPBg<X0EH
z*}b?97w=G)NsL!8%a`BNkwN{U^Rtol^Z(379h7puT*$zbNw!+G#Jt>qM%wCN!>_v5
z=_%C0t!=rc>(Tehf_jV7?%enh$g*en=Lk@e7b%OUy>B|p_pFpa_OM>pXx_*ofP{f}
z&bDOFwHyHl2QfSPNU*}Ibo=ot2F;W}pn~t`?{`lQxEENAq6gDWdRV|!%kvXJbJkjp
zucCg*vI9Kg5T?{@><V9-M$Lu8uwmHx)n@p?TQe=h2Oftjq^EnJplb|%Bn3sqjFoc6
z4o0Jo7H`D)%5rNl#u-#Q!)im2<E%OL`SVoF+M??o9|uvpf2@NOLWxpWEW$tgRZ?G}
z#O$e11u$@8{<4Bh<vbZ|zDt;`XB&DjEug9cP%NHR#JZdi3N9nNl@aAo?XO!dUf+rE
z+;|O%*U;|TXkzRdCje=FtO&%g8XN|vJ&5bF;&NXwj*q?~%%Qw8b}y@q2NlA79T$-a
zkiH-H8FwkvuC<uP+zv(cyWF2Le(W7Fd%l1;;=`#nUM#dD<##)!#;Be>mK6FR<sUKL
zN6&iRY|3oE<FHG%YbzG^UVLp|z^#f`(talc)Ly)k`AAtoK9=8&wYo15^7E{X)qC&u
zzFo802J-0m7=3E|6WzZB@0c%#8{lJA|Ei|8lNFe?YhW0(&k+o7QBOA2`hj5t9nR#(
zo=QAd!}al~qM1yd+wpX^E5RB)3VAO=V>nKjPgT|a`?o=aiptFfhI<IEZ*@cJ(wcK@
z+n=PZZ-~5F+%9%{eyS9VTg}tuZ9h%GTp~gC@5i8O5wNbW=~3AU-VE?wVl?MBQ(g&o
z>S`Hla3j?|oIn+E#|@^SNA6o!hP~7Av|WW6I?v4&Yn5JjTCl@+YD+6hy?p~u4%8!&
zzU3a6|GQnh2j?lV#Q7Z77OMNtks~1}Ke)|4qldlkH?88YWhM`>?Sv!0c2G(9#$x)^
z-jKJQfasyUdH2v$yt))X<5{9+)0u=#(+rX2cp5^3acC}a6<-C9S65ry0tD5MI^)=3
zDvYMyVbEc>eqAV#mo@oYNks1a0=`FzQoU5a9SId}zQ!EA$!7u1jYTV4@)n=^<Rlta
zl13B_0vOcs-4dDE3FMU8i2jXc$~0RcjR&|Zk)c`6X}c=Md8gOtdh!;Utvt9F6J8=F
z>GoH2zvJ3>p_?HM&9-~;UdE+})J8sLuUY)9{_ov|bCz>uS*#RhCv)XQ7BfYGW2=<Z
z97W{+9&?ZqgZh<~5lr~)t@xjp43eI1^dJl;bEEsn;YY1Y3R9;Kp=uo~1Hh`HqP~tP
z86F5T?Mc3PmcscAkobVDOJGF}$T?{O@VazZH_!?6P1G{C=Gz#a_MNst_uO^6w2l%`
zd=vPHD=HGs+T`&=pOuf3kAolxP}_iS=~U~|?cEqIfB*Ji)X|6A_G_NF#WK8Gpz$1p
zk#Qm+zx$Q&E@Qh`GMo9^;z2tIl=(VqJjw?aXb;e_WQb?udJ+ZM3Cx=S3hSq_RW)D9
z?`#t;4ZKirB;R<U{F|HcPrfuX69qLpG!S_20o)rbiCv|L{nk<Z4<CKh7||(*PAL7L
zV!&W~tf1)8p1zRwv&_4v3tr>taeOCyv&L=H7$Z9n;CNg(Qhzq_L1VoNzaW0If>K-S
zSeG|>aLAlwWbCtu_FR39ahbu|WLx7?a_C%rabRtZH@<eN^Yj6Wz^<!xP9q0M7a2hX
zI!`Vcd(M++bBlrz<I6~tzSq6ZrER6>t-V|-7izgdy9DTDWuioj^;an7(1$Ef1PFg$
zVq#d>mk*Rt6qG=3t`wC0$Ul!zQBN-~f9+2d0FI9l>#F3xlLji+nI%NZxnu?kdOyIr
zxVYf+IKsZ%*;J}sD|LHKL{ps&#lwQ6jUUw5lHb9MHM?7{`fV!w!akKUVUkPn!~Wfq
zxf3o?z#zJ6Q%Xuo7n#e+a<3XpDvX-shkae84{rWlUCK>%$!+vr^9@9IYz{_VE86on
zh2@EdMgYt3?mPDkJ~w|h^U1=)nVz0E;I3)7ZPQwzx?UgQAQ1_KAI_C;b_F7QaduX}
z=yv!;l&<*hvGH0bXCx<qv2YFhcyayi(qU|u?#@v(9I&LiYCG^Uy@`2iuN<6`!k!bZ
zM`k|1*N@70Hyi2RQd6a!+~^RwiyX#LfAF=9kB<){4{5DHJQPEC<5x&+E^!nL68_80
ztY0qolxl3(2nl$dzeYsJM{kn{{GP*?4?(ooN_)>zHF@;HLMSLuP_KoF(a19C?y%@O
zelQzHlZZmU#pgnykWC2x73{e-9)wI32?DMKM3KI%!7WPknr}slgFeO&J~tNiGTo45
zcFSyya{c}o3fXyfYWDwc5h0*{EvdqG#FYH=%pU6H`Z_ptO6t`Oa1RZS7i;IMO{1tU
zM3M61{?{ELl$+PE6;98?NvxgCW3HxBNi3f~w0XK6E!G~w)G)}y{?mjZ4tOv+y^_jr
zUjH7=Iv%QFQ#dM_1hu%p)Y;K$D_#nRwaC-MZF{8z4dw57c;JGHA(!@l4GG=R*9U`$
z$Dx{Vr&Vpz1F>k`+dL`;_2=J#dxHZTEb?rc$zTotyJhw8j_x4jiE;x$E*_qZ*-{-Y
z=OgTw+lG%8s&&o^Kb!scz1dKq8ncPG9Np_}ChJAE5`%VqdUa$xj<-=nf_1Y-Lk)l1
zAx<SFhJXCR!#w$SU9w5m4-dl_^qPo`d&6CC&nf1OHXm*;di>19Ql$3F|C<CYD49ft
zh*F*Uk85pS4Gz1J*;`;AKrGX5wWtdx|97@flEEVpWSfX<`rWRdRPZZ@xAitFxG#70
zm1tJRDTABBsi>|F`)c;kK=OP(#2AbwPPpxjzcCp@jMDb9{u#VPNvx*5%Z(0po4tr&
zp_lD?wo#~@2Y-FCa__)PS^AF=%#ab=`l$={un7Gbl@&^Z<31WV!a-<gXyhuSX>%jv
zbK2lrSy^@E3;$$Cfq1@Odp=$Bd43Zbiw^!(FduBct<cSn!%qV-0MMF`2H$<H#f(US
zLRvJdslv;p&_n{ButaHb`u~hojf5E9t2kK!C;ZRguVG(qk$kz{@~3hh2IJF63MaGU
z-iL|{A=E#8V<COHL9oVZ0WFcy0GiQ;QKwGsSQU@m0^{M~fs~A##pJJDz>-0UWf4hq
z$khI6wU*@NTBtZi1HzeN4UIuhTU%Q&t${h)?5nTK8g%_L>$VcU&$KSbi-9Z6F0?X)
zFfcHVd*iPO2?^J93otOC{;@t$A&fg*H2Z&^#ljvfH=>J&y%WFK8BwaY(W?;m-W^TD
zG-w-uK%@EV=Y#db;Mik#yvWYPu9zhfIE_>v0D}zA^Vb}RVLe$Pvd5oI_Hr}-I8-Yv
z6X(S@M7(HDTm8$U1=fzL3bkVOU_$<;D<oZsKmPNT1>#7@WWK_v!Rwymy;@N&{=mia
zVJIf;DR{6SY~+j3|Mu8Y5<G}mfo>LoKl2O)>v?x6efDHBl8iZgKNR`2ys|QpDGVtV
zLK=lg;>$8z{RFl=7Am92DmHLq22fj^#1Eqpjp;q1n6fk%<@&8f7!Mb_V?nL0tsBcA
z5QtMlZsW~AP7C|hw<QV2UsU1GocD)x+=GD3E+@KNyu1OJZ&eBuUU(UXfdFlUSQimQ
zOV~#H^Amy(K`bF0gntLJu#h*TH|(&{B<yoJIMCsEfKGZ6=ykrFZHT{I7_Hajs8r)@
zG+C(n{_mWt2l!H7mYTf(O#FM6h5Aul@2^XhS9!sKBE#pY^L%@#(&yPbj$VgjD_$Dz
zpB|$^G}>=>f#c|x?Lo|6VOYov3=D-T1xS&E{J7cf<kf$dO^oox^4xZ`CH}MxMSS{j
z>kv-JuWfy0y6XuJd0lnZOL5(8zH(?Nf0n|J$X}>P^(V#*!NpM`ofevW3WvBK#G4BD
zD+?xq^x%JC0vBu<bJ5W`f7}NS<>d}#Xk@~-U~|=N^VHJXEZ3<=ay#D&fyzlF`4bxf
zcz#@V8`P4QbL9q-?cR?}#=UQ^GWg9W$y;2{NIhNj-KqWoBd8ZJLiv&R#|77SNxuMB
z`N5{AVz(ved4p1*n9)0#C;Od7B|l?{*<26h&v4pAd<iDz3Y8#R!fd*~As+ZFjgBJm
zyI=Z7MU83PQi%NhCMhy_G*NsG^kkU-_bB))>~fP6!tqkQWCER59OR5_B7@OrD)-<b
zBU%@U=(|5Z#D{p{?_j%LKRJmso5=Qa10Bv32gOn;N;FhT1^oj%_oyg=PatDz_CJFI
zAo;xzQZTF-QV4<3W*q?V8;GSMb2^v?l$B4C{ImABuXS{E!tpp2wDz#xtHOe1usEPA
zuyfk&`4MC=jwyxo_q!GF;KjGaoT3i+GrYI34ULUiVCqb*oD0kHss9xtBV#x|H`+i9
z1$yw`VA27P5{$)2*yMa%^t74eA9Q<_mYVv4OD&$alkgJMqJItLwJaDZ*3fuj{y4lC
z3RENk@3+Bt+8nSfG%V%^$F>)u07mj~VzOwPKkM%=B!yy_{<v6cx!M1nOtaPk^MwjI
z?oVbZW(a`cNkOV8rsW?@@_S)=P|0-rMSo&Yc*kn1hf>uiX*RR*S2IOwkPbU8z*Pdf
zjbF^8?|1(}|2)`+rlyZq2QyqA*K`8j55f4{_L)jK5(mpNL3G_{e^!zHW%+AV7$LF#
zTmF7ntA#3Yt&QsLC9qF-O--R&FV%^MV>6SUSL6c!ws1-STM;Ora(W8klg{V%q1r?~
zj#>#Kh10h8M`wb-?{@(_KY`kTg<7Hi%z>*gIOYL9Q|1!=d4vL6(al@eW?bAX8~~kV
zI+otxe)S#<gZ_m|InYYlF@J4|xQaNMSQz3e^~=SMe2wKC64<&`+MG;AQ$t>y>K*o{
z)W1MSdWujwImFVcrcnA;=D&s3Mp-Yt0Go8r-b7BkgO2(Rt45pDmZ&dx3^G1fCiDpa
z@LGRs;0JRKkEvw8T<fpY!54OCm<9$~U#HQ-C|pd3qw1@t&H$L1vNo_?k*F1^3KqGU
zOa65!5je2t2BAeONBmjIMc5V(5E3{l_q`BF8Wqazt%UlKFfn~7@7ryxumCJx#IRGT
zY+@7uDwsvU3I+l`zS7s%e_@i~MaX{uo<<6;K2nkX4G*Z0ubBa%*carj{{dhHDHI`}
ztDzkmqe0uRo3l;8`ns=Pv$N6T-PL@nNBu@et>y=`QbIpKcMzG$@DGECSTCx#gnT!}
zt|2cyKNJB6^%6?~pII|}0)m&D<e3N*!2#ey<wx;2S`DP7dRwgY#|(Cw$+CY)EI8Ie
zRFD_{v58jLo4Y&G&x3J;??)RL1O)}(67XPLUtfcbuta&ocDVud_I#TaY^mQd{CXqs
zy5j;FrjFRaeioBW+Wqb&7@gyxxb56YhO;SUi|sAdm)mD9wYWLF#by$sxdCH&SiS99
z@=+I<Vk%_{y%xVUVF?iX$F_cufteyxkS6Xw0ic3q4Xi*Wvc*Ngq5lO6;yW1o9z>3}
zwxEcJi0}oy@o6>6#AZvh-tY*6$J8?4Jr`0rzp&7`u&}_TJqotHjzKWDdqtZv{)HtK
zY&JPY-QQj=hJLA;px3GMeb)H)76-yxVnP<o6anC1h|6xV4({O{Iow~m*MTeoHhV!>
zGjoamEZLqcljeem-yO=?eu%~SNYn9T83%{uQwLZY**!hD(P>r)VN`#=SBA5{Y{AB$
zQb_gne7QjWaJCe3l;VFBI>Pq>S`+l}bTm;}o(={qKZ3m~mJEXOdW#u&FiGnLN7NS@
zdNo{SIz}iyWjq@5Y#JrHi1mAJN^oA=g<6hO(;ELrNno*H_u_9e1iaw)&@a+mj^<Ip
z)*~8<K?AV?Zc$Uj7n2UmpL@V$L5*W)+LUlh^{2AddrI>c521MRka)RIP&mC{2zWU(
z)zdf@=ZD3h3%kA$g-QDk+=<9m4-lz-i!16Uu}S$qF1<lU{O%6b2kA4^Z*;)+<8?Wq
zP9~@uo};~J9BS#|mrrKP9!BoFz#?q!4HE<-E8L!Toz;R^5|Qt>7weULv0k6Z?kU(<
zO8HPfNSOCwkKkp6qP}R1vPv6+_XZPcKCe4AJ|~^KKVGEC4i0|x2mA4zaQ{gkIVjDf
zF(hkwnu|sQFq7&EMkRxwEA=rk$U6=W6k4jYj%WY0++f#f^a9Gkx-aZ|D;apov<QY^
zDCHeXFL%%h0a$;&P{ieSAA_)joWIQBD+z>1Bl@>>nVJ8_`5ns-9P3CNc7{#jZ{_N{
zb!HZr87{^X81#6~{VB+hnISw^$Ym4cXf#|qslon0QYH%jV&%WY!J{QTSr-45esPt+
zWF!)%7xj)l`Obg`>}E%})xmM(e|%ysN~i`{E@JV2Q+<SL2%`(84d3ePIc?V3O7*e9
z@h9<>Y}WH5&?gwJr(%B(%=7{T<Pq?Q8s}cJ1AK|K7_+<eaZ&$Ya-hxPzvO_s*(Dj5
zHB>wmH62|ZUcvs-WKcw3ks>SB>|AN!RUfl)FZFIceqYz1$4=rl*2mShbS}G(m^+(b
z=byP6Y1y-y09%{8nnYV0@#)S89zMUjD)+XrtWa+>(4$wQdAD|X5$N}>d@Ls(Ko-W)
zzPN1ebjsXdr%iBIh0lRyTcj_08*w-@n$#82$K`hZiT-_mPNP7eM9WVexd^qP0)oaV
zZB3soZkBXeYu`mai{0U~VUYYwSZnxVRFPwddMNH&e$Q$;+A%UBuLX5sL-iD~eHbxy
z!!vnecdR{8VD5G)d8|ECCd}`yt!>t0$DhAgiBWE=gAjDdNt|o_?Yg1%y4IaO?E<^D
zajJJ{7mToK4#UrIo*7^fh6`q`%Cmq}N;D4GQ{Bq34ACH@pB^VYVOXIyLAx<_yN(Hc
zQAEYuV6T_zcrXWLbF3|XzpykqP9E>G(<CZK4pL9H^YFBx(<X}8taaG*Ha}ZrE<8_V
z<Z|;C;}yOlD-^&k(BVOjVDlo#@6t~7*44%*(yCd*p-<n`Z}*NMmyRj4T<GZ$=hykt
z@_!T-0Qb6b99glN#0wS7d4*t4->u$W>=^Y<OS_(cFmCZ<Iyk&BTHHW{(wn`niZ|az
zeeCdYJil!WCmenY7`x*U%aTjA@!aaI;X-n7D-Ca~5NOY1Ni|7L++?ea>vi*&nFA>n
zPu*avre{uy<`8w7lNIs_8yV58aLKPF2n-v~WWZ3Bl<OcIM)ld=b#mBQkq2WU+8Lh?
zm0fQbnuU$s;F`Ey-$8!t4MtkSdIr?g6Kzn)g(Iak1L|i>$yz_2gZVBtuL&E6_2O52
z9D?sNH1q;q555@EZJN0kLz)=rHDl*QX_ky8_+wm_j8@wEqutqK1wB8Z?l!7AL#7VK
z2hAlN&hihoS$57XuabC@i-%P8BU$Gp%zmX5b*nztcrGffG15my@#Wk#PH%7t_BpI1
z^Icw0<taPrUN%KsXxL<G*IMjN)o0kx7H)qx{w3*~)~S%p1{a1!?=OFz`(+>raUhnm
z>v>OcHFF&1_B}J*!NpZqHqG(v>fAHZUOD?!^QoJXLD*bQo(8n}*&2GE(+`V2<(Ap9
zvk6tf@C&nD66-0|l{NH&+VUBrlhn=IoDl+$K}vg}l3OM8x~bR`f$QBcQNfr&S8Gtz
zc#D>UBX9!ps6mbV5@r2OwV>&6$Fdgd>Hee6mV-=W+*v<E^t-WaiN`nPYu>*OK^cNT
zfvyXH!}Q=T$|-N@W7K&$&r0B>HJ3edZN*c&NZ6*MSBn6?wz+Op8@~W&LWRQawZBEc
z1$I=%<|SGxkN=C)fAC^bsCTqU&9`0RlJ_A@tx!))(BL&~so@f+G8;pGXg=qLw-t=;
zgc=!qOeC^fyg<$V6&$#>=q%KykcJ}Y?Wv~F_J0Yah5sdx#(ZlS!Ny66I$voF!&74(
zC?}nhr6-KHRj+8pk^u(z+29Uypy>~<qu0wsZQaA%7fOxxpDpq8dzdRBvNXmBHJeQk
zn>fotO=e`tlpY%14S~5-2P~h*8N4h9uxEN(Cr;cPl-@dXf-Ow-k*tGPNg`(K`?uPn
zOwkMIFLH?-ACaZpiHg9&=QvCdtF)uaDP_F<`PJpY46W*6R7rd(Sk3A2)QJigd@?iW
zLghY+kx-}pwEP|04?#eYY1bDk(v`0_*PQhO7&D~8zrB4?VqFMea6x8dC3$>Nm#?5G
z6;pjfj3o?Z<V`W`60m|^8Iq&yqi`KuVd%}eWm*EC_ErjPb8!UkiKw+&u<d#BfBh(v
zM-?m^FAz%}K<@EtbH8}^SisQp)3UzOx=&$#aq-*7W96AB5#Q396gHc;ic!Uye4pc}
z&wVgXQELd)kLve0P@JzLHCsNxee4a#5tcl^Xy*TPDnyqg2veQBEvg^1CD<-H2`Riz
zMYg{42zC6#6LKf7W3;u$EZE37+47=noL?B;LisD4_Lsc#4<-`~cMwtVML;N-xThN&
zg6S05$}Jo$pW79o+HR_7>76{Y(O^x+cmUDT78vR=OwCJK0xLeLMiDJF)R@G>%IP<u
zlI@oEg8$az_5~^h00ORcFDz)?j=Fgrwrm_k8j41rJV8OZq88eu*l;;Lv3PoULEfzR
zh-i>n#X{a){@Oc>UFU8j+~|u$HR>ZFz=9IN^Gmc9%v_A~>Q8T{R)1)vtSL^s`#(0r
z0RatU?;K^o`JaA(<%P;uObgXg>DwHPjsdd?-XD28u-{(Yhs{q2AX6%&A+5H49N7Q@
zqdtnm8>`lvWDnQxggkW`AFk%|K-ZQPr~bUZy?tDt&1<w3_vN0#W48#_)tcNajcTU8
zXvgfA(v0N3n;Xr@PpzN*c~QKv7WI|M^CJ5B#*2xgO#hqC6mBrfbBa=&|7@sGi{?aS
zsi}F1ZtJ@M=*#+9qrsPaCou#un7+a?Nl%Ik6SELk*sckL<6iS}JS3opY=}j@6|Fm9
zu41LB=Y1NKPF6#{gh6BwKH6&#YF6)!B8=qsxPHh~_pxHNQqtO;(=JpZWo%+AxiiK~
zoGl^rv&o$XPJ$(2?~OJ(+%qhb!4u@#(v8RHr)^^P2hh#E$Ibo}6~2HSwD;ED)$x)G
zm%|gOeyjK6Bc|jeuB%4STK7FRhoc#}IPq5ZYJ~tbK+3-n<h!wy5cw4uQiZqnp0pn^
zVV8Zbaj|R#YeOr)1_}QPB~&}s*Q#_SR4>;jAZ#M{`oDy#?}rnyF6l%P1PB1)Yj3}i
zbQ*qW4C${dN!`ejyVtVFc5Bb{+89*`w$Xw2Pr8jDh`fCKZo%X8&Y8|C(M=siedF5!
zUNeJXLg{|oN7JEI%c#n|LS%*;0ZEAzS2G<p0vXdPPd)F4%M4uGau$Wc1lG?TC_flX
zPxFhp+zxR*ibeP<C&gs7$y$7pO%vL_b$FJyYP=yP7KG6~LNju$mGj%EK|fiF_KLE*
zT|^X%ve82@u<XA*?k9n?Gn?p#m5i!zdzN1Ix`eVmRGrR0NQpAuDHjItX?<W*oV6_@
z<HLBkIYZaQyY6^Z6<)4a_eCgiH9E%rCRJ_T?&hu0HHOn5rZ6a<eckMSyA7^i&ad4C
zzw2mb=}XTcM+~KmuLZ^VCde5Pg;!m!#pC&PKmlV<`lijP*9}Ibxk|Q1b+X^$)Lc*K
z16nC`&RhqH*?6X(fOHq&?aP(#TO+YB1ukXc2XQoA@<0iNq=@V?H>DE8Hm3d9=L;;U
z@3q7n^hptHY$q+-7Dm7(3hGor#UC|Ti}8ef&hQaDm;QZ~dcoo#>Y-^j*$Ve5j#&At
z{dwFO+Bl>*^&Bl2K6eZ4vj&G5kzZkWRK(BiL6KSR)z_NMhCJ`%=x|ZWk(YgLu-yh{
zyFy~A09#wRiZK<a5as4))H&6VCY+-Y4+?<YHr9juRBuR#9L^@A+wXV{TqkqHvTL4~
z5JSiwnW76NQfR>4sf;I&0W1WNLAM*Qpg)d>WHE8Te{x~+2R;n&D7k5qv@Z3+Z>`wn
zi=A9#R_z7p=WNDyJnhcap(x|-eVqp8#<tLnDhH6h8W;biPrx#g^hBMiNnC%7HtqS$
z*~_b5YH!%eJXS@(G`?LX5%zy7qzmnZ70Ej0*}TMwcMvavaMBFe9nX@%+Xgx?F5%k7
zD`A$(?|_}B3JVw^IFH@n@m_~QHx*D^GwglRaecbZ37ICs+g6ZyP?uefrqh2bSDJ!2
zO0$u@U|_>_V6NTxr4JaP9d1^$N5@Cl*0mJnNAKgK<4y=>L0NX*I+x|^;l^8Abvs8Y
zu2-W>vt+7_DhF)Crg*0`-#c5#_LOmAJ2^&uG{KT5>J4SM&Sqj^CnSvFr-E;&3Q1$*
zM?CKDe5{x8-cyfgHol=^&_#%E@|K{e2MTs=_*`eI*MDzt9@3y$xq93aun|bGU5J8?
z4>i69rhPexfmuX;R>`0sZrS~Tr~N2^X=mKmH7rapUZQO-XHfa(V)j_s%{@FLDsWoe
zF7fDxU=6}#o=oIUKz_&lNw+W0<?DO^_DmzHHxV1k$Gmb$%7;M7aO}wI5xlRop9ON*
zte?>Ec{i^`UpZ9Ocl-GykeI4ZA+!kQqIX(t^$%44;Ibsqc0P;>ipqe>ZWu_TXFX6W
z$Jj&9z%WQGcO$hvD91W4N?^pKwq3D^|1x!t<occIn+?f*XHK8ld>fU3=Z)E!w?Yyr
z4o|JcSJdtx7zeiRoJ%c@4&Q^q;l+|{z2=0bviB>qRR#<Zhj-6p<(Aywo@`fPXnizI
z>_Kj!@2ZiBwH`EothS&`7RUwnvVs8pwfB@MX*DAR==J(S-6EsdDuRMG6B}ajXS|)f
zL047#;haxATYkt*yVu;$`km|S1xhTk@8xWiQ9?F<iA`zU+g<M%drb%#Ilt5EV5Tt{
zVSB_!r>;)@znsY+Y`;L$2WZuoT>!smGA7H#Pc^iLt7wI{2D@<|8#J^`)!d5nY;FQY
zE>gCXcO8)3Hse67!<MP&_0pX@>lw6Mo10Pd6FHgo&Ulx!8R>P1l065kE|t53WEIMe
z&{`6TWw8--T{*_9Mz%*S55eb>;!ag%Cu*~B)yS(|;23*)zCxAfyCRgZ(|evy2q9?$
z>M#M_2TO!zvn1{3RW?8>Q-ua1f2!m+*+sN!JN3nwA&RE57I|Bm47V>WJp68-K)oqP
zRvKnFuTBd!mHRWF+ygu9BxPi2@`5Kofb#h;(4$5)vqA-p%}zdSBLlPbaCD`6vDyW2
zaxVf<4!F}5G{EKYiEP6w%ak<l^X>UQ2NcX+ePI2BnGPKx^>bK#m#j{&uNH<85XZ-`
zh#?&WuO)B1_BC!ge1=(Bis<4zKrdz<TbqJRcalQctLE!f4i=EfA_zu8Z47EP!NR6R
z(ZK7&MGxkAOJq?xP9rRBuTR)L<g$;#EW0xtbCZclt{jSNEnaH51KTs+=vp^Z7aw_*
zFWm+D6np0!v~0-Rfr`MLFK1^<)M6Kgbm6;`n=YrFP=S32+L>H#L&z#-MwtqP0ZvhC
zhsYL@Cwv~oCC~gs?Gp9Mk!C^{Si0t$$q*{0a^hj{=>2ubmv?zRZ<szthX{vwE{t5)
zF8E6KfA4&q?s7sCc9e=iIGIJVQ@#{(95hOOc%0>^U-!XAt?weSoY`;@?MTo~Sj$0!
zCLI}4XHK-nOi8V&xsN^%f7RyzB8EyKn6$Cl(5z1)7?TzHo%MV~_bx*>yQIXhK{NW8
zLF?;8qPd(N4Fsgv#0SW_ai6n0#KC%k@HB_FGC7TwF3~a127K2ZWd}G5d2o~Ya*^J#
zSew)0$aDlNkUv$(1IJn=&A46R-(w90Jr=A06d!l0fYQ=E3CGq!qa7uk1knr3r3WA|
za@kEEhn_Q6d_45FqLKyKv+YRxJO*gBy9oK=C#kxrSf{+1qXs+~3(kOK#Py>216q}e
zwx=94S}J@`?@!MT-!vVrwrSRVCQ69d55&!bFAPoXFY;dMoMKx=_KR~$y#c~;a^+Tb
zCI|*Wu?=3-8a^sGnyX?qN2Q7O-Nnze*nNZGX8kzir>-d`^CY;Z%|sz?27=`Me^h+)
zcO6<6ZDS{mPHfwa(I#n}G`4Nqw#^gUwyo1NwynlGN#4Epd+&$$59~2E#@utQHRtr$
z=<^qrPNg@cj~;$E`ERV=V!ga~J5WRIJ;<N3c8?!32jyL80fD2_vm3hIi6*wMUsa&X
zF*F;k#l%<eqoUAazL@yi<wkU(vf1vaoPDuOHmolnIP8^+p{i5cjUYg0m0d9l4V9Q5
zp{*yo-g8-qQL6)0LID^WtgD55xruiujrBcs)OG-H@BE<^`qa%|ED<j9-X$uZe%^K(
zM|`Pm@X^RMo=8P8ihLHDdfsO6@;c+t<snXYe6Cg%VfJ8E*+_EINy2J{;daFzGaQLg
z=!YWO$8#e@R}B}R8AUe)N6K(%RK4EGAPr#72k`N{-?}O^pYp;-m&JxQ1?R5%IdAQZ
z_rS&1TP*Ua&}j#~5;uO%_Ho+{(Q&&gRexSrm3nXo`$-IqFf6lqT&PSF^J95BnY`2c
zSA_*KMU2g6l-=5B+|w8hvtG#Vi|+t*--fu{Q7``a<g55%DENN6CN@Zl%X;5H-t5EK
zS{UdEHRgamocXAiCJo%$Z8|pn1`IhF(ufJ<6HV`ekING{n6eTZJ^HuH93HMXOxk3E
zPG=CwJRyhfE?|H(xUC<IB`-s~rjVvLI&A5?pD_el-h)WiC;NWuXsb4uML&hSOJWY2
zm+nvtMP9MmMbcK^>Nb>@_BQEFL|t*b<v*m`{omE{`d76T61J!z=#N_|815LO8qXh-
z(@)yYDCNE9?t~N(R12CRxSSZY3QCGs1kY~1yo_gg-wd34e2=<_;;cn?y_@qF>T~>3
z{UxvY67uS`=VxLJ-2kG~jMMrCuGv*@OpOU$Z>P28aI9s!+f)0S5hPvCF<1@yG-*UV
z?bwF*^Wae1RPF;#GkKl*W$n*f_Y)?vraNpbBEK=h!*X;zOBNl~&*VRRjyl2EJtb_!
z?b0Cmg~!3x5kt%FE3v5|o`&WeXOTr<PReQZ#g^l4>tO9qJ@5O9EGWg(O&XrtJivZy
z&V!=#L)gM+!Z_Gheet>U8&h#qDy8+yjCoH(#8Cz7OmM^WiYvlQw7J+62A(R$=eBxw
zJ!Tj4#`l))Q<a>jeVmhq3tVYM-7$+TE=}JJq>jsEqmw#j`br+U$2MSJEX-RSdAuk(
z)4KYY8&%G4$4A-wv(4iw&dC3D5E;?ox9S`;9!1Uo|5EzcFZ{GFeJ*%ZOd3>8g}%Yr
z0~q@^Q%@^yH)O4U!cS#2K6w3s#3@C4D4O~zKvxWIDJi~kBQH+3mJ$6iyP~H~8F~_i
zu{MkzDKtKXWS%soqbEIgJaeg+_3V`A&>Re7KDt*;!@TU2$<w4!vh$dea_MNuQj`8s
z1_Ky}o^H!u{O<JCWln-i_-pvhH<enn7LrkG-?2FYu-k1iFF%i<wz@w<oZH^#f*1!f
z|CEr^EubZrd`mavb$CZuW9<8)t#tK7iil(1Xkg!PKuDaF1g-{gxO}YZ&b1*Ivr)$o
zq4X^4!)iyF7m7pZF>1)X|31)PGj&*7C*l1=8E{;rUjFT_SShKE$*OL2DJkWo%QA5&
zC5JRC17d^5OZZx6FtB=3xZtXrqmtyi;f$vzdjIE)MD+eYg*|K(4*}MpX~Yt>X;6m7
z_j`WjCoeiGkLVyv$GnuI9*evn483j*w>{KJdBHGg|GZy3w=#Vl@fEM##63SDD56Lu
z)@&UvKV+0w73s23++64a4w>$JQi-DhDUWU}A@smHl3q8Qxd#hOVQTi@gQzU<H{A^|
zpZ1(+wrRQQ(enDLCRmD$zDhb>SAk>kGup^_*a%N$+1$2b4;W!Mw9dIb4)1`lX`MK0
zZePVFmG6W+zUZO&PoqLgE8LYzkoMa>qB4y~X1ly~Tc$iq*FsTNRJw|(7Fwl@n;~&U
zW)&{2_cjaekn$b+66$-Zb$X3oj;dD!LyIHKI^K603cc_I?OPft6>`ZD=6Y?$&T{n`
z;%?1c!-8v2k|<;1vr14{a+K3;f1ovrDTpGz$)i;D=f3I{iu?a17oE2%ad?cP@siib
zUpxEeR$G`{q2G{kCJ}Qommt_9sMBOldUe6gBL}#s+v-;Je2K>S!mQI*G3!erG%)H*
z^V@KkQHxPBtQY~aUuhj8H3Od2>`kGg$gvM$T%(0)ltr&D<@V@`B#xUZ=Bc1>)MzLr
zP@l^C71uBO&)2;7GR9GR(b8Q)!?`Z;n!1!1op$&+)Y!8%#(d&Z^SjQwFLOoW%1fOB
z?ho`>D3VW|w$DHVWG1OtjF)=L{hu<Ll$SK*_3qbt?FZ9w#gQn{W23r`Zj#|L#S7cc
zf6Jx$9PW#isOz?0Up{)g(X^i~!gFn{E(CoDUH#4>%UmqkY!66HoZQ5kn0-@kvyr;=
zJHoa^#|&53Vrdmuut8v8k<{3hWEv-1*u!ZQPcwh9+LXL>qJEJPUdb;rao@mh-q<df
z(qy9&x1z@$hx13G11i&p31^(&LDfsfN#tH)eRh&!PZr^RqPzRn?KpgBgx}RZWh{65
z0eCqVbH?K*0bcq&Ua`Sr3cGkr<n@1pzARTE%ojO$+i&$Ytk^U4RZwh?y~X8-1h;PP
z9%#W<3z$bgC2T##TsGAB7iq7vJXKXZdYH?035g~MtD|z3tB$Coi8wCz*clEVWIp$p
zkBM|~je>LnvW47P%^a>pUS@JxuV8METD3?xk4A}-JXly*1fHIhrEQ%Lw|+Y>t!l^5
z`JOt88zwQ?8r2WBJpX07?;?b$t`?L{n>YTJtFG#i^rz#tov2*`e0e!#QhGd#YawLb
zr&Q|%Tgq_S{tR#Q`^OupQY}sKz1~#Z{_MPZC3uA2@1*Yxt6NfWyU+uc70stBH2l=_
zK3IC$|AvE<!OmXDCvYp(xu0260J$VI1iP5=*sm$I6{qA~k0V}bv}v!Gre$RCvWDMN
znQra8osSfIu4$SkD*d6%uJZmp{d$jEOQ&8A1Ffh0O1t>0-8XAt;;&56C2S_rx$tEc
z5Qf`_ETF@xI8S<dT(zU7L15}}1?JfIu0NTMk)zs%({3&K`S?q|f>f^=Mav{#AIE<2
z1fwujn)JK1ty{G1Burv2;P#O83`;2t-Sw8-l;($K6aumThc-XFaL5jy@EVn_wW-Ri
zyw_%<o4*?O|6Cra3##TJk#|mL+UDc3->{SP*A^fVlK8H*Fc&Qv1#NXBNiwzz_ffZq
zs^RN5h7|@Qpl4%b7+lqQhgMWuRiyP*F%j68*W*_Q&}?qkaM5ifi9CM-x6me!>6|9*
zT~^_=`@MHgxJ3RpuO_YjS`Qw@RI8Eh`JdeC=gvll^SC9Da4tfBBZn0Rf#H&*&!?<?
z+;~oA2(A&*Snt-T7}<=$@`zVo6!826-Zqpth8(xF-l6>rwRsu4>I~dwFjrY@x>>1)
zT~(w|r*G{2jKDPczL}FUiZbDqD{Rw399!+Ug3m8A`7C|9*7$ds@YN=;_v~JG(Qf;8
zw({z^p^#4eURP<0g{CskZK2H0gb6el4HG&>qUdW9?G0!xzSf=nm3Ha>sK-rf;i(vy
zI*yceAm%a{ioIznDdihA`i5kB02A=iOt6c?SCPX6^DPnP<*F+FcVT|6mXF4@o5BM>
z>~3>r^#;evu(+GjiMy^Zx?=hCqkL`i7YsQ;mGo>k!f2@poo~)gi!}RnI+mCQ@<oJ|
zOm@qz8)Vgm4GsHy`SsswrxrYvsbWe}_WTo5`L!$V$Prn7u%gIuWTj_t!XE8xd#-YZ
z!nO@ey0zvX>mu0;fx~Y##}TuRiYYah)eB&d_3GKEUOy~O_2Tv6bojM$pk_u;IHHNQ
zh?OPs`1R`lVIzMTKd3&u0b4Eo&c~u^dOm_zr1~g6aybPf_<42F?m+Dg&9AAoAf3kv
zhofn{6c&BGMZCGb**u|8=9VZJ-k556w-pZmkiqMTCLE2HebNv7Rj3YjUh*DYtL6`5
zzN-ZGk3zW1H8O_hzMpPrJIe8LX-^c#wWAq!GpX2g47=f9alG>w!@U1StecuzZU-{<
za;?gy?XKpW3Q>nwJ|1cX)x2un)TSGGHdP2*OB(*JwG%hB6tdgZG7X$fJ7L6*ZYN!?
z%Wa5ExsI($6YEp!qbK&fx~?rcW_6dODVENIX0lEjx#mbKFqoC4)g{H5)$1*=pESw9
zHi`b}qO5aob4aYRKAQS0=wDHNgj2hXf(GoGB=`3LrAaG(+%9SlyVn@&Q0{;iDkmWL
z;ULMGpu2rg(i~LzQAFAIhK@s{RM*k=c1~q7gsZ32sXUj1s97x1#nR`T7r^qD%Reuh
z4bFY!Jm!io>+z!U${bUZw3;W3KQ2z$tR~!&Eiqp1(Eru#(D^iFUczfz;?PuGQ<TkC
z3WO`<DR!Uz3#sl$j_RI>2ck0oqFu>J@L>~m<#udXOC`O`yHpcSp%w3tD~7P1&dAFw
z;MV$^Ee*_zHfNQ#^34*5DzbL+w)Zo9A6I2?U=dA`LGK2wc6a^i$0Dh`hh02f-p6V$
z+~QRFXk`cNST7et>d4v*e|;=~&`O0dnyTva`+Diaq7jsO#@1Y3^=(;3&AQIu6_mvE
zBfnu#9n+uH)2k#d0!x|!tNK4VICRt~m*KEnRR(02^0X=?y%VJ#-w(jOa@6xdm>q$C
zcf=!xvejT|_&;#DQQRt&zuUL?=g`K7DZFBKeV|U4k*I{$b~NoypU#(SHktj>q%h^%
zcK3`BqU0HrzSL87DnA}~(N&`_xvrcoQJrJx1*I5eDU{l{tE+ABzp4J{)|gZM_*2ws
zv>Rwmn?96=b#MPGl5uy{DhC){dC)NQvoXZn+G=JnD|L^+J4HF|){`iy;4wu|dLS!1
zFZh0NX|dX1d$A8gV_?$hZGwT2jOtm!Yh_P7bTcLG&KbXtBxdU%OxVhwZCqo$2D!0V
zDn|aWk=(hC%ZjkC6JrrD2dX#nQ0&?ZJHEZJ{N;6^R#JU@?Cxy5lZ$M#*}FL-aS3lw
z@|G9fd!oUNBv|IJ1o%|7oJLy1Qh#Sk-vq#f*{Fc8Q6kt!Q|@cBFmtB}{7CTTg~zS7
zqf@rRYq!vra35T=i0g{hR{N2NyG4dnxM}Qu%m2{&WvN2h^UAYM$@#YR7?~&KHZT8$
zxi22@BkA{5U?&XWT*l;WSGe4p=_@ldBFSqHJ^LS@^H3baX5RTdPi8mW=hxM|yGF`b
z9GNLZz5E?FPdMnQa{pO6`htO$+wx6x2t2xeXa5AZppByAJC)|~#<x8p0k0c1CeG8l
zOvEmp)$-$8WW4EX2R2ptz5fmRh@lI86%6SZ{?o!d{U>BBjxc{Ii?CGC&<F$p3rz~Z
zDb;uJ$uK+&KA#jw9IaZ!5?I{TDxlhZTadlI2dl}s3!rq@u*xdU<33<a`XE%ozRF7`
zdFW4+rfsj%N#*oUePWeu8n5rPCmB4XC8Cbre~)Qzik-9a3P;#9btuWG+G#WctPy>E
zy|fCO3}5lms*3ZEdh35=gkn$T4WY0|`{Yq8+Qh6MR#e}BwmoMRC!_MvS}e-?F&dT3
zqy4Ye^BCeMCb#R$`#tx^qAJTBcvz5}rMI4AV#%ogx_=L+Jfn&>?ud75w(gI1gBX5{
zw?enxS~Xxi8nw=;<C^=jsP6#FT!dB`8umx*ZK#VdW+WXT&sW<m63q@W0CIloOmiyU
z#Hctuq%#v)%4;49tix7L;#ypmj&r{cdbq)~Ty-6Ek={P9Zj+&dJ1DYtq+pVL?|sQC
zWRk|>=i^JQS67jIY)%+hmK_Q;w=fS!8{sDnwIl4jD$hsR<nX!Hh<@JcsX8$a!dDP#
z)H+W0OFNLl_^K)oSu%Zmhb4vb#K+*bS?lG9^_c54)^OK66%yjPD4Z0^P2ve*uv>4Z
z(d*h*7~b(uvHTBI)2<PvX7>I=)yuFRd}eJ+sr=mZ#NegvUWHVedE>$0x$`D$q4Y@U
zr#n{?ebsJ>Cj6E9CtNi<{s2Vr+d{)b4Ehi4UXOx8KFyIP=rUVh0>Q}Ok2*Q;#b{ps
zGF=gASKGkgJMn0I?7J|xB^(5-Bk_m?c;%OR&I&E*Ih+RN_1WTLb3)kW3hO1dt28K@
zIcG5B-@T#c58K7T@FnJDwkS6xS?yjt+3hCpD~FXi9nGW^6JOu!-hr%bB_XLdUigiN
z2XVLX6aayF=xe}Sath?<$%OOCd<B;u&B&YP(-e;g;(K8&bA;T|!<E>zQbo*n&2jA8
z9vxD9=IZrEj{J|Pl4!b4g1B;>R%oSFN{N?OCdF0_!xR=xXsFIssJnVN^)qx^RE9gv
zmC9{qIeO$=R;au!As+9Kq^M&Q=5GD~_uLa4xJfJ~T$|8dzhqg-*G=2Ml7&l}byxO!
z^%?@U$hSw}#(+T;@oy6-h>DAvo9aQ&H!kL6X!3bI@YpigqP-id!HhoVtt=i823H}w
z5b`^oWI^t~T;6Nw5IYZ0S5Cj0n5K}aMjh}tzL+^Q{+twWKlSIKy39h3J8)(PkKVoy
z%PY&6%3=Ou<ej5vgyEm^+!Ho0;b2hY@|5XPS#<)2(m7`6g)zN$#AH8VVk>wvEGG6L
z@|3~n_ef<gIJ`HsbT;UeghH!r98pe^$hAIJ<OlgihS3N;P<XGgSiT6y?~eM4>SH96
zAuzaTdg!!D?eEdyVKmgca%)#vgAU_N$o!nGpi)Qmk4F1bIAzg>Q+|+4u!FZ{+z0p(
z{HV?y?Ds1C<-6QS1pT8SfOK~4%-`02|J`*qrpuyJ6-hCZ0;!0m`@RPe8jE~K)WEM3
zA_3_S4a!@M4LX9t5yK}tE_#(3o$=hn_$?mj#AQ@sP6Ccv=@CoS_lR&AyPNhvfFm4H
zIo{gCC#+F<ma~qHkS4uY8X&OjVeqGM^?~_urz}x(i93uqQ88%9uoS&g$Pvd0O#|)i
z;ASBXb&G1YOE3r{yIry7tV5vuM!NG~nHL%~qxIaN1W5_bzq+^)cvkf*me)2VOyd?Z
ze85qLx#<T&Ax*{4oEuO?bWVk3p#+t+>3T?%&=H;>E)9)&DgHK<`iL5@G%}yaNC_zI
zJPWxlO=@oLf*i@_Xb;%=RO%F_m5w!(7P4+GCA}UT8)|8$ks<VxrTcGwI?sJV`NZ*6
zg>vbkl(fKAX^7{cHYOL6aw{vBthbiPN!u<G5ekdJ3P!lFw(ni=rMA+q<>F0vMje*E
znC`-KI=Nd65<f^SXf~ROJ(w#vdsWlQ40jU*60D_B*DQiMN1;dT`bt&3wNyv6xA96}
zs!MoUQ6xec?E{3ARJsU_Oi(muH-7=@Oy<oG$hR->k(6aP(2pqjIKpc|*{4k<Jwig|
ztjYp6d&tU&hiwYgLfyZV-Oe2UK8!}AK)E7egE5{M94|@kJt|KU8R_mPrcwTdt@@-e
zI=Y}FXh)QsZmnJqM@XRc4M-2S-jLY6blFf!8nA2AF=NUj=uQ+|*x78&oI+;pVSIJW
z)6lEo;ppi%fEFh}QPM=sqEEm$_}u+{WN-W|MV{(LJ`-1rU}zBWC1G6ovp*W~f>q5K
z3oobj!Pb`FC96DYivh;F%PY0q-|f}J^m6tcmB~;ahEYzc0OZlw7Zi3tn2!D0UA|}W
z)`F3a{Qy<^Q34S1IR4^%W<3b_c<_!LLBFNzZx$5HTKeT`b$S%nU_8>FToyJ0<LpWK
zaZI*z{TXw@Z^Pu4ZxcBfl^4deli8KWG?fnSfe4QSssUbY9%ct!xGa_YI?Q;kiW2?9
zDNQo#q;ggo8tmVdMhg0(tGBvP@gqMdXa@G5nJHy%O(FEB?9a;U?B82;#mwiSGUxdV
z;hET7?3W>P;gHow^7>f8YQ=w;0RKD&hr@{`a|lh*Ki4%~Py0-BP;Bx3V}MvJVSf>E
ztnr*JDc^g?0(2p1e)<(Y|HB-qfl50*Mf8z4Y|NIb6r^#`-HeO)TclZq%4O++%X+;5
z$PfMzz30y}nJ?5WK3pLt_OkP5U5!`lo%oXTgFvlvvG!nK>b_T``4R8!>bd(F?>qKo
z#A1l&*?#hOC?w~Cwk+?jcuG`a-<3ApDSO|97?8o2S|2M1P%@Yw*PJwd3jpVd?AJOo
zdTc!l)(5}*Lvt(ao)c}apGffC4u;Y}_(OFsL|9Dd&iBrG#C@W-&y$_{RSNNF$^#}J
z-Szn;=kYc-wG-Z2Urjs}KnDYOMQU;JX`@x3vE9T28PYOcrHkTROy@_Pf?OP&RA7h~
zOfICUpy>UN3u985EHMjR`edb#bI-@sco8CeeaVpHnQ$=4$w>SWh4WdRf4nC*1d{SQ
z7P%HVV2f=B1iCSK5l&BesopTxhh}tt*Q4uPsuWvXh;)-n0VM+7U{w{n#M+US2^x|&
zkq;ucX$zJW1^n$a2odS)ivQCLf7ODA{y0Xg(JCSSZ$cduDRO@@hpt?<@=ZOa=Kizi
zQBt!#zA_Q-!Xbrdojtv4RB;w?0Be=^B>6MS7r?z&Y2KiKA)&Zbo($4&i_Ffe&|1Oo
zv<z&Xv?{$BbdGuxq74<|k1*wxc{rvnt6`kp7W#}pD_xkVoXwo2nKEZ^IC}8^8Rhq8
zQi`2W*?-KUz4x>{Kg<g9f@;+G3?}^32&Lcp)7;KY?TavJS%dGAKiv!qHJ5XlQZo+-
zp(v>MUO^PmFU@7)7_Jx^c)PmGF!t`drargeu#!?gq$k%s!T!mth|d;_%z~&fn1zGZ
zt4jTOj-aj-&}6K2w3g%g7Z7Go0cQ}ekB+Ul5n#0n8tB_ZG(wo!<tbFf?|AQY3fjmh
zPxf$!G*&dfvTWaEdxr$w&8oe7O(fA1d(c2>Ke6`u9Ed|5-J(_0|L7xAytJ_b>T4c5
zzfb~k0z3>&$aERJBj_#Fh6oRk2L3voi!M{jSQU~tQeIf_teJ<_m-wJlG3jmaQ0ou&
zwR}DTQ+tWdRe#3CpjKwxEsx)9)l$@x$$XS#KkZd~$`ESuq4ILz*Fu!`BB0EzF=uc%
z2N<iB4jOin%sBbHMlnT60CQ*Uo}X}Uc1YC0FQb=Dp#sr<hC5Nqw)-{^#{-u9q#Q7)
zEO_?$>#OxLhDks-k>+wm*p^8wrActt@EovBkJ~p<WcNLKl(gs_XKGp`l3M;xiF~b*
z@0Bp3L#uc8AR}E8uA9cc_dV$ZOnIT;<5p~zc;!UpCgH&=i%dJ|-(#D+4sh)3a4yw~
zz<HXokL>Wl_;mgnAs%g6?bAYuM|hbS$8$S2n^6~M1Eb&l%ddK)Ue&QGorb2B@zh3n
zcU%gJD_)I+AvN~`ibc=T^5X0ATO3Q+!)2wAx`04p-(a0ay%E#vALf7rM23>wi}}6j
zE0kuLa7Lbq8ykte-{!j?O*uT?5X>FOfb(z#Tr4^Ifs<$=uRcmWU%6VHr-h(N8WD##
zU!OO8@rNN7e*Zc02<yC@s6R)`m51ou@`Tem&5(WZ+4#TOgc1fOASBwV|46{_cZZdS
zc?=XvXC@8MH<{hJv5KR!s97WlI>X*uZ7cDQAPw#}uZx?wxQHypoI>r%!4!(emb%$y
zNGJ1rl~YC5vWUZ|{KcvO7A+jel$LkmGDP((%`-HFVPI%OL#{0>_soFmS-!>8TKY<K
zIqn4{dSJhdL%Hc~YR>2ZmMic5knxtir_1{DMEC+}U?_6|rLOY(0_Pgp0M%121F83F
zK@eAOjB1?ifGmrYc;D-cKp6prtgMakR7l&m>(7r-kz11fK~91FwgfaLOnFCD*w<T+
z9TEe!S(hJ-c*-;(n@mH`eXcQ;<e2(G&B+0$&V1>hN{_eI$==<Vo{Xj^#2D;X0aP5V
z`se6d?#vGT8Bx(fR3nLAo5NX^Z=vCK>~m3MR_1+ClJ>;{An9m=Zs%a^yKbKTF-&ga
zA4MjA)JIhWz&_Kx@SMCyhWQuMveC%h+T_``7)+68v5?zkb&0^&=>>tNC!JA#cC9n%
zxO7ci8Mm{I`UGVWJZFerccW2o<BAb8a2V<T8fi8Bv`}Z#ikbfdE?Lrk#PuD?i5RI=
zBW`BRv-)6lza68{b5|vvbg)gRE>k#4>2+6RWS8JvUl0c!aGU-$)95sr^ufHOwYTP=
zxy6H**TD3UgTbvR3G$h3uIb>VI^VpkbQ%-nA~;=BhjX-SPSf0kw)vwngbn+Bno>!O
zKXEYg<!n9~M$maaT@onETWcE1F%Q+y0!~hcjb``teRa6psKCbggvN(+K5*7lyLam8
z)(+%-^|Yyzst|IS;!;h);o+omsnMasMwPv0e_5`EP;Dyg+YV#ERd`r*mXhnfk7qSn
zcXeS?vVz0P|H2Q?45SLxLdNab4CZ+bbQ*5^j69mbSRzwA;h1GC>jqZnqI5quEWVTd
zQ~?#ABzvF-=6>Q|d+e}x8?cv}0g!7`AkHToZ<Um;mZ_<P-bgKeMsuRvm3_~gF0Hx0
zo4bNGHe~R3YoCsW>K-3*K9T5RYqpN53oX#DE0RO!Z|vrllgeM5KtlDgjoJ2v>V_Zb
z;wzndm>A@TyP9u*oMTQ7B&ahI3B8Xp2Cna)$NM5^r0r~Sr=7zKR}8!p!P~`GVS&QG
z&!-!Gy_rKhUxAl>7(BMw9Gf(Zr4J*jE+w~$q{{wKT@dKt;D^z;hihY+9aDRJA{Gn`
zf6^^hNXH-1ZKUG^+MLeQRXyEAFBvlNp47;o>eP9^(SLwD52L@j`<Y|eZsBZ>%~Es%
ztDB=qd)F!)PQ%cMg&x8=7DpC{bcF0KGA7ag;wc&Mulue#)m_uDsA*Vf1Jj2#^p-~#
zf0rn#9)5Knc-1`-utI&6)e8A%StTDXWDDKpc8z}CyPa?`daPNQ1!*Z(qpR)a1OnZj
z3<b?jbUcfki>iPdV{336N@c!S@5PGOUn~roi13sN3<jDJEp5kft2P}nCkIckK-BS6
z*i>HIqJ+%xs|Pcal!d-?+FDX`0>#%H38>P_%sw9!{KeW*d|Ex$KBev^alf>fS^8S}
z#i$HPpBDmma5Vb}Vi3`>Ub_zp$JmEqqq5Dwy8AOUCHa=;QG1`<($00{5}gdJsBd@G
z!QSGz$MZ9S5@i#Zq#vcYQ7Vw%{d%4`2ySx1b$fSTr3<9rtJ3|@&C2#xmrASJ{DoN1
z5Ao=co5qH>&vRKXrPOh_42I4PP$l*2ed+clF&D+%ew}c0jZ+!b@tEmdoA>PUWej(L
zrjxW30e_>Ys8>>i=k!?j8w5HTeEr=f2B*Zuwg}7g?sh!uHCP)ng>r0|u{+^5z)!g)
z6pu-m#m6F=gm4Ur(mAdAN7E~pVuj1^uE>Ml<pl66E?{MW`utvqLEhlPjC+B``YQMM
z+F&KD=|pXyKEPCT*4$TWVoGhZZX(KmTx%$33CWAue8g6=SWU=#ud}Z_XY{cu4YPWD
zQ(E4bhSLe=?owgXF)`)vIQ4cT9qYE*g*lHGo3|JR>|<X!qd%c|Wzx8CbSsO=nMND5
zVkJPJ_zcYjtM+4g(mp;?dvlR_^~`g=f?6X`CnN{-;k$vUX^WXW;z=$`4TSp4d4!E0
z$%$EUKy6(xcLQ`i_q&=DZ=%RTsKlPV=dqow8<+4*(K!`P^`fQ`KjxdSxS!~;b$LA*
ziMK@~_#@f8&p!V>2F!h75<Cc`3a_nlOeL*ePcl9bI`AIIIqA4=5t!}+sR#{OZd|~7
zpz|`AO_nHWq97GA#BKeW4O~Toy8crB!}e|iKw(9$C$?ifxk?>)8)(q2!5R;slAo`I
zJeoJ(4Gl^ZLmP=^oUo7>&F3mh3<89Ekdg9W0G@t1tfdP43#*>!5dhC{W(pi==)j1O
zFUo1GEr`mK_2j~+bdQ>bUY7$ht_KY!GPwack!HiJO6`q@f=Jpr3gxE1tdg1e60li?
z>mTX&L)eCU`oP6?lc33kNAXM5jMQ40iL>5=gQCB9s%j&a`ruoECO<b${k<?X84!a+
zr)&*-rI*=a-|0i+ep`thE$2r){own1@-0Vf6kVv$Dc*$H0_pS+Xg{lcOxbsZ+|O#$
zrjn6=&9(^)c^l7LoO1r4^N05EuLZs+M5DBM4*Q%DlrmO>QMZ#DLCuuG)Psz>zQLb9
zy|W>NREvRp4$aDJ%%e>EF0jBBW}U#bm%)gOS(I@@NQ)Xt{8Y-x*c1|{Eat-Db##OR
zxNN#$+inDAA;(u5`;Z8By1>3$P(z>PuJoH@iz?8$**d7^(X-+3A|PM?@5962J#lWW
zos3RUG$5gS05v0go@-<oABjcfg1R%Ovf0B#`hNeA2y)|E@0_{lB!q&VUw=r{vJ3`<
ziNx<>c0AebFgWO;P-k9`tb~4x9S}gQ9S1&4rwwcc!I=~X)X1?PTD0zuYtsH{j+Zv_
zbDfc5VgN@U%mz`v(^PV*bw62Z5?})Ce<2<|-|S`7Xm7^>nAK_H_x3>B`EcO4+QFE>
z%ene%57Y0no*4jh9}qygL;t8PFhU%$W^OZ!VXT&e&i)4QK4(8^H2qf8<Yf1DTv=23
z4l~qAkDrMUy<B861u?#t0kByFMcD&-NAIy93C$~P<+-VJ6ehsD3&#*leynvLh5l{~
zW4*ZqYFw(6D~>vC)Us909%AmC6jQJGx)S9UZwqwwMmF}LBOHaKtA5WhYh?YAG%R}R
z-7kG8nWd@)aThFxRH=s2C4yrc3lhKrs5v;8lHbe6=?=qUBFZN?3_FFezhVdv8(E3s
zseAaEl>}a0wcQsBf0s%LI(cb}=APFn#FhVQq@A4m@xAH_YB9WK=zOs}WO8^2q|~M*
zE~{T3bJa!rY`vy|UFLf;iw5w`t>&~}sfA<5GAYYb{GGC9O2u{ej^jELEv_t4<W@C1
zu3FhiKgG$TZog5?w@~r=XK-`9RV4O^2rCqA{+0Dp@S1SPE%@vyeXTd79as{D^$Ylh
zMuJ)Q!N*wqYNg5?UR#sxH-Q_J@XhVA9NSkGOTD`&-;xOGN=twQ_#!;EB=`oUcwTR8
z3{byyH;`)H79LeQI3~*n+V|?UOtnd&o9lGHtI#!R7R)GR@+@|#unMHN(ZF2Wi*+Qs
z@4&X~=m8lEG!^-+iU<%(+Yrhn2r}JRES|CeAc@inyybpusK}vQA`m;*j6D^5q2uoT
znZFaX=6n;>s(3N9xgpi-@J|?(jPu1>Ds031#l)Ng_IM#}&|UbccIkXA)E2_m1zZhY
z9e!Z&V1VayCO|S-Di$gKFbB$l<1-;@&wMrUL5ET4>Ti?7oME=A-y;I+<=qt=Dje77
z7jO_R3xC0(E@Uo%WI+E?nwClcG|@Q3t<rUo(=Iqip^4$Wmf9b6+uPjUn2L34bJ(ZL
zRFBKrrlJN^3(TRu`!Se8bd}NZ*4WrjqPQ&1^v9vsnv+EqW|&_;Q_aOD7;FH9j5WVf
zUMwyYr*QkID*KpQu0vZc7zutqz-n76){Sc9Lg5|FU2TlGg(?8i82(R<jRYyCOGHSJ
z@%NAJY52cBO#(<H9OmIIm?GAKPUUix*j!m%@4r#itHz{EypD@~78zCtRGKgi7(Aj;
z9+QxI!0>Y;Vn1hj#TS};YQ;H93Ue(4q4`pa!<XG8I_niDAt<`T)gQu?w;7BttDb2@
zb1<i&9lb2u{@)k)9aD_;<?jlZiY^$_arlFhB*eP*#V7sLQr0nqjwpu`dQSA#&}Sm;
zj$mVhYj$JL(dd?r){LrGf&}Sy$5~Wk80Jw&C}DV57JQ=rGRGwt)wT@n8<`IXn9V3j
z$snC3b|?xW*aK+Rc>7swymnu@J+4DL!(=Zh=Z+T#-@kbm{z?|Im@kb5U>C&C#yvip
zD*;8oCk1Tz%fpzytMd|?!RD^5%7|RPJ>0$|v_gx`bjei~tx@Kt-jGTs;VNMLUh?yg
z&y+>}^!NB`Btf660pRm12k%N)xqNLj?_;KL*5S+4vKYG9Q+J%O9#zK6YhG&$eFv%=
zdG*bm^-Y1JNgu4b{U%x7y@f+*D4>-5C2{X<2Vt_A0Fk6n4IRNBzS&|6S=`A?ESV3M
z%H85V!E@O+(ptDEnIZ{#XmVrft+lZl1muS^^6f=y@q2D4o-Y4?Spb+U!rM3(F5EnW
zPZ<2iuCnQXoBa`0al6+I#aXA9HUFop9;bw<C9tJZrXJ(cn&F4P!dW)vqG^=dj8b0<
zA0yGCufkJV19nddJ35su_8i-#&{bH;l)6|7^XG3r`5Yc}ZuienGA<^TLLZqjVj=$`
z!kEUSl??eXU?6j%sCsFMAzU@%bl~huP*Dii71~?8Q5G%@fu9))c;g*B<6V+rE-o<t
z7y7}^K!;nql;aRt8R7KMd$iwlbuEu~eRWO9&j!l0yw?x)y2}UWq4YhvsmKcEnaW_w
z$M|K7&FbOCnb&!dhQhoNtn+bu3712aMSfKzH<%zy;n)L~H=f{;uwGwC96NN}1_F$d
zpwByU+~HYUlBxWmOV@UJ+`=woz^`63lSj8!H$|4qKX1t&=KB2onWW?XXCLN;3~WI(
z+{)qnpV!?0TEo%$g7KR}TmrDcNRUMH7EXOHF{q^+?-k3Ks>d-G%$|Sjx?I*&5;0m(
z7msGQvIDns4(IJx|4q$_fZ?5~B{9bh;V(@)Mz$L}>t;M10^L&S*$&UB^ft|CD$%1~
zyiXN7aQ(pXu;H!uv8A^O66j8rftjZub6Priqa?tfS1s<rXlSV4zre_>IiDRWb#0K!
zQ6g!y5^2~P5E>E!#CFvJwBn*5^mVJn#gDq)K#EEN@OsN^Ye;k?kSR<f(jI#xjv&pW
za)DHz6l3vAE}vjV!lBB|`e#9dINuLwb&Z!<66K_X6OSL{J}igRS#-b9;rQ1B?lQ#o
zi^>uaBw5i7sQiA?z+uvb(aOAk)0Rxnzk0(BOwI2+fG1J`Wvkuy;jJYbwqO6~QWN!+
zWf>w-9t4SuMkyPIc2qbW?XAEE=U!wbK^HTEu2Tk8l<M~~{Pp*DmT{i**c>#C>2ZBG
zO9#`i&bIk~)(HABFN5${T3KUfp2pkg?>gj&)%7qrg^yl`_@R4MS0EaBX~vJvnp=;Z
zocB{a{xy>5!)qL4YCjKZcv1XwV}P#)BTRQ%(ILXnNyCxyYOJ^1=pSZbbVQKzd?8~(
zVZ{p7Vu%$bI+ADU-9Z7w5T({T*7B4-<o$gxx87fcj;YC?qy_CKeeK;4Q9g&vlBsrS
zD842G1~Y#~y!%G}v+A?1+DF^qcDH2wf5AdqX4Sr7ranO;JQAdgjQ2w?34K-G%U69W
z@{YlkA#H?Rn>Xs~+QF07%<mrCg%*mxXLTBfKgFO|FlZSW*Zh|f1_#NX%wt38{4)Mf
z(R#{o5|~d6Tvl=02T#(U7Q_T-oP@LI7w8auwu&EYw@oeEC17s6eJ{LCqaB#d>64sw
zLP@dt4oE=W4h`_y!b0+G!VN14?Opvz;Q@JfwoPS~mM<ol>@BS5BKZA~D_xH^7^POe
zI$CA;oEhWe>(1s8`inmUn2cT!;~&PmLrR*$Kk!}Z_P>w^GSV~;vfx&1-ZiH0vtAM^
zNZr2M?Etq3ZO2yx@H{h#4iPe{M(kb4K@N#h*=1LxEe^>Z-HlJ@SE88&db|YoKuc7^
zZF76J$edT=?i*+X%WS*QYP#P=%|&~=2X1AT4}+sfPd7U+{8&!1K~V=eZDzIo+MdD~
z{^k{QV=4Q4R=mn##a(K^<u%6pnp5|D#cc#yWmX*C0b9!(zHY}i;wY}`a9!u{qF=Ss
zfIC>4#A`*jye`LWgps!EEJ2E&uC`Gs6UiBje!<W0tLQ8&^sM3bd5rS7>pd`k%4A~K
zHYIi&ww;p?TW1M>u+)&2b$HkNffP5q@3B;64kV(pOjGCelS2nJDYZ|csdRa=u^7(e
zd%kV<qk7YzG)<TI?3Z#Qs#e7JoTueGd<bmEe))ur$z?<PwAl_W^8<|+*vQ8aEf7>F
zXO-geI$Tixl+6l_R<EOQhNVfN3GcU^FmxykLHPBb=j(OB1n5?)6>*-&J?FXex9*R7
zogTCsB_Z)h)`$5Zq&}38p<>g+(rOUlTH5pvd7pDULZ@tk3LM9Sbb=m~Y_C)a2&bgC
zp24Tr1tOe=(3}}|P*1e66LL*14&4MqHQkNxD<WpAUjIPVQ$j+yHGPutUS6sbv_P`G
z^W8vaGPmQw8WfhKg1pZif4&tl6ce%YF^k|uU~E~B<Mt;PePZ$&wfMY0CU;Louck~+
zi(tC;RQEW~{}-%9nnTf#Rsu#xAckxrIQ)fpoA0!qe7875bW7np>G9nZMeX&#zbEAf
zF!kAT;^k`CUr^EZh>q|=^Mv78<szna5rey%A`~pC<>7IA1_f)|8D)N8k%mJKAif!Z
zR(f2X%|$$%D(#=0LYp%z>DrsCOWK5W2XE~JjA7c4hYIs!`^h{+LjMe~$FuIVuRe9(
zsY~2UWbGZeG2PKb@;Nfe#`v5iL2z3lz-M<G(2Ai6{W!6kF_4h=lK|-lv)91wPH!uJ
zjppfmxD6z<PP60s;<!Kj19a4qz-mVrquUOs`OJuWkN#OC-QeRPzBH8e7c8sMOx#MB
zoXGLp0+-7k1(5vXGaeNMyX|wqG`s$7g*T_Pb!E=h1Lt+Qq4JHUd!K0awaw-T8amv3
zdtwU}?n5)oCX+n`QLLzRmMru`Cx&2yz^%1{&1sS`rFs$i_{r{dvf=u&>*-gAB0>H9
zcBJUr?n`v~36~2avrcc=;lgL-p$yDF4Cj;AxC|t&bt~GB8pykC>%<B62FVA$5(DI$
zp}77M(Jb8y-x0B9Fj@7F*)&pSh!|Yii>C%{h$-dMx;=NbD8QJ=6Y=$CQrkL@rCeU&
zzE&aqp^mLec~Bj_9^MtM^nI6==Xc8nZN)#`GtMBPDd=?%^)HrLknQyF^?AI}m5_*p
zH)^sLb#S=pw8yGBqZvtTM~&|%$TAT!QiOi=<9A@kzXQ8icJ-ZpdwWQ7J!}wL<R?B4
zujNfq+^9_ch6#iZ6?yhQkNV{xZ4LwRowOLfUOd6!#Q*yx6y1Qt^<Pq|CyHWeT%-qG
zCF9_AU*t;3CDw{P<1Isy1L<rL#ocCIKVq2i*N^UNXF^IC52L;AYpp-%HZva)#48et
zf<=|)KT1=!uW+T$SAlSL1)71)5Lf~?#^R!}t7Y|XlN2g5=qp|yteS_;K#%VOhfT)o
z(e6Ow5h;D-@&sb<I>Dcj$P=s`^50Z_{hSQC&+c_~sQAd#PW1B|W#YF7;8yYL20j-P
znF6aLp#;bsPny_$#(@m=m%r(Xr4q4XtI~$gh{D;rosADqTt_Qw4kDKEZ2f8i%%G?X
zeqBxTxQ`%tE96Neh1|?#W1X~20Y-yqK!3-ku`_;msg`TBD(H2h*I1{kU*5tD$-7Lo
z;Kf{qgYwNzg92;Gu0WWdI4hjStWIVwDwg&UkmYafp6w2orH!r2I0jSxX)VdG45nL?
zws5~Dx!|&vX^Zx@eV+rCL`(N8mx)Y326cObcj*qBVaKa%;MUM%A}o><FA@Pa`qqoJ
zaLx*w-_?3@)ob<GMX>iGt<!j{wjUS=Q>jFkSzIDTs6D?Z{q=HOX{c5Pj3twJ&|h8c
zV)d)k`rjqu**vCIWbuT)XmVt_ZxL;*X7N}DHr5}zX=04V$1AmD8=d#T{pxb+5Lvw7
z_kjeScpBgRS#7R2uRq(Yx8-I|Y!Q;F3hRTN1f_j45&+&KaoGxk-Er!Rwzu&Y5mHzL
zfaS+>yT7^l#Rm&Mxj4bHLbpKY6~~WnN;GlNu%S6ARBzWUx1=&z!x?Ya*axi&VQK$V
zd=K5hlb!EDPs$g4P71uLmuh4Lm0c_QyLa4`d2_|TK(dFu>Sh<6Ax9$|m2Mri&nB+=
zwuAB)oepq|xA=_bE#FY}yW;PJ6c&Si2sIx3Q!su+QE;=Co7q+K?v?-`U=s(0@OcNB
zQs!W%P`CTt@HrUnd3!vLR4jCc{}Mc0?bEpExIiyQ*wJch&GUb0X{DG~%wr&*x0%v_
z3H%!z;BXQS>`2IihfKtPXoWZ0`i>3-`Tp5anDR&!sYW9X#yx3#EmR74>-orM&$;n~
z1xdj6ZvNZW=^w3X7r*DLwlROrLao}sw*RPd2Rckop?%m3Wre6Usc^V*GWS39`h6n6
z5cYM4@ccpjl4gyjhi?uy|2`b?iga5nhnsMe!pOXpkox<w%$~LKC$(&Ku6PMs9W4u>
z9?SJxH?GR0Sul|<qJihsbs6a|Z0%i~ZT4w*YbK0Dp8Kx&?D)*$W)^K$Zkr_PAKAiY
zwCDF_#iZuEhkiuFTkfjQgAdv#ynYbfR~zb+k8knd-?vW{QQV(Yj>jZ?-UuHodKKNH
z`lVNLOtU=K?_++Ivx6NdPuOMrC2}Q=Y-N27^^N{E@qqEYdsFuS8pSUSVl+rTJ%2TR
z&S=L5+MdJweoAqgkwU^RYluAYMa}7r(5z}D6Hf{mPZO#q|E{m#frzwh>yZ^(_lB|M
zkJgDy7<h`5R-h9}UN7~1vfB=--ucr|!1GZS<>+7)NleOFCX_{|Gs-SlV(=~vZH==g
z+#8L!(DJx<^7VbOYMI9ajFl(I4R*t&DA>F%jT`m67W5Ev?fs6fo*qDH4mb4f`;sLg
zk^}fxvBZ3_U7hlx^I5Q8=D4#$mqL%>aT9;1|I`_Zo@}1P<u~@_l^Bi1{#n>R@>>sV
zk=Oj=voY9(k6RG6W(Mec;HWkKX=br+`7gCziqO^HPNhlt)2p(?Dft~M`QL8BmW=!E
z(`_#P%4v4I<Nkm%cCW`6+Pivg&~g!ug`xn|m9TfL8xk7B7FhjS07F2$zi98|0q$Pb
z&LfTc=g5L!?FinnuSheuj{f(;(Yax^&;f^V-uKwhQ$HiI=%DWAzxk0xa*z+TJ9p1i
zQHhOtz>zT<r(l8`Fq*czT2DG#>Q^L&JNG|72Gk0lc1Hp|z3a$5!5ZUMZ-==eRov_(
zhaKeXpxzkWug{+F(0<l$p6>~1|H69``PW27=OM4CmC8J(#V*Wc9}v5IC9i~1hytY-
zzH`zXHStp0OqmDk-5QlxuBW-=#0O%-$?j&(G@l33P%D1*-<Z2%1@v?a7eu(IM=zUS
zJk+rSf}}#8;EV_Y3i`Zm@ay65`GhrP;U=VFVFKQZxgjEH8qIK@K5fZ&$G-bvqWet;
z%^Sx%d8pO}oPX#FZ0$PBGi;s8^9;74AKXf%-|11$$#Ny?>!+Fo#F>o*xTE;*HyTaC
zd84t50xJpeUK_k$CHud0dOYFSCMFkiSVL#}EoEe>&6^d1IjYKFrHzGr(o7;Tp@;8`
zL<I^L*hWHu69WM7oXKw3W=5MGjr74(-1LHe4{~-7hGnEx*@(HdkNo*+HTB1q<5!fo
zIJ9xTQtvzXAsG<itIteihywKB-Sq%eW96sygD~#z79v<z`?@c(*9LW1TSvdAljiWe
zml)NSbFA}n%767?*Xle`(pP63MH$@YjGS;A1x?(`&c^Zd*(>LBwq}@1<=sao%}{%?
z4^!~fNGO0I4-b=zo0S8E{XpPEYIU~8rO%L+21&!!${Tgh&;JRAVH($-+7IQHz{Jg@
zUkO+&4UO)bc{RE`;p8tgFb!@#RmFvQe=y!NDFi@}oO{;R3#{cUuU7gUDk_X{ZC6R7
zs`f`>?e4efuxj&QDxTR~D9Q_d9#Ne?t|>q6ybAgc#}F^tWM+1;?8m#B%a)cYWPXwO
zC?}?!IWI8WWwh*O5vo?AjEHS#E1KhYSr_qq3Av+*!c=~cl5u(N{_-~kKrL-1kJ7pU
zY`sf*bs$zRJAlyn)SWU{?!s8n51<}8Zx_q>^sWIJ`eGry*Z<Tw`!}0S`QLy^EP9N8
z!$5_R1OM7ZJ`DxrhuH~ug(@86G-f#e2{4N;Q+V}%tc7B%EVAzNuckF9h?K_u%1`Uo
z@PxtTW*p6yhHHJnhVLEQFuY~VJqqx??s0EW6#t3@?{@coJnfafX?jXJ#eVF`?0%KP
z!Et?fwI%3$B<P5L&%kr2)>piKs0dSNZ11Oiw}JxGGXAzb$ywEjHWR|DczS~9yw~JY
z++Znfc01U{8f%&2^R_F@hE~%FjC}*}7QP)i%G&Cw{5`yGCni7QEsu*f<H(GI30x6q
zGpexryLx))9MH3vSz7qB1H(!-(~FupSV;MbcGNf=OdsEQ%%VTAq}5ViqE?((uHL~}
z%~a^#Zh=LXZYk^c;eNdxO;0jGTSk}3Vfu&t);i7cRq+@6tDIhlsCiwXZE$4<TYI(|
zcHdY5`S=w$v8TEHW*s>T*_j>d(Q;etLP?|967A>oD-x5Hx{~*eH643B($VkqyiQm*
zWzI0;aUk4<CusCqYYTgUqE@%XC<oTJbVJb}_G_EqR1=PgWIUAyQ~Gk<DGEW=i95us
z^6&<unN3dHwF>bNNhGSMGP|-Yk{m^!8D81U$(_mJ#RQK`7owWoMv#5)c`U<lM_q)y
z$k*0i37g{cV{=!zC)aP^<Ol|B^C_~p%PA_EwOSy2f5r^y6SNhmr@Kiz2lL6Gy|>$k
zBa&P`b+>Edx}QSbcH>+hh!<K?t!fMYb8c|msyUh$j4Gg0XGAS)0-1WPpY?7xkoDx(
z{=P<b81K_`hZl07LF0L>(y2MlTL#l^{>S9mxY=nSPV-%s*-&$?y5q;La^ZsGcUSf+
zl5X@6*D?}=)w!+{wIL#+W0Ek<N5&J)uJV<VdT=?~Py5b@Qn5=+?XP+z*D)?>#IJa#
zc2gbF(8zFs8EFJVc-e%JVj|=T5MNnM6uMbZKC9a3y$I)R=JEtyk&m+W+2{VgN?9;M
za^IdEQxq7ztUiNWcpcPQ_DC>fkB9i7T|#k?c)nhpKanI-RGnFe;oz|+D{c=ALWf6+
z0@n{gxkN;hIKyp+i20NE$t#Y6Sby^CqDw!lPYdcu(idi9NflH8%D2+0!|ZK-LtRRF
z6yEoHT0ek_x@cbF>Ne(80+npfFrkB8E^Laiv2iSsvl7W_5>4-mU1)R7AdlP9nYccG
zk}%k<ITa%5o{n?3{t(EsZ#1Gk<ZUFdJGJ&-e?D1BK)7Z{U$A3w<lC%6tTcZXv)(ee
zqC7AP-lIIsmXDf_U&HLyscbcv@fP7MF)|tH=j`JhvPOPeq6>R}z7j!S{;~7V&TT8$
zKbjzVtwtIPGoO*+<3;8MYuOfbI^C3^!<dyTd9-{tMF<7tk0$$kuLC=;d1cfRkNv(i
zTj;3EjQs}s{qD5;l*23pWJPqL5nCM+{0Y7P9sg!0iRNqTL%vlGn3S8ysEH^EQ}0eH
z^ZZtjTNOol|F$*r4d+?1TmxehC0c)lLB**F-SXQ?_&ln{7Fj-nZ2Vn}z)jJXFDZ-P
zqK@8ga;U!KySQMYbg*L1iL;qYIc*j@+$YsGtBueT{8s4Q?p88<jKx`?9~N3b38&(d
zNBldEHp??2Mc;Uezr|RP<49{#n+iE8x+Dp_WU$|ByMj{+21Jo~e}dpmFXsu2u=fh^
z!Iutl$LTX2id@MDcS`=%_|H~GPUl8qw5t5Y73QVqxaV{_2up}NSTziJmjCVq26zv<
z!6ZH4t!XC|C$<XW<SQc)k|6t&n+qU?weJlL8>wMyjqEBP6DGgI$OSoTK_?`!LKlbo
z4LV4lYRL)?IlWF2p?^uA$FV!}PitqaXf)Wji^=R$vv*c6eCBO+ic-@g^VTlB9*;VI
z@>zVLVcrH<`og~dKHjzXeaG$5&Tlw$Z^$}@*CSG(CK4$JCtqMWgZNt~Wk(zmFtssv
zFQhgM1cgnSH{X#8M%@4AzUVUf=RKeT1uqox`Y0iH*pqOW^%WdNj+ykM6_r~mT2kaU
z5P5eO;vObTcsTU^+}OMIyIj?C{lORihl#g}s;lXmMsas{4esu)A-KD{1a}J%+=J^z
z65J)YyKV>?+}#~Eun&3O?;qpbt<e{2walKgs;hdzCbP<>inTYGt+G0<brUB}ldI$F
z=~dv-z84$G$GpB>e&u%7jl`oV5~I$vx9KhWOt5?^G+(>TQUNE7h9`uNv@Gk-#UzNb
z{}GwH-PnJmOo#!YzFKHZHkXkx5Uqye5f`cHG?AC4JHMIy7TS~g0ynQB^sPQ7{PGW9
zbfZb~+@qvSgM3IPTwwTV{%Ho&&lQ@{fOpnsH%esR`<^I2l@Pz?)xUe3g_$~PKA&S@
zn6vZ2=}kG`@R@v)#TB@rhVfPz{j51@Fq04nIHCR&IFEYEB`AFa#HcDlD`s$4tQHzG
z{{iwt3{Y0~&ZK4YvnxvH=%{E}LJ+V=m?)KBVdTtev+O(SyCAlFhy#nM3NpHd*BY*c
zR_HF;ng=%Q*lv*Nxn!BfAX1wmr^S4NQ08nALBLpJ@b1vak+uvjdrJIs4jX*&?+)Cv
z+iX+N0$uACegb*LGLzAGqjWC?CjE{jJYVE%W|y_la#UVC)F@h3J@e3rexapV+*4t*
zaTi{jCbD&QxWHKBZ@qXV-3DL4R-cVM0JBj$roMf-9K}&A0RRsQ&?b;w1COw6KwR>c
z8UOx?@U^nMQafC|Ul%A0F8s(7$};Fv=JiR*v7sl^#<y&<&wHWBEVPq`T}rCU>~Bdp
z(cd_zdyQ|y#|lK<XBFxog$)*~2>AWfV^rFTGT?(qmknx-V}U|$+{o?+qm)O^c-0%<
zh_uEB#%d@c5^`9FPfmWmT%2?G;G7}vjj+A>BT<V0%}&;6Y*(u=r;XZU^gO1J@xF$W
zgxf#Nqx-&~+|xhRmhdm*_C7N$UGFAs*7?6-2_pu^)b{{k=Rg5c!@&$#kQU+v?7mZt
z5`m(wYmGVqzbCV<CWdbrM4R0Pb;S0#<a%mugawMdpBSt@R%v;Tj4AQCprV82j713+
z7A;W=F<~`$!ESkfDjGnXCcXXxHaT|VJopU+^?sy_*^SfW4Xt+KQL80gX2}M2!Vj#!
z+H#s_P=m0sb=~5J?KxLpP=e{US^@aNf9(r3x4~`T9Wm?=JGQF8{%^iMO3X)IF@X6g
zA?9?Q(e(v}Qy6Mkmh?w1@%U>7@=AszxW{=bazUqWUdT~?Q#{12(MAo{WsfB<#<oOe
z<B&6+(Hyx<-9#XR*%VC_i#jD6w~rdhYt?SZYIc%Whu87wAY8G@Dr0}S%Vw{-_QrZ#
z!XN>BvJ;Iq!B%`QIak#AyMrboz*zc7e~e5nFu!o>#}S1oYK)$R{CcMG_ci`u;qc+O
zw@8j{n-SelT>)s8CFFNZbSTn^_3pD~%O2Z*H&RiKE4@kZ=D7pJxSeCI^|eacZL%Mg
zxvJMx(7kPL^1+kNzbuE{^ErMJlfl3hF91_yrd|hU!IyyI<Y=EeJ+!>82y@m7nZ{c_
z%?>2~t?HUH18%f*vIEPz9)_6l4Eidq@UCYaC9?0IaS$@Yb{ceSXkX5V{&t~Oq%+iL
zEvd3)y2+7WL<nxg@v=(23!4_-pZLzrIBm%{opPwU%rwG-OmOHk!p0kH-<|DBcqX{T
zh`Hw^A>Xc<)ig=QJte0b>EZ^VO@ep&z@IhQl7Mg>i!9o~a3sQBNJH!E(uDAbR@u2G
zbU++<n|A&ci=i_H`<hpaQmx^M8>=S#d;S?#3wHZ)$JD9~=_3o)_kAw9B8+2pO^!V<
zlT@!_cp_!oAp3Cd<z8&NZ}Fu`>D612h3QEit);w8(DNCf+>?8=pL<RidcW=9l31ol
zLjg_o&xRz)tiW|?NfT3~oMq-E`(<6(B*|OO%TkHA+k$g=aSK_>QWv8*uh4$gdp+C=
z*cV}8<Q(JvuZvjIXNKhb<&vr=GC=2B4Eu{at;RVDctvW`*Y43ivwN43pPtpf@YY|f
zrCQPRr!;l%qru~?XuD>~Cnzk)Xz>$@tKd7SIMM81@-QDQAx#pd$A5nDw|Y=nbI&<X
zD&?sCQmQ4l%K-@#t-8{v1oFn*_T33h;pdKY)$@$?0v;Udf-L6`Aq*~_&HWHhi-vF8
z2uY$}Eni!xfMiN$Xh9}_hPZ{!emP^yVVF#LaaK#39zDR9Et_HZJ)GI$B=M(U+OCpI
zZ5z(i1yi{7j%nyP7uQ)x>~W9D%6a9syAj{}#YtXmYt^eS1&rkkhI}>Y35+xkSKq9|
z!+HqC!^}qzW_B9^XI;^_Mpf4KCb7?-yc0u3(@Mu|Q_i$$$ocrU7HgbZ!|}Byl8c2N
zd#xGeGQ3Np2`akWtkKNd|0OhQ0U|*M>!ZIEnm}xH`o@}H)Lf&{@b57MKtq`&^P%US
zBO|eae2DEdefG$s`0jjn6;j2rCER%`2o&Ac4@j52UpN=5=1Q12Ga*Z^#Y>cO;5MNv
zuU7e1kooO|tGG||H_fuXj-(0S|E14Q+30(IIo*8R-k$v3zno_s23q6rU+pRM(Tcei
zt+|f!b}L%1pGpY}U6s|aWpLV<?Rc~5u`z8YrK`-`Jfz%q9;6B-R=am$H)2>pP+arE
zO;0n}%`gd?WlhrgSj}(J6u6l3{->&p%E}e#2>({}7!mRTF1zlKtYOyIRs?^0eIQln
zOvxExIAHisAC+7v<^FRjV_7gAg#QzJspvU(Iu)ioA3|N^`%#-ilsz4?teUFs!!Xe+
z$ZcDRoEZHrBR^LIQY)#GZMHVvitLI~hAK%7!xa53ImMcac&LhitF`ti-+G+%q0ENO
zQwqZdz5@n*U~w2J;FDIaEv9&kbT0wsmiO|t0UI^btO=U)3B&7|9LB!#XHX}c*`&=R
z={>=QMF;*)-kJoiuz^XQ<KTozpUhBwZEY^<TY!A>e}lq;9FP!iW}9#srk11nsL}E;
zoZL{mJznt6Irhy@P%F&c2)Ach^$oFi>EknC9Jd>|@ZLL#%{i_YG42;vl<Ze-ytXN~
zt;@|w!g+#y(|>V4FSV6eKL9Hu)AieEw%`qY)ah%&)Wo}D7uqlDfAgw@p*lMY!ScSG
zsNIOB8c<dh+D)qjM=g~0e;tbch6eS4Q{XXt89q7`pa-)&aL*KOdo$tXl+@+BvfrPc
zZv|Vegm+_ZlbD@~2S}3Nz6Jn7mlZw(5U!|i7i?<g?*yRF)ou;{AV1n>6|pXNxa1aO
zjY@8`pEZMN_TM7WTU~+ZUlU+W6+Fn*u@WKNiHPTJQ+!}hI>HNs$$WfMd;64_{(S-|
zZps)`&O%a<?hS#{#N(*nFh5Kw@G=A{!3=b!g+~r_(t0|w@Lkdove@$#)Heqe_=PUg
zB6lU0jtlv^@cX{n$i53l)DO^Eto=;|(16pG*e>fC#Qe+E{#7m}7<DizceOPw=-<%)
zz(SD>?}`zQQxA7>7FO4m(}>k{Bs3(n^@V($Q{M=DMc?&)-ilHoAdBn?!=8P<tKR-z
z`=uNO&plDNKY^H1U+S&_81LE?8)EfP7`g0qjwngBQTW5%n4A7<`~lf-{rmX)Q1UGB
zI|yi7GUEQ(nKDqZp%gW;D%gdf!iz9s9eKF7!=s-2jokF&%oTab0vH|Z^Li(szfc_t
znU_r%zznltW~K;K$$rQFDS3Fjf7|=r+%|yW_-+nElkqfKZ&xm;oOSms68azrv`euj
z)P??+M-(}$1Ke`#UI*IZfDu8I$SFa+cCW30!(`U!bS9d8w9e(DM+|2Su(TBkoXNt7
z46^?W{57Ihlf*--QD7j2aj;U$dtriNXNaWKMY153-S`OSW%!_D;@&^;7YSQGV_}^5
z{t;F<HXycbQ)_md?*aLpC1+#9BK|FZ^!YOkClplq<9RCPOp1Hz|Dr486E(Z}sgq+j
zN!)K^Sk;9oC}K94!ozllEh8i^9Qwz-u*161EZqATS$99TZ${}>bZ;DQz0gfpO)_1A
zi_88@IGw6}WSa3`0tC>mKdBh=rXmnPLP3|X_Fe4aUul4bT|DW(jeP;j*1Eb6p}|pS
z=I4ff;FN+e2zUrl_iLnsTHV;0IvWlY-9Wj7k?A+Nna4<F7pA?aXfFIBP=U!%0vxH<
zjrvD;uv5T50pzwvt4Ef-X3-z>rMFmM#}-mkyroRdvPtj1$c#+UU-|0v!Rx~;e;$5E
zMhBRWZFiLEG51(&WTpS#Xf-9piS9HIb<4Fs-k5}NMAa*M$a5I#vhfH@a&v$Y;_QI)
z8)xAwY)^mxZV-PEn2>oa%BW8OU3T7E#_MMOG4Mg3pi)SOK8i&rf2(hfexAuXK4Bl(
zeDAkVs}znP3BR}Ysd*juRK)X00Or#U>h+~f-|tElm`(3eB(Ec_1vPK^9lvM0;HqL*
zkyyV(PP8_Tib7(f12r{Y7n4|(r{um@2A?HyjhlRO%~A~hGAFlvcHaZ=);Yz&So^s8
zrvET$--&w&fwFrr#skrtgJ7}IUnq`tK@UEZg>~j@xyqfQDw1=4-QaDG3*qQgxHkgZ
zvA9%8)5H2(-THfoX~Pa=IV4e`b{nDjnz9&I*SCDOgR1oq>ciSTMio}gYvve^t{xdL
zr2MvkiJ&vFK-dkdm}u@fkGs<=9?_uJ?|PLj)Kn)t`b)QX_r6su@XC^;fBbK5oN`+|
zu+g5!{A+{pQ`z)%FE#O;R>IO&(WuU3zL#+oI?FBZYrwY0IZXPc(Xh{rupxZFyy26v
zu$d}r*x8CCBhfQKFGxk?)HEpgFUbf$y<y(hy$L3k-oiBSS^LH|5}>vpB-T$Dfpe!V
zm*xjWY5V~@%vH>{{T-e7!|11M!iyO->)!(-RSN8f8gv^ToUesZ&aO#C175!oBI|4|
zRLVCwEh6leiGI0d?F!%6wPD+O%0hN&V)|xR99bYq_{QOZS7jKvf67OF_7^#oO#>-K
z4lOhgasjQmbvSnoRR2LYb9>zKyjlnE_PRk^nC~mCxTkp2y?m&rF*SyHB)z-7;CxZ+
zhH>Ecy&bmIQW7Zt8aUJYXb}G+1rAzz3Ds4#c~m#y6BIJqT7SKeim6>XDXvj9M}x23
zwetJXG1HyCA5>mtyZMB#$r^ez1$7Ilpnr$?V8m_f?GufH>eiY-9#h}=8qDp(rsP(m
zd?qINeKEGXAj8f6G=lsTXF({-O76mZY7Eprktebh?4Sk*ds}hjL#AqijJ`3P?*O@F
z1U)s*Z!Vwm<4!q0z;C)T2v~QMi7tBL``fnwU6Pu;KmxI;_6*Bn?Q--<%zINQ+4%?0
zy%)jAk*@Xs_6|yd_>I^b7ymc6m;V6*N`&t{&)<w4YLiEN@ldEOczUk9jC~0)Q-P}n
zJtm&8_;1EY#r^|3yt+X#LAJT5I2D_n;O`TV+XSu0w2qUFg}Omp%rxlt*}=Jzi<2gZ
zVULLV2$>C!A!Wg0+(;8;Z#6lTN9wa^G^5L6UZv=ZKMD<XmY>ATPf3>##uF|q4SVjN
zaEYcwVs)IwVk*MGZQ)<uw+R$wUlqb{OwneyUZsrhU%t0w4|^Rg9tzw-Ng<=gkQld`
znz4Z%rAG#0v24h?#QFM5Kqr;UKG*g7Q-jtQPhH_BBN+aeQz9o!<DEt~6{n-$r737I
zeV!SSjDnt_(=DsW*DFcgEP*lFVk)vemZ|KM!9pH_(<YCUQh^%W)oPH);EgaV%TXId
z1jK;{YMX1&I7^v98+TDu`VL5~k;7#LOOtk>r_`TP&Fg86M-7ZacoUreTT@sCT8K|>
z{<MV(!dZS9S*_+d5+4TcM`-}01M)UQ@cIb;LJx65)aX)_!(_H#s)r-6w%chB{Z$J9
z4O?|<{xpeu#Oue)x*7DXZ!?fO{u!y@CbtIDhmb5Jp$mTC4;A__+HoX_bp5T;5&(nj
zIwt=)(RQv&?1KtNK#Ft}=!LChs>pirZZyKcz*?N<Luc|jBNEM5#n9_FovwM)zr^Q?
zM}Rr$jfx0Lrj0_S=%%Ay=%$UN!6tM`^_Rf}YaoB=?x1+M^Pt|88TOCfAEN7Hn2NwU
zIen*ly*KdQe)ANPo}rR3v{;-k{E2-#>7Mf!TXyGe8ue(ixUonwup2z8@-yo&cyzh^
zsb<r~u+tBF;Kd_jIDkii(+uGw)PfSpz9bCvGZV{l-{K5kC)}dV2|fE?D{pvis~P&>
zL~K}kIG@>}^ET89UQkE8AH2yK-k=VA)3Qlo^G3sInUFF*_KqsChHsb=5+LqWSqlsu
zTdj?GdCdQAf!Mt1IYe$$iHkS<zWLr3#Y1;7^er?VJg^9$7JVZ*(yK)9F)MuETlbOK
zH79I+;b#5Q8i7TsY{f@BNlTQgWJ5kQGe8&}_?viE_%YG;yQkPfT;ZjN06C0sZ2z%8
zrKw{(i(mw5saB6bYXgG-hO2>S4BjYqao1$h{}rZr330!pQnVWN%KupWjO11nkPFEJ
z1SrU0xJ-qg8k-CJlMh1hy*DO!CXE&g`T*QMIFmFhc{FVAMtoAU)iGiEz}}VcD=yO4
z9~Z(m1?(5K$NKuIjW#bzH#a)p!%Aw7_7LeNNS?NX9}xU}(I!;j3j*6h6>!bC0lrs-
zSGFc=g{h;p8iPn_rFny_J2lSI^E|?6#yDGE0XPlg(UX3ix4vfI@KNugD?OD-Ky3}$
znDBTY35&0Mc9eyi!^EQ!$fQo4Hv@pU=d({uk?duL^lYpzJ%c?<QZBo$FvGmAks{|V
zrg}25WKu0_=EMk$7&T|*Y#%6}2!Vv(9oBiHOKyagM|G?iL4Tn#bn~EQD5Wl=Zqf;@
zG?X#%?GFxj#{OQK!PDcY!0YnwWIukPgM^3<4T;%q9u_zcXDxCIfA8MPGQMS*d7dSP
zWv6&I8ks<}ZqU+T2-T>5iv7N7r>}DrRpOI14^zc`J>BcPRnGTFVgK<)?y?=&#Pq<r
zw=2CHoN1!dOn0?!N=n<|Atu(~Us=DD@k<~&@%g6JP26|Pk{G*XQUJp-&$m=VFY6;N
zXZ=ddgsVE0HJwQaKjmN^$VEQ))ovEKpB|JM)--K}fj!5WVIoDt(`1bKg_e;M-#CA1
z3Jql@Sg&4$J(WAa-+Or18iZ2SG*KwOKC-kwNYqZBvj;svbQXp8lENo!5fSb>XseyO
zT;I?q@Z1LD17DZ!?fbgi0;|?qVWom!B&rid*6F5<<;ku*N3q}0lQ-+UcyMZ05DdJ@
zAx0jeoF>~Qr{?|m`;egrhE<lvHCR7rs_Xrg*FLAjVPp_wk#U&UI*e#?Q&hdzZ4bZO
z35oV(el7)R!w>1LFBC4BWmX$WT=E?XtYur~@y_Fd=KYHk2#CSdwpJw7nq(j{G(2;D
zn~O-(!!(FjRW8AY`2txQjxyKJ*UA?&-lRACoP95%&C+kH5^Uid##zWcDYa)>q|=KB
zN5*rc64K6%cSlA6218oql9zh`c8et*ZU(2sU+b08ukAEQMg2!kp&i|y?cj5}z^Qn!
zYtlkrPt8-lI2RZj6Ku&q2kvA8N$7$dDr)H$g~4_5n3s-x=k{gl{NyZN!b1;Vf?CP0
zNOnE>?<};akpIV|q$!3lli=QPA~hNR2Y@*+O{LmUeW9B510WX+SyboNv7qgR2C{E-
z4}P;y3eV`|Ar49vWt#!-Vk`?5aY}xgJ&V0qC1nV|;kJ`aH*!uP^#I8(0*olUz04*L
zxtDB;`Tg%5Vm5F2q9{~e>qZ}+tX7>oG3Sog=qgC(9J9h7#r#2l+rVgj?gO|XN7agg
zw)UDIWis;;!f=7+`63UsL}+m!A=hoT0$512n2u^4y=BaG6^Qur(vyUzbUaL-`dw_j
z$5plt`8%=q)XSg4@u0kwPWprm{|HjcD;uM*lQNWwmtm^gPf(v_3JP2BZ}bi%ipfU`
zw8PcQ492hrx}6b;$1U`De~T@i35WjLOAFov*B4Bft&qyNWgTD)2Xr66LZl1giVXc_
zE=ByHY}5nY_8CYv+6+dT!X3FE1YAz=BKRTZP&i#qqOeJ7izmMuVWi;ri&oYhA2v(o
z^Ar5+lm3u^wI{T~d~qK1B3=co%v6B#DG^xUa_n#&iZo|_a6V*no0O0Uq^KX|*^yu_
z{kRB6PLOd(Awz|M4;6o+8Ce~Vd(wO!BdEnxY#h%>Fk4G^qgPpWqdlD!suk=!u)`C)
zyCG$g%kPi5f!jIH`1jlNZ~&e$gsIw*-=wWXyvVzWjBHH6yGFC5-aQvERY>^(w$P`G
zDUIv}K8h5o5SVY#TNbUrD{`B)C6&rY{qh*+K6wI{T4pjw%&W~9_RCtdbZSb`vvH>@
z2<t|3ej7Gva|C5TNFU<Ob(x^**+wXc^r=_^W;os(t3fhCyx{9tXdNYVK&;q9v7eg3
zJfk)QG;unPFG}xg&Bn6JK}bP?S#}QBqZI6U0xiE4a=GpKppd|<(c&l47~L$JNN3%q
z6jY6aIu1biX`@<h*9*^3V%$Ne!=v4xK^z5Jpv+TV?ntlk)j1PB*2cR{6m6N;HI(h1
zS)w`spEUI);DD1XOq7FftcXo$y0a?l195kmn^X3UjmK?_jY-%|Q!)#Bzz3e->g?i`
zV(HOi{;%vXB#57Br{i2I{NH9E#D{zk-h8*v>nj9qyePVJeLWP(CBI-1-l$X*W9T)r
z2=+w60Fig(o<<Q8F7pg=8=MKJ;R!UwpDLOSM3%6lCQ+$LpBBqc{P|7)LuG%A@4~^m
z57=}0putR)0M?5n!_*0OZDLI^DQBJj%3HEU&2f|co5fqt7uLq^y2u&T#;o@+6Ox$(
z=P?8BtG(xISc%o~E5FyJKmffpg#R6W$@|S*zYKf~&6QIdlqrh0A$l5IAdfv~@(+Z%
zzPH|;Yh4M$`?G2wHy^c4ZGZLo?U)oRwy?oH=K2ETWylJyu(^w3&c5b3<19&v0O5@V
z_QCd2>La-*(pSH!&J4`i;a>ToeYKxPVfOUeqY}(mgS^&OY7%w>Lhw)WW%RqJ(r|&i
zsxDRE9#AV0!$jF^`SB++RJ9uN|H&q!X6B&-IXZ}FZ-IONdN4pY+1teh8sq{RUo$sX
zYDwnVryXZ4y(jQ2{e>Oy+~rI)I9M<qv~hm*zSPnrUP}*W@We4g_mSy#nuQr!U$;%B
zc;GlaIUIts>XVpF$!$#gW$#+>W%Zx}2=g*|Hgn@(Hq#^LAskL}o>FZ5I}PC{0hB;g
zaL~w>U#JwI>ptti`t0CKD}TK}S4wbrOmRz0cLn|w8&u~>aUK^nbVo_%@OI}9zS41D
zxpH9cZGi3gmm)~@PbcegG|`Lud1`xH@l!B25o}h@Aze#Q^;f-4lWT!eF83|vW%~S@
zL~VZ$pvW0$9&jwaSs>8&=e||j8p=7t?^3#+-tI*GKQZFz2Eo=KvO56WAopKO5+%mg
zx7IVy=|1CU(_)W2z0&#!dW+>{@XkH_*{OlV)jAgi$sLctO;qQ-fbF^t;Mi4J#D^7S
zD_f9s)ONa`n@5-ykM~7|j_onjw;}erh98?kct@a&m3T{txq~@$?}X?sBl!{hvy0<d
zT(1;)_U`PWxCm#<LB_A9i`5+3MNiGZi4kJaduv)_(%A&7P1GOz89RvitZ&nN7M7d&
zE6b5Lzlx9iY)$x~+H5RN?mu{<j^N^=*E3T1bUyO1(|SvfktT?OCE@mW87wS}MW>@A
zQ{F#aG!a<JN>_Q>(JgWEkQlW=OMgcg!@pX7%ikB`tH+qmNHsiYSF)W|XG?wNRE@Ee
z3GewFQCVr#H@r^36<qzI&$Vr@5^Kj)YHK}P1DWPs1WqZX-#(xphoYY{&(l#iL8RCN
zc*H}%yWb<D^e-2_4Z4Chpi?V1u}H3XZP2iVg9#HFJHQcg*Hq^D{^;Z?iva><RaA;5
z*aWXIGV4v=x%Eq;U-lb*4Q29AMSxwo4*I(bsu9GAS2KCa{ae>N_Z>LwPnea5F<K;*
z`I{{Q<6c`iNzOEp--1H===J7}H1E*^n5TIG@o%op8CF;!Y~c{_gQx}K%{>^wqU54|
z+hqb~@Yj2rQp#@mzL@u9@X_e`VXF-!otSTdxv3m(6h>3tHz}f&F2LG1k+F|C8OMgj
zNQ@4*^C>xRr48NTJfA{B(qD}f-WMqQI^ThA3qGV|7WN=+hYNgo2eh8Y%oKt~hK5N%
z5q$Gg3Afqzq^)G%!`C~LPa6j?5+jGwcM$&H32sP}ggDXN9Gv&xe^@%H5929DSa$?t
zMA+HWNjl=FAjA3keGaV34=7)8&kyTgIT)Aia|nYOQer@|p6Ihkp;G+C<u_vFL6`s1
z!zMC}y0b*C`(-t3B<3h6GbS=@szG8sZ<zCRM#{0{1!mJ{FDy+@dqoWc;bHGmXO1w-
z7R~IAy0j@!v6h@oABR9w&Y*|$LRhc^_Tz(I2~Xb(k?w)7-~jr5;kR_@P4`SM&ew)>
zG}#&R0SWyrT?r0|R%iI4Ei;SvgNeSf^RF@iFSN_b7TsI;%AK!&pjG?S@-^@vgCgW6
zUy$3T+;&ch_A#G6#l;<xhbQL&nbD!8OuyjZ1v`mF&FNhBHp)|Cvhy)en~A>2YIS{|
z4Vnj3+PPjZaZ{|2E7jMJ#@ibYb|`6z4H=P+GHsI0rGvwiKq`fvfzEaC!HyqPZQvM*
z2H9c|z1p*Js9A>f9F}jo%A^hDMnin<I+mBhAf4G9`F&Drb?YFBe8`C8?j$)2E9CAw
z1X~n3DIaNDv|y=Tm>|^*3o-ET=rpnoGy3RNz8HWfy>>HB<B+Y`W6Ye&^sD;8Sio)k
z>HIe?!t>y)yGWk{;-#9JU*p5$SPM*P0~JcEszp)6G08gbdjiPUj}IdMG+g7P|9Yq{
zjgQ+U9EDWc^7VJ?@0q`pKU;n%TMuL}s8Rq>hu`1n8mzPeX(x5-NW$%0zl|`;P?^#r
zTs=M?Dd8XI2{Z8%I8wN-G)f^8|4m!3a{La=ScP#=-Wd4{uQJ9Z>ilPO>{#SwNeO=@
zjQ<HN3c}!n0)z=X@eHl3NvmG7FR0Xq`l@~(z8SeVUH0o!ad|aJ!GC2$`U$$Z7ecXB
zh3a(YIa@IB5h%x7;_*zQ7A&9~ysTBZIXc6e(UIc=jx!6P%_>K1H>H?ax~q^Bz$WbD
zKo`WN(2zuj%EkPbLO*B)D%76>0B`GrG0neIm<VyD9X}gJQnc3<XKGW)dCsXZTjNNS
zKY%FX$mg4m+BRBM278B?cS~_V;8US4jIGJ<Z^li_Oh%Z=aosop5B|0~y;wIM!0co2
zmFHN1w8QSF3J30iOO}`&ke1G@dv%5L1`B(>t=F2E#sx}+5kku22EL}mS|;jAS0`yW
z*_0TK-E~Of_woAzgn$BARNj@_LWXt5)v)rQ8nwK8e6n3~&b3CIVwp$*r<hoeC4~om
zh4U5WHu{=pl`!)~0(G6$95o|)Y^)GVs@~k^W+uM^P^Q?>M@6V+VE-D%X!V^V=lMnm
z4QzDRKYUMVWMKLd!(ypwg2=T4)bW}a<V9h%Dej`8k(VpQ<Qgmr@!T-X<aqKoqYuU?
z@FROiuYi9Cn)nsI&Db5?Opq7ZDevs(Kiqovw_YmuIa&}m{z-IuaPfvCN5th_t|kzJ
z$SdEpejARpp1kex8aK71?KWcgh~NbAH=>QQFPegz=ua{viZ*j4dQ#j)6ioopK&*i0
zbAL3PIcBw=>=BRWAq>H+cyONX`cHEK`S)gh`{2jmS)j0gb1-NZ*uT4xExckA4~88Q
zBkt$vxja-M4dY!06-!<ANwQqxWpY^a6FwSkU!0JfW_pfn=@rpG$;Z&|s!dl>6VHCF
z&K8Ov=0Q*re892ymFtU;;X3A$c;8b*nr=`!p-`>#37V%*ClA!l$AZ_maxpbTlVsW$
zGpe3-y5^V_AYS_WP1FI18doCTxT(De&M$vvzViP)%p_114b1n3Wu~2euFC3>oa0P5
z(lNOgt<-asjYK+XzoKtvQ#*gyUWdY7@hNBUUX};uKUS)&;agaQ`PB+?0JeWfPYlSp
zNxO?Gb&<GEfdv-xg1#voY6?ZiV6(b+Xeuk?H&p*$@P5J$0aGitW=c<uX)@+~`7b|z
zr>b-O%Soayh&h*Y@_3<KmVc+72)zUs-XmkzVW|4FlUDCe=)Z>ZNr5oEZq$2JHiDy(
z-$x+rclIHGZ++0|#%J?y2gwCibD~qtnf(gUl%GO&3#3?qcqd6*tnZa4DE1R8Z~}7M
zrI9y0zKs{XtcbdKT9|jk$TW+qa%*DgEE0PErUtI$uNm7Ln+WezeZcWuT`+lp`UvIG
zMi(gKLQw0)-PLPwkx**uE=!VDOYk%0FU4SN0<4IJAwyW_&Jumr2}MY|^P6(v14%$6
zwGd(V)K~$4V9cZ#J3_s8{etDRZ-h;r;6~uc%CUcK{gK}E>r*9tUs&y`iJ?V;l8xl8
z7#2-0>87|PWmsOUktk*^SE18fg|X@=*ZhS5+`Dus)zB5jQ_U9}O$IFjQUe~Is6ND(
zoDmh(>BdhDbCUTbE`-FD=g7-3XESv+cWp^5$~6H~4+<+4wy>BkEjFdUO;1Q4Y!ffD
z9t*`gw=rvB(7Rdrf(<RSUlglvM-s_h4>zPNr+*OED5JSfQ#TJ6s9e6V2-O3209E>p
z+9s|Mz#P#)!B@1J{xa)doG5tRd*R@xA%|4jfz}HlfsE@%0r(~>U!rT-c@FB@_rcL`
zTBWW+M>JmO)rmrhCRD-sxgYBi==Z^KuKz@0mj;>b^evsU-JHW4-k@Sz@wQ<-el4u+
z8QhEZc3Uwr{Y(ZDMrW8sP}~SY#Y;1=7>vT4&HB$GB@d%c2zkt!JPz^Y{-38{4dE)g
zs2ifrwMC)lmm>Z^r0DEa3-<yq$&mxL4w?TFeg1>%f(6i2=XBX*QfW+Tbd|<8U)1%U
zT!%1`g%=oX{A%h)wxXB2{XJvH6Pmrm2E&K-c&O(aKB>8Z&&-q!@MSLwkq{kM`Mve7
zuGHD=W;ju4hsfR02`d}8@pK$-ei&SENty>yf|yrZmGOc=uyPpE7)JUuL)ZR%rUxM5
zC?w@Yq6;?kd&=_mjt-Pl`dIES&I+m5P>~Id6f}3gQqiuv82}*NZO!eJ`q$S?R$AV1
zg7pBVqQTDzS3+%AYrL!?J9)w;XP0OtIB_i%9rlnMtRm%Vv>Rh7$Ik$wT~3(1R|DRz
z&JiLgV>gt4b23hW(FNqPy8F!X-wDSD<QYb&gA`RsV73!R`rGxdK@*1AEvN63|4t`F
z%1VYK-M^|L0+sH2JHrL&lzD9egp-8)Q{`eIh_~G5STz{_?@bh#@^=4>f(q1(_n-gU
zyOvZKFDIn=-F}JY&nbAe_KGR}V<ZwDw~dL$Tj!DBsM}kxYAF8=oBEbsI+Sn2KzQda
zs#Xn5#GIFvR>7Z&8sF$r-IS?$RXLUsD`tc>UmOIkon}_k6mIABQ-gyN1f1HUjjc6Q
z^1=yL0F5Kdt-Vo)z~IFKU=5PJ8dc-Et`1%d{wj?MlU*Ao|9b=_H#5p)7yzd&(A--+
zd&u^dkrT;=%?S6|OBy<Lb6Ed08+yJZ&9Ku?3g&$O-N9EUb#ubbliIQ;T9V?^Yka{f
z!`BQxwe!^RU41+FMxC1!H&J&n9Oup8pW=5}s-eT~+uIt25q2SlJ69(v3MLGH4`MdC
zI^+E4G_{$`nE-F8Fghnc{^t?+DA9$^J-MCI6z7<xQkC4l?KdMm55wx7zLS5At0*)|
zg3}RmYE9s~6zfa24ISq5lw19UZze|qcIvX{UA_N*{{pzNpn#nWB)|)w1#b+c^&-?S
z%om%oUWGzwY}+E;-Y(_1!=>fk?mVUsNn<)I83pu{JJ@xr*kKN&>>$0W4nZ!^`?#3@
zb7&pay<(+%--xpyad$qok}He)gkbzR6)NN%qGD$a6RlMtm~*#q1VC@T31Z`ZX}1_o
z`WpSFqdfca7!}-O2XJ*MFl{C!UV?ysSm-2MtWIn1+2+#OI*vtPCY;mB!#COa*vv;_
zR<3dO9r$s5^<VhdJ&(l(&%p$-BX<%Va0}-1Un)`1=p7Y-D*D@*`fnv8z%X25UW@4;
zS_thGNY+{U%Y&s$kx!4rfvcnmUa&Lep6>ItYBHxU8SfRQ%~glX$G14&y)iEGHcD}c
zzqK74vYLb$h(7iS!DlblhI530LXW>;QNt$!Wap;4>g<k5h$8u{*^I)AOQ$LM_ZX-y
zOt2(8u#4H+Jm+@7T5gBAnGcKsSaVbs^G4n$qIhJT@vCzNrO)5}*~CJYaF)iS%R*^?
zvqUa7qW5p~<D4_%Lc!0hyY=Z@5(q#zOQHLBS-o(|y#+!7sW#0oytVFdcaP3Zo4~gi
z?<ZQM%`qb$C^IZJugYtIP{t&PnRCX>@<&>^30(2#UBaZBNCF<w;t~JEXRi74^prFN
z0x~Ga`Lx`y2yYW^x%rNEY(Fi}PsLs1Lc#9nY?!T&1Et^@%L{?M(R9Q>2mZd?(8qVz
zE6Nwg)(|7Hv?a|Di5zD&h*duNLxL$_dlVi^(CT(r{stQTWh%T-O=Q8u_{8#BF|+Ww
z#Uc8Fx*~oj!A#(kR>0FB_)Eo~wsQsZ&1zK-V_%t;b~VFW;rQdWWB_zI$hj>N<&324
z@lQ3Vh<g788n$-L>xp^V<gxEHzC*3FNUjLZ?Jb`TieCI`^unoW#fk>aiFfNZxy8+(
znp3LP6fMxgF9-#4E|nzgRajm$8w56{gt~9g5sr16I!-xc-JTp`%Iq;`IcM}Kxo18T
za)ZhPp~=LL@7LSD@VjiSZ<b6+ya(_eZFX_-sU&?DMwJ2?i|r0XCZ71<GTEr~CA>w>
z7C8W5j^&RPZorIRrL~?0jpWixM4$psa7Pk7V?w(&^nSk^>cCux>|<kD6#8v8iyH>u
z$adH_=Q-X@g>D{u8$aM1gTJ6{%1oI*-)l+H!@^(uej?~{2bc1}Wq02^8&#TKHbZ#i
zdP~<TTa0r5esr^FrE^2W3W@vnZEpsM;H@|P&#dyagAmYhhDBI;X@5VlB(h|T?b{M+
z@=ph$7rCZCnHG#!18-2;=cwr4Kbt?>1)J?HG9SGx;Z})O7By_WI!IvJc0EBU#EXSG
z<K>U!=M-iyXs`svcisSc-oIIqihQ>TB(8-z(p$jHfbMtQb^YGHsl}~5J%`!}dK@zK
zS<bTX*zGSfnIzKaVL(D_Z9YrwX?F@he1{zGKp0>CBAED8_hTPXZtG+{vh01`<xId;
zZJqvP%3%KG9lvVXPkgU>GY~}yln|!TRr8eHxux^wFK?v7Rc|{-)<noH!-aevp*VEI
z0$T7jEOi^{qhD)<7mUvg`9`O*U{#*w1G<2=uHHz{_@J1d40Z&9F5zK{f96oP^VSx7
zx)d1xL(t!Q(}-Kteti}?6ZFv(IV)SLRjby6XEx~gyjQ*22^}wVCoYj#(SI7aAHj0c
zo*fO#v{bnNxFUPHD0VmS2k6(zy2P+qfQyt?$%3-^F433GsLs>%&-T4ipMi1?(COvz
z6LL9AMS$qEc~ft)HAEi(J%~5Q?AQ4$z{Tsk(mne>cLWWnZ8Tc0;aZhTZy{68cuh;x
zEwz<gc|Ouqkn-?Bn-l<JQs{Zk$mg@`u&Md-mq5zhXQM_Q*}=LRsO`-Bp;{xC<+nA@
zVp0=XAhm!|TJD^B^p0AXvI%uAwXrEU5qn%U@{NmUmwbBPl3)X}UnD!5&o2o<VUz3L
z*f!CNBcCeYFI|Fuf25dbPUUpJQY@3U4Ovc&|B$gf%#XNl+(&eTf`3*26<S1??w6BF
z0x}wgLcVIx<xHCa6C0=nF$805r>T!8Q@|1o`B>%t2S<xaeE)S7;H<j_H=zP+OL&}i
z4KE!0W%ycW{1?9o9dyvlL<`m1mK{!azBif*)ka}tOEE&kYMRGoPxU6nDBEg1seaz4
zt^PRSn-Q$81b<Nt&vQAmS{cDFv<piUd33cfG{CdPAv>YE2lVeQMcb~UIa2Gq{s*;5
zVL{Ss$noljLm5xEQEacHx51j!El(9hT5a`e@E=e}d{*+ohkyWucZI3pagQdqD+oVd
zCf7CJ=Q#Mu3C)QZ82Zm$Le`jj#N}FOEY-RB?@k^_UX6NM=Rz1K|2&`%?mk2#qu^UJ
zY>@pX!BSwrkh!Gv&l9~r_VW1kB0aePU5*L6&|Rj|s4(9PY^KMz(7xR>p08xx$6}Bc
z_*GK;u0F)=?E7e7;Aaj+m_IwW1-fSV!dEm%{G1<MI-GIc9y4yY<jCy^K;672PjJ+0
zG<~gmAI+N}5kn(}4vg^W=^fdEg;Ht@c*BDdB|Zo1wnDwhDvL`P5Z}EOV%Kc8BVX38
zQtoK|w8r)nqyhM93;OrV=Wet?zWIA-KNf!KWJ(>{OAbr@67X_|i2nAPO!K9m&r>C*
z1t;+pn&Ww?&t4=VY3owyd2J+!PV<QXU!;(8HK5IXi;c6LDS@jtIx=XVgPdPCJI4E_
z203x(142F00w0I-A4~_{Dt0TLxMI*m;e}%dJ;8_BdabfcAwaE7e{@Nz8QT*wn4X`P
zLn|^zdBRNX%)(F)sY`SWZ;niXSLW-zS~jZLD3>e~LQH>M^96h^-!bC{)tbb~x;$+z
zyAXRyYuQU=vOcAqI$3_oQf#9ZGl|=?9=DT+`GX0C7%cQfF6D>Rf2KTiDDc}OVA1RH
zj_amXZl!+?Z!#RXcHtSTzq^%J>Lb->B_OtQF4y;2tSKKk(>;H`>Js^k3i$GMkP<+U
z75IeNUDPtVF6}z-n)SL{aB;?^>naH08KY}g?xz?FVMQxLOy7)jyffOXK<|j{l)J>T
z)a6y&9zfrp|HNcBTcl9fDT&Tzey8h>n#YJEEweMdlt%U2nqT<|CbB2|gAVeG6NJlj
z%Y9Ax9>`-<0GoUay0p-EY#1es<xBnw7<Q|EMc__<{uirwMR9h<l~}wNA_<S=q7+Po
z7cFFhVN}yz^jgb=tI;JTpPSuuxvY|OlN|LMq22{tkUGYSd;BL}5W88b{7*P<c8VV=
z=_=!FQ7e4<imk|J<sxZ$nmp>S!5M#p?e-b)!%301c82xF_wfb9DA3<>&|j0e{CTDS
zphc8Vy!To6Z|_yujb_PrpR;jp0rUR<QA>d=vU`@?9dxSP8LR>nCcFUuZ9|t?L+UH`
z!E{^?gitzN=zoz3j<V_Y9S~f;yCgdkbb%2OYWc#RROSEYEc_{ArPkmn(st>8TCn`x
zY`G^Fi;?3iId4$AcD~#8#ZFvhWD8RUgWFeiz89`Y=(BhJGd>vMpinWoc!i{(k{d!k
z3)TlGUAVd2S>y*q``q!gCw~pm$NI3U{=?vU9r+O&mI(o~^D)ri+dIL%!pe5<J%t|~
z7-ssAKqW!DgkWPHn+zcG7r2Jrr+iZdF0@XTp7J}mI31Oa9rBOlfgv6xoYTklHvc=@
zQD)FkkX|#wR<y+t(5Rh%Fm$b^q{<)s#|v;lagzI`+iaPuDWqiF^4H(h;FW~RW#Iy}
zz(bqpLZ<E;0@OI>ut3JW!`S3?O##wqaY1nkGwl=)c1!{3e`0SSRT5slj|o(H%pf8l
zvGdM5B{?7UWfI+lswZp3RJ6qj(XsQ1pmH=8OzGPU@V$dtqpf#Kvvr<qu<0KZnOzP1
z3w<o=dvU68L6XH$s^q%vc+Zo_$FTEw#uM?iExy<?*ZDIt9(GQ75Jdc1kX#wDEHeI!
zrD~p}ZTFUNr)_db*aasVtzLJL5y~|&q%qcl3R@A|iSjGRiqxI@YKS<{h`=fTZbs$J
zV6~*1Fa7IAWbvVMfx>Z7XpqhFBa6(4$r8Ov?p-=Ti;D;uY{b@|a}X6ExRXwn3^x3S
z9LNDlmN-FTE~6sHx`>oj^j+e_fkv&?%}&?gD3L9Bf;%X|P)0=|&tRc46cg=vwn-MY
z&}~f2UnXg7K?to_t}>PD)6aPyT#-LgCKCPZK)ZY6nHpaUX5%i*NR@n<1tjC;z$*gT
z05q!ADha2LEp7b*@DV3F&8XSH*`m2g9+3h=e3$E#JZd~7V7D)NfwLS>MI@Qg);U?a
zi-=DPYDaNzJc9LH_ZsWk=zhe2dlke4nnYM!?9CoMI92*+br-ag`lHMCe9cpxozS?v
z2=__*X%2f_NY2+fd{$^OwgC6+mSc)fc>K2jd`TmgT1S1D0))GLHtMS1o`l|R1{y!u
zM0u4IReG*{J7(xo*m!Xq1+?BbwBNA^wA3jZjmd0)+<Pjka;m!zj9@>f`GTAad0{K9
z(BD#%GSr%oGmoPEuzsnuNiLp{BF0=iji>ny2Us6621We*pOC;7{ZXKUKQ~#K{Nvdu
zFn{{7*@h|9DZ|>fa?-}>;vPDiAA9rIQK~CB9k~`1(~jw6YX6xC)~6($P0N=fhC1)h
z%*OlMGdE~sg4pTRhbdbSR4Wac4oUR5immM4Bx+IxqeZyhR62g)M>EoKy`SuqZ{S|V
zcu`)Vg+sM97N29`qCc<GsUi5h!%ZCk239V=6k9^Nlbo1fuGWTc(M%Qpa3`A-zGH`5
z>h9(^Z^sN3UvYK0=|e8OB%LLwk6S$t+kFIBpr}C-Hr8$P*AH7r4G&<5!84}2;md-8
z6Y~dhyf24lv*TWw&#E`gmOWa9aHXw)MMq18<?NPpIJ<}2$vf{|aZ#Fl(E-&(vpXng
zv<6j4VeN{aU4jD$07pQ$zhY2MZsHeJS9r@}0Cr~mBYX$u-?fLhyOYRX>;j;7jzz0Q
z(HYb3ab!W~{^&bXt1UaWt5t;r=LbfhnVJ!C99K`-v>RiBnOmVdLJrpaENOf(|7-R8
z$w)ck2dd&CyXOCCp%&^RxUMYHfp{_6cd+VQ_|=0mu5&YW+B~nK)|p%LD)KFzu^|jm
z3Auui*(`$F>!8P?zfsX#E>DAmio_z^l=kY?Xu&oL=qt<HTH7hDGRMo!gxZKbDSVG^
zgAsG+m@@SiXln@lb+sac6aywbaznEQGVOk|uT%t%<{P!fN%aQ4bivOrq+Kas^80Hw
zRB=5Urp&CtJ6m8D3$i{<5}I5!GbG|cug&G_uk}jW0<ZoGmpa1|TtQbp7^RZSAy8t)
zCW!@~$v@-g*4G@S=^NMTPOL>W!J2zOa#u#_t^s8v>%aSd@?k*{?dxt0Sdar!;yy5h
z<#Gdjd)3`rt>m<qU2Hx=IsSXtd|S>plfe?Q&w+tV1u%rodT8sY*f#58?P4g)*m>oS
zIIa}R?@2nB?2TGYhi?(cooW2Ec`zEQnr}wUNd-+y!j~`QlA&?O3ORM7-nR8eq07E?
zKjNNCq@Rls5<9$k<o4Bb-|Wr~_S*JjkkJA*@yQ;4k;8(e07S|~xyN2L@LiHZNHxR&
zcB90S_cCvQ3k?MntF}Nh|J&|dit)sAG6!wVbkogf{OS8G*!I+KLsw?wU=sUM$@m4<
zNps_?`_lP0_r3uBjtB2LoAZ=9{39W|LSs+G%>hVH-0g2I<INV3#K}hWnwYN-7XHcB
zMEQuZy!{)C3U)=JY|E|076-LjCJwZQ=eN5A*W1AiW)*pTmYtWI^$aK4j4bV_D890}
ztKeuK2ormf&&P6pe`sKv*_LQOBJnB9d9@1j->q|(oyJ;)fE%x8r=h_^X1#F91h3ZC
z;8>r7Fsb|^KG&mD?4Sc`hJqR@$Grr<>J^y-%Ub=UO2gmW!)WJp73`?!bMwpin)kn&
zc5nX^PAh^KS^lS+c3IzBmM*<E8WG&mB@jdb{5_RYyRl@}*yW3-nDpyfUH0LSgoFxp
z&6cE(72?Nt0yT!Gqj2>X-0>Y;+u3aOrU`G+*M@&O>0Cu@6z61k;<&HK=RdO*ZR}E9
zk7U#VnQqX2^qBxGt>KAKEkuahcxp+Rr!a`#;Ti>X)cH(;Um)-Cowjt4LQOhl_c>P6
zCI}9@28eoj0Pm!}9E=W@jvZUb5AU_5QrE?5G_zwxu(t&&>*gn3$xZZbe;)}A9VgYk
zdkNq$BneHsc}h_$k3J$%celCOV?jd8Qmnj~_lBiPrF`JPIDE)$%+zI((_<Jk-o+06
z!@(Ww_cFa@ENyE>1!A<!iJdQ+Jipr-%+A{=a&H5p)inXo2lhzr^F0B#m>RZf1rn|f
zihl|gxBaH<W=#oZ*A-X2$(=fHEH0}{33?{r109A@LWd+Avkqu^>p!4E82#_|+~Anb
z-M`qIt3yzrG83h8*$y;W-uVaKJpn<zgPC!giUK$3EDN)2Ef`=k=R{0NQ)5W@^Aq~)
zbOa6QmuMq|>=%)9w-9t0NqYJ#qzYgxWV5F^JiTY!@nm8ghc8h{n%vg_F!ot87&OkZ
z+ywq~0DdWMZ?D!ehKcx2Nq^VIK8+4&R!{AU@uLovuV7LS6*)0Uq7#LWdu#XejmvM5
zHFC5_CSK@_v6z~sroT85H-3J|{QXLPa*30=4skM`TBh2-F8ey8&-Eqm_o!8ga;fGH
z9^tth&?s!<oT%9EFP6SNscqkO{}kxSZU^ctQ%IlLUQxFRGRrQcSg$%p-xKv*((e!9
z3h6AFihdATokZ}-8qJeG=qqzJ*?EYwgA_ORAPqSI302&!)QA#pSVB}>-1TIm+*@o0
z=9AM333s&#a*uK}@93(oUXkA~H{E7C;$;ZqxXv3opq~8OIc6;4iu|`S2jQ#J-i`LZ
zJ1bNwcisgXNS{a7>Mz-E_ITxjvj52h3q*?av~a-v#dQ#YfG_Qf$7vK%+C=4q!1lhH
zB#8non;-M5gNR^@6Km&N+H2E)KWj+V5%G3i`V9Q#SmixslOPj$fxs5v;Ev5fnq$al
zpCgBJmS*R>h_Uqk7hXjz=H#?H1)o;sE@N*N7z-6Ie9kTrrXQKs{E^xy9f0QJJvuYL
z%_+~ih@$EZSPO##&o5PKCX7K+L>YTOjqBN`YWT#uCwIL81u;3O4c5~?weF{PGYES2
zIsN9{B3t|?hoq{<1@ZF}(zm3NldhI@gXZDSnK}xT5f`Ye0`?UQEJP+Aku16z-Hbm{
zv8*`J9V+O3boMh$92w+8=FWibJ~6t1BBh1@#GhF<;C-wfa`)>64tx5qvTBBx6v&{5
z?MQ;gIe2Z$HEs6t-LHt~LhceWw};7c=4QQFaXT>{^g0-P<URdyVdz7B&!}Tpo2Bh3
z>N0`W=r3d&2`8s}w5%_Y$hg8{Mt@ftnVEANPOKe_xZZ&E{=nHEov&i=k75$Q0TM9D
zTE_l#iu_r}n`Wxj?7LCXuShmqzGe*ScO7-5t6P)&(i^cNtt4eOlzaDYsAoe)Brbr^
z9!BA76j#7Q{H*H}3Xaptm$THPnO4CQ&AUP8+eIWKvIAS4pCR6asSGDnP{Y%@ZE2l<
zMYhe!-+ZqQ*}7(ej<`LZ_t9s+hm9$No`uZTbUf&7(7kQ0u77=si~B)xlo9rD5@Uyk
zU6srJ%?qVU>R9_tx?WEio0)7*?#y_m+F5=nAUN*$nd6h=97s2&tpn*K2h<(3)!Pzt
zcez1$?vM7}V5og3>u{_2zCx#uo^#PZgN^zfuso0z9=7J^i-X_iTsmbC^TJcnyTFhw
z>F_?!@z2OEh_NaBMcFUiI*w3zI8ev6`gx>LTJ0*Cs7M%kJPX9{f1Irwcl5_OQ#)gF
z7YO%O3ItdjhEVG-+iK0Xnr36;!Q#gkaepQ{c~s2)R#>WmE9!?7M<3gDN9WWNx=5tS
zhbnvqV-nOdG%~0do6Ps}4UUHWpSt>fNF}gqS?aC00Oo}MeecoKOOEsEHN8PxCL?@5
zy*L|tOj_BxCEM%%+I`#3uWLsAXR{@!$-4uJ&i(wcDz5?f&^X%CN=(9vhn*UfdfTYu
zTz~0DL6q@Zr{PC{wkQ{lp$G1ePL07PK~4O5*BZe^bZ(e51vLc)>I>mE>x6&;3<|+V
zR4Sn8{DM@Gxtw`fa)Tq!?$3S98QYrbv*emZ-w5^9G?v`d`7OnJQ)Z^(H#D%3E9~Qi
zjjN1tJ#;vljp*xU>p$xbD&NEzGQH_MENGb2t38@K<Kp=<sR1CUc{vwLpCgKpnXb~E
zYv#_M(>@%LhSwDCiTVGSI_sz^oA2#Q34+qy-QC@tN_R+0cXvsb2-4lDAl<^DyQI6}
z&>io6e7?W6UjMsUtaE1e?0sFIJu}?;%a&a!z-VO_8Fy2k>s*#Wf&CG(Rkf_YTv^^s
z@D%U$Cy2n0p;pCJe`7ig*IU}%=ob><MY~2O_^IO?A`T0FDVv?1QR3A;nZ!yClL?4B
z{+pl$N?wh$@=$)^!n%D5iQJ`-_q@u;kY-sd!zyy#T%UT!=3S~=%)2;OgJ{{h4~H-+
z|DmnD3w}If>;0KH+;l8%3+q}Pn~+F&oQF5TTHkYq>HFqPOS!wu)rg};KiHDCzk8%J
z@lP;tV1KIo3A~4Clzanb7mmy)4<Tq`c#0A^<n7t>LJ#-a-oGmWy?c2z2noZ)&t-V)
z-2G<4mRI*36n7KfBKA<e81~C*RcIp<3Z@-IY+|g&o-waW`_6V~H<Ja7`?8^K?#qU$
z5MssTT8Z|euS=-iW%(|RU#}|=KPZ-|u{tP!TCS2dF)O)M;YhCBS9$A??II5*nnM+n
zY)B5tRJqtzKWV}WyQ0BhI#hc74IR!&7Z#CM6Ur==qe{ayemqOkBKr74^PxhYmqd38
z8Jea72t|?NZJUd%-wI=u&FTuOo`^nMhclD8w6lpVct_e<Z<~`6S~@;(<MojJL64M%
z7Hwbwp%zERmYBFdE+!=G_vF}=`eQ`U?$h6oB&oGxdU;oo6^5!qV~L-HSePg=$1Lg7
zU;Pf);A%&sFO<EK@|mSG+RPUAB(!!3{X@sFws`ROe{UTL@kp4)D&*Gqdk=GPtB_)O
zIMi$TcsG#<?k0t0)rXfmZrYgpcYM6Ny-umC?uk1G#ZAm}lX=;A5QlmE_&kuU(Lh?N
zuXRC-Zh3jh<9qfs=Gr`}UwO6<@oF@K??%O-BNinmWUue<Bo4C~il4+e{!5RV3eL*8
z4-`o_^^=HNd0OjSEvB+>)yOyw;%i3-``pfHe)~dAjiKtLCPono!P+lNHNQss`IO<Z
zO#?~|aP#ndS2V6+QGIOwyU9jBFxjz?uOOL*T2Y+m8s&1)zD(dN&4;tPX^{o>0k;`8
zS3SSt?a6tKpmFH{Kl4n`Tf49+n&13#qkFlHUI(|q0#5V49N!1Z+uwie?&DWZ!66X}
zGtw?xFhRBKQ>R__qk04H@U|$}qW2RW-_?gb!+}IuSax*iiz;f06+1qy$UrFaE@H6}
z0op!RG+NUfJ;ysRH4LBHjXEiX@btDmQk3j96iU@m@EBH-#e|6eE8c0C^t&$<QR6nQ
zpKY^6tp|VDxxNzOW3U%x!K4U9uTL~rH2Ln4xz=!%A1ZxkRz-mS9wRfySRVGKJ>~PD
zuc!br_lZc5Jn?DmSOvS_AJyw<%U=<~|LsH-(zs25qv=ivOT{ro;p21qi;j5me6n0e
z%fFk;WA`R9RU}+rZW@2SA`<p`>v1X)>H)`s$oRAq5FB+2+2Qmaj>(b`nTsqb2xeE-
z#O%5yayK~FYEMo^dmX4U*1z<@1~Up0GS-?pw&N~8oO3G%A%eiLy%D8>^ewv4S%Y-B
zQ|s!o$bKQMRL-L4LGObCuO`M(8+DXb$&H&3%*;@pw^vhSB<r?bpJG35$q7Cxjwet3
z%x!n?s%DbJ`ljur=qhHbZ@J+KMI!vHa>V~T3l8zs{`2G314erPiB~3OR`^VK`+m$I
zVCZ;Op$shNPt%z#&Uznb%_u^=Ql{vJi-rwa`7pay!?w;|df^E>U6G(vdRXm@wX{D@
z3=4Dw@p|5re?{x+_PUm%Ms&<jzi*NDIy3o7E_C(byjn7TFiiwSIlDEK$6<|JZ?3YZ
zmPqL6n;aJNc4$b3<`=r@_T8SYb($s@a9$~>+e^gvsnl)MmG6O)FXvrA3{jrS4$g2g
z@q?2@)}+wYmNd=FR74)y84Zf(_bKJ?!Pzt~nvoTCr5TNyAtxb+)&?NAZdXypcq7So
zTl^)_s%>wOpN9M0clf#7u@!AAl$z=IIev=t-x%7PHOeT<kQ>(n{btAHsxBW}r09ca
zV1TjQ>r>%Mg@z?e-F6|#ab!q<@)G)4{cw?{>+vdwtG55UVF9n)@2x5PBGrF-YiVO%
zZkelLr4jtC<Aa-FW&Ft)uyRVLxHm`PXEI!*Cr+PMRk8M?h1Rfh&OnQoVxA9&S-8U}
z8iXH5N1;oIg+Ypli;scWq`dn~(Ysvmpp+zS*!|{20X=l2^zK<Ow>P(8r<V-j^}VCj
zfuw%N^PA@q#yGaI6{P;a`&$h9Ud!sPX&f|d_Tm~V^!o><w@P|ZAeEegbS)>&8Ur-7
z)E9}i=Xfd9CHD{IRO$_tkUI~QFhOA21>2b;o<Cv^@uo|2R@R?u9*VPkOs1#!FlDut
z97{AljrY|}@138-hd9Q2zWlQB)vR3miQ$r3)6JCgSUZv-1N)nZqlu;IrBDY6F8=jG
zCplDl^yp@db8<R*gxv!6E-QF#u!!<QkOl~BGMOh(+sN712Byboi2d>@=ELr6|8q21
ztU_K{YJuU7j01eZb56vWR?Q*#M74cUx6PRGFr3lyYeAKL;;PMDAC&X8Tm*e}TH`A=
zaj2v*b8NDCfA!JfktRM=$aO=i*Z;0*81q7wa$^~Z4?R*eTv8ODD%oR+FX6a4r@r0z
zZ|<t!sK&;-oR1k`_=oOEvr|6Ai<JLPm7y<R#H8<p3GDw&h*(`R-3j}IILA#Yuth81
zZ8Q=I?;o6yHeGw@qll%Zq}Dad=+t&+qeP%R_L;fWBu9u!+Y!aKvymnwiom~h{Sxj^
z)Pdc2{vc!o8yae2cFV2X$Ylc!J^C}^<ae*7(@P7=uETOHKbFu8KQAhHN9ot3-HffI
zWcyo!v-<%(K9hc^2675{v+Y+tiO$%5e}}ivcS3Kk_Ve8c#IwWm{AzN}H+n&mlVWwC
z*B{YZtO0$xq2WT)y8A00+pzq%L{UHsVmf){)Z_WORL?h%0}x)T$6JDvQADpOCr8~&
zT^mPfSCivcy}@=3^qg%n?H*>9oA)-P-(XeOmnj;FO_!uZmVMBJ8cBo?G>Gxf->zVj
z6nNacUa2P7m)tN|*L8(LUs0>$fBV^cbW1Wjd*3m(><2UQyjC2{!JqW0b6uqzPKtZ|
zGEcvPU{Eq2En!?GpDg*qe|Kg8RvU!<$eju~^ot0*dMi)t-%@pNeo^vazI*vL;xzmK
z8|3|er@z-LL%EVrE}y`o5wEPiM|R7egiwaO#v2AusW3eoU#d<TULi-os3+t5{b5q=
z<<;26tIy^oPi!FJ^X|(~^m`g)*qGsHWq;&S`|IaG<8Q7mlJmQ6n@zV5(W@PgfjSk9
z(c;PDL9EyF-<o3rpK>^kUqfQHX-C$U)~*TSF0!fWQe(~{6F)4uZ#<<O)gz#~B;Iz_
zsW}2rlzpu*OOKMn8{JpIjS$ko&jCZV?LzW6`hDR~oRX@1HG0kSkg4!p>E2)UtMgxL
zKKF9p$fB|(ct{xj9{w^~V0f?ZhtCzg3p(&2JAArBCu%onC?8q&GtjR&FrZN4y=r)C
z`UmF%k9{+Pp@bTZ^&@PDuOvAPoH()q<cO32cC|Ag{+8t0r>b!D>y$@L0P%SkiPw1i
zm+kMRk!guCFALOR8`<OPD{*`eIzN4p4L$Y-8N>!TTHTzsDOLQH-qGL_N3`+H==|jf
z2qV<tC~Ll$g=0@7N2o2Cm$?|=#*U<{aXR=5yL^izExB!Th0lYsc9j!YhcrhfsEQRp
z-r-{>4dobLlSv-8$&7riF+Q<Fh{?lFZ_;F8l{;m1;Sg#1^8AQmJ@c3fc3*5Og1pyd
zuoGF?br14hpo1*4T<|2Hv>q7IoXu;S)a|82CFg8noJ8lRnPNb7e2(xY*CdhLWPISE
zpZ;5<`FQ0`7Xq(E)7P)k+P7+>+63dPJ{bNgMkb2U>}-d%ApdwOnu+wbz4n#MK3Wy4
zw`xa<YhE8D(!Ys!EKh<D-r2DnQN0hkC3)>_^#G%5>dYnm_g6omft>p#s0pgEkgH*w
zu!>4b=Oi)1OT5;yl%2D1ovCRZufpdusiMm4#Wd2(wu;@Ew^8TWLg!l9??h9?y*{N4
z8Fs#g65CYgsf7iX?q~9ZAO1UoYkThApdTB9@yI4;yb#Lg{QcB#MahvA_olgKcciRc
zf&5@-s>3{cx%rG42Jh^%#1oX)eh$S@W*3dhIXOAHTIzGi7lW@ZkFoyD%L`WnPfci^
zmp4L;Usd#@Ovn=$wCoXQ(%7&celS}u$24nmyO1lo6j#<&$*tfAbL4<87<6%u#h;#`
z;&XC6=8?m_w$!}T6)nFpG`_r9FI#qN`8W>7{F9%<-2df$h~VoPUGYg%lt?5<*z$&v
zt(ep4$UL;uE8eJ+j;Mf<4KMk(0h+o%i{WGBB-v4ad>73krMZ3e;wWK&osPtR*RtP3
z?tHQH%0sT*iK3&`HZ(|C6`p(I)1ZZ$@%YHO^0m(yN_cG6vK*ydkyQGR8y1`-MdOMK
z-Ydy37~{dsn&UM;$MK?6?2(VP1iYQ43vJxa+T=RGrB|U?T}{#Y3ch4mu<<-O`kgp0
zCqo*Vjpy$kBZas<9uLYJk1^+$%a*C|)~K}@B7G9@;6!H02>l@u(B=1PT)FDRJIHU|
zn}#6hbp!qTKI!xvSFK<~P7d}0hc<9Mp@gn%iqx&$6E7~V9lpS*QMq)#X!EB^Y>@q$
z4+1e)#%E9mMVf#Yr~m#0KUWKzXI6KfN@7ljO(DR-bzT>T?Jwd#GV&4HV!UNmZ$Xp4
zG||2>m=Q_bc4vrbd!BLcx|h9Au9RE*v_*m)NtF7`mMiwIp;tL(%uG6p^el9BcxO)`
zN|Y`Il8TX{qz7xLk^7RG|Gn&{j%@wlAhM!EWssnTl|7)?J+i8LB%Ub9fJYql?t%$&
ziOu42e1pgi2wvk7JP0fT!p)zKRH=rS%3Ya?mP|>;3vfG^%p%v!AFOC3sXO&WMcf`^
zj(N`Xx|PF(&hA9>t9tI>lt%}@<h@|kJ?kPEqCYEB?FpuN?VG%>e!lA`9o~{z$2hMu
zJpUl^gPpC>3beM!YV+-eKKhaeT%Z}YFIpYWH#I{vo6Edr<u>tKOyfz=kIvUOr2b+@
z>H?M5<C$3ACq*F}{V)G4Dy=~ar^Tx<*X$KR)MiyHN|y*Nu8eY)#3*)_Rcg{u8vp0b
z_k_YuZ>ObElx{v+GV4a>Z58L4RDHSg8)5A(%dsswQ1AXb%rnvMvF6JEFhjyE0(?H+
ztGMnV1zrG(z~I1`n2XBt2eIS>pV^s$|2ty9cv5|$#(y4{$~!)frfnlDm#ThLTgflg
znqq!YF75A;XtWk1#pqzBu)afw)n=zR@WqXyyLvD8-`~#|p+i|uye3%<cR;)g7h!ua
z8J;Q%tIs6I@(YY}j66GqA#B=3u?LroTl@dF|L=J|i{*-Lk_evoJPG%0z#%-vY=mbN
z`}sP7YWmgsV7K^%Wwk5wH-Zn}Hg-4Fzf%Ruf2RuK9MdQIDBfZ2hH$RlNvVTU>$sl?
zu-sk3ii*sgHSiQ4sf~D~z$7i77jBt8SST8Klk$+<nX=-6A3bcNT2j<S5Md;;H!zsS
zuxfwbP4%)JzWgTUi$D0ahj&dleYyQ*t};?Ni+}TPbZYHD@)eO~c$`=yxjpVoMTK`-
z4ej@VFb1BF3xD7BA=<z<sB6hfejXKB3S$e{{2FwpQ;;KB%9>uz(o5mEyJ>D)X|QCq
z`)vxt$=yBNt6H#0RGrmuj6Ndn+o{B%8rjEsW8$kaJ;e_)lyD85$B*xsxp!%kM0lbw
z2VytjF3X2&?)<`X_i8@Hoklk_NAinw{P~U4VWp`5a$HJ5b&c924wdk0k254gZ*w!X
zX!UF1zqdmbkdlh5a13RAzx;-P%L?88fZWv+*;^<kf_FPG&5U#Rk&<$gND~!am|(^x
z;HKVE`ANc~r?$|v3|G%VCDBsEXVa%u+jx!14yEdgJvQ#&kB?N>nP3H8$d-9d(;o2<
z+Dd$wwI-V2{178U`uC)I>>(k!vdjEmdHZ};`toHQ2tpd=>fI5q&m_n>t>O4P^7wz!
zI<80wZL1uE-07Y;ouJ%NcgB)yanfq7V7v~@#eeBy^*Of0W`DpD(0hM0cc_3eZuKDG
zbbxJ$#^Q8+I2&uB+|P}+dtpCYj<w5V6m10mVr4t6HVzstfW7+|Y8{UK(m<TG#EnCI
zo%w3ex#8&K+{N<jLqCh|yZV3o&;&LBm8-3c{h@44*E7np-j4KQn9tLhwqWJp;JElf
z?cgosbK4z!eOV`of0YbxCRimz#MB)@PPS1~_f(3qsO2t}g$|cud+o*MwRkBW-BS{<
zcdih#=8PxwrCF^@|M@CfxO)R3)-yW2k8<=I{(!nXBk#qV6(4-ChX#E&^{hO^NCP28
z`sS-@7{o{gVHX>0A|g$tH@Z$rzHGT%a1PBG`Ni2t*Qw1q@5xc#p!JhP{-lal9$Ail
z*~nl0=65@Lo$_Zc+x-0X*Z|H`oexK-sa22e^t6LFy0w+*6!_ONu!)4!=jY_%5!Kpy
z^14@8*rse*%xzbar7!$&vB~FkxxKr)tUcIMuJ#Lb_!r~z+I4CvbLEqc$bnb!+|XCq
zw+aEj;M@x1Q5XEK`A!4gM|xi$-j`;!TOyM&&ye+o7p!%4EaY%dh=;oh;)o<tNO$FL
zWU+bfI-Au@ep1%?CX2@E;pjuDa{Qb1_vIPh0ptAY-n8N4?OMQV-uO(>g|5pVn-Rmk
zKK;&zo0AB_;Er6Kl|xzFH4adHGrO1&^Y%D=<N4xp$KIQP25`z5vHT~Q@c3+gnxT%A
zF-NazO2bTT<f4gyrz^hV*XH*!al{&^`vq{Yy7jh^PNveE3!^0#%_~fp)03;BU3W{I
zqt#R5_(qFL9C%|!ZI3Z0d*A|JcN&RG(0aE`ym{!V-3A(tUQQM!r-qDud)Fqnot;5L
zwF*Tp+601UWT580&%4Ai$ip2EbY3Xj;T7NIr$%pcTP*$K^Zf8jFm;IEvr~R(Z^2}d
zV9YNtxXkhDhGC;}PBMeVIni`YK=84==h{y2LIo6siLH=f0UAazIzJn=G*iUYFrSqn
zAQROsv}fk2_3XnBeTUI6-LR;_HCwLH+t&PRGN9l2y#r1gHe*}#1vtz(_|<?!u)BoC
zC~=Aoswed2x}3Y+$mUqce|7lKeU#+XoLFXr)pzyFpRcKJ8otar|Ji^qW1~?@$IRk)
z6YoWCueH0^KNBKrQfT9?JKtG_L+Q=ra!0r!ywXKm>Ck~W>-=O-v}P{oHgFbCBJw9#
z+M(WN9+^(H&{6iiFZ=#iG>XqdS2|YHEqv~`^pT}#l>{z#h4wO|1R~y&f>}G~jn+|c
z)~=O1)}I)O?u_JuIL|)`d9YJPn|gudlF6;RUUi>PpiXaWe%<j>BO;2NGEV0MVYPyV
z#)krN1R+_4y7}Nk#RfK_((2aJKD@^buI#<tyaqX}(id<*jfW<AY;5>=KAc9UIx0Tj
za^vq@W>YY+F#Jj&EB<(#L4R2z=uh)_r=)$VrZ~=|UR3Y_p$#i8oMBuj`n3JAK$3Om
zO|iyddFQ<G>7gBN>pDsAeu&4}+23L9mUAPN5LabPl!GLi^V|{3@VsA|x%SO#g4@?)
zgC*lel-2WK;ctHj=T$O0{a!b+HoM8kudp&|z317U+8;<FmZ8uwY3dw0{lBKtz9Z+i
z?0$XOyLm#0+G0Jq%-bqCn_putp0pT;kIjmDm}N*;x>VTiW0<FEA8&M6Gv_Ma^DkS<
z9-6aW!D8F&2L%SfoRsOgucXpzl;_F~bIbnqq7wFZYZX3j>2;rJwgd~mJIt)`yx2|d
zyou7>@W}4OsTUHyt#P^UX7~$o=n+BbyIM99AdaDbWz59XN^gUABWM&IP3H=k9Ju<(
z35_3=8ww;fe9_-2zgYxW7uZFQ<FDB(Ex8wMF1cJkv5ydVK~aH!CMmjhj!gx9ZfCtM
z5*L4|%X1DLbt3w?WHnts6&#xq@8N#M;3HHVA5EEb>O{>A+29O$kqW`Kn`pLaG<w&~
zipiwQDd_)%Yl{66;)FE+winHtQ1vvXj4EvGK?o#Bt<5V?>-qj?v64O6tn<}g&04KD
z2r3i>=PjNp|Hkf$%!OdUT>V!9jyrxLjjx;}oLE9Mj9shG@S-xuy1{Ru3-h#_ow1Ur
z-Na+DvqD4mb35Op2>4*PVPf(#>v?$FpvT$kgxUU>@?oe^VK?aXgRhu#5DQxVi^k*5
zn3nxm$m}zs;9cb?$G$mGcB%c_wi%wD+~>Ha{n~mWA+OY^1iKs|jH)W3!811SSM7Hd
zObYZlVvoLGjV~1w<LR8(-0m!~wRwIWE(ded9zWx~RR2h2(r%b|pI*zl#ogv~c9%V~
zt5?sy>6MO|3ABQRL|s&ZG_=<9oiH^^(AxrgZ9raXsh}}H@x4uny{ye&<S{XobsMZ;
z39*7C8V&P%(~tHis&Kroj(1pD1%2;OJ5!m%Z0RsQtAoNxQ`@j(169!O&JH-Jng0gH
z!jGG*1%>GT0IR(@p?saMF`6tjk!V_NixZ(imO!)|Zku$UGGB=mmF&7+h2cD!LV1TS
zU~(Uof-E+_ZRCh@yoM9KH`+fU-JC$7L-_hddcFb;fgK^U`4YMcn@Hj$e3aF9yOea=
zVdYb7pfsr`IqH{$gmpenA$rJTDR}}*fWO&D(uXW+=Dwf|mPA&0+6^-ehO6L1-$VaL
z!XOL}L({G~0d3zyDvMF0P-><0hqfeAeRBD4v{o}sZ0@HbF~}j*fv3dN)C!D@BV0KS
zsM&LGjvkIYNLVdLxj9co;FQZaOs++2(74KPj?j!TjJLuEa*W!&-P#_{(3F<k`+Gqu
ziS#<Ef4&lr_A?zE2D3w}Z=>W05)eInkhuI_#cDlerc|>7ddH+&$(ux}5X4Qfw>grP
z>KKJbe4a0LWm?tJBu%<c=J;}|@ZOy-zO2>Wc%;yJf1*xpf*|y?PFj=#^%!*m`=Q%j
z{yLBox*I--#i+yhrOhji-&%0B%{yp$@H5$`?SwnZ%InY4g=!zm@0YppoNCY-_tcus
zzlt@y%7!V$k5p_f_mC}ndF!MucA35Je7zt_?0t4twN~T^S325l{uFb9zooHobY=9c
zX08}t{CsZBfBD!zVtCzM_2J0$yalCmckvQes2K0j^nFI$Y?=k%Q-%Gk%yfa1Z)HS&
zwTc&cjmdh$`Tz<I;fk>5D~a-|SJ=R0eu2{CnAK`N9J!?9abIOL4$J4O(nLomC$U}5
z!NgA=`=9P!J~(}re2g?lNH7%Js9>e0>5Q^WM{M1AJxb@;RqOO>Y<6lb@uX3n%^NB!
zDDvFmD!RC^qU&Ufw_Z>+D88R*F5<RDbErmkc~-lhN%T`Q*-Biv?2IxskjHfp40t(Y
z5|4NnjnpI3^_PL?Q{88&<_apcV#V^W!f#<<vxWVl-eFK1Zv?}A#i7A=HX93E2<AQv
z*d#_3_T^<0fBEetLRwMkE%I@GsD?8Iv^9Cs(e*QiC#v~qfyJqENnc=YGxnJ+eDu0E
z&Hnvky^@RDBh%OiJ&lON_-2=zr7Mq9cM$}ty@(Dm+M3n2-a!|f^Zf%3PWu^LgBEwG
zHqUdj^cPbCSYhA0K)mS^)0>#GyYp)fXX^NnyF>{j{7Udw%F^ZcjTTQJpUZI%0w?%X
z6Jk2Dd&XU%=e<c>5PR``ntcpUxLu0HF--67p9LeAgW_lQ(Jc+H-CO1CKr2v%zvnFc
zw5!eDZ&@|B<$!xAxzEz>o#E<j{P>`vP0zy8qi^$dO@c)U6e6LX@xE850?Mo2;ix7(
zc~TX>tr!Sxr}rhJfgA)R10J?se2v&b>{bkTIztc2;Bnae6N%B}vWI-S(i+j^wB>y_
zp*F;?St$I*tB}?_;b1bYzxE-Q6UX8i6i3vXClN8x>|$w=__^#ohK7$lpU_Lg_gxL0
zVuGb>?aLaDnhVbq|J@P1Nb}#efBiMSpkI_D*G%bTX?Z;br&#Nl>h;%Ji^N!e@9A&R
zSh%vPOuha%b8TID+ZT-^`;8VYiBc{&lh1`5fPeOymrJCe@f?w@_3n@+*8}uwqt3X6
zI@4B~7CJBgW;oSb7_9Mv9)mtM6sggWbb(uWsjgP1$`F?(^I`07EJnmvaH5~whB10c
zT&{J89cT~ftjd^G@T6b^rvI=xthQw*F<sBnobO(wALR^{#2v>V62!kJ*h|`bQZW%c
z(!4nS%TcKK109duOw_+_cz6=02a~z`df;Bf>!Pr&xNN(mC(&i>C`A8pw;F5u@y%sg
zD@&;-g?s<e-Ab#+LX#5(A|~w{8k+bkk$&Iuwr1YqA2uM}w({{St1|y--?_t=m*>_q
zty**}diAr~j6VGp|421cO*d_ROTkxyfW0p-O{O6xtHV8END;8_&`E9PDhINKT8rn|
zcn-MZT7B<-`jXIatq!NMpz;0um`~4CY87oV#v(OJ5|7EXc4!r&HXq6sXJXpMW?R8I
zedB3Q-~W<ce7;<Buv*rO4c=b0236FhjX@o8Mp?HsxJ+}JE)`q%ySQxp;INdJ@;Tq^
z+Yu6*ExIUoxzXr5m;E$PS7WeS;QMeTw~BOcKjwq9##H~$n4o+lL>aD5?4(@-AWy9l
z>xZ)y!54**zEJv2j(O_%e+_7!^CTo9kl~u6y=EZ9rY#*f^oNkoIXB?-rBtgXdnYC+
zvu6d_eE53`QbEu%)=HZ+sVw=~)4j{~a4L#H7?x*+W>xSuE~8sIW92P)4J25{p$bW8
zV<Uup59g!H?y%**zbuCxzUNCte_p6Hk<Autz82@#_{ZR|PO$zzxZ%!&_D^m84Oow&
zE46o{H)1&c_Tl=-^!4Q#xnQJ`>wAlEtf-@OuHDk_Df;E?l>0w9*cK+R4H!(%|7I#I
zvAzY0r`4(^`~vil$KUQH^G!i<cY8Sj+sv{#WCbq&K0T;QDp0^<vEDj_fXANH@8N17
zg#l;roj2j>`qY248q5H#n@6H=sTr>TsX+|r66LHU?8d@7UXRmEUpe(9)@SuP=1R!O
z8DOzl3>goaij^|3Df=~w<gqFZTDfJfG(1!(j9CyIgF2;S@l`)b4v(+XvVHy|inP&a
zK9a`P?d0jL2q9q@QdBxs%B!P!l=*4{2_P_u+;+O}i3M>B<^PAy!32=OEL*6@%ltQ(
zZN=ls0{EeCbRpS7KA$glM`bg4>v9ff!nlCBtw+8&Xliji2=07+AwZ*$evckfso%^#
zSEYwT`A{xUhE|IM`*OF1VttOcd|)L@bv17kU|pf<(|qPC0*Sm487hM&<AG?F0J?#J
z!4C4#U^0yY{U;-hQY7ntOCul&$t6cs*VKQZ6f*{|whw%xOJ25Du0G&*JA_r1ajk&N
zY=uHJBBp+Ui0g}bg7bIj7qCn`@gRzRE4+@!M+#$jgao~8`8+_;;C)IeoVNO<Lf(F0
z^?&jd5H<7=`<E4WAE5Y|jcSdd+)oz6vIIQH-4188f0{+Qh8{JhyIBKd0(kZvf>oIR
z`gqW35aek9!I6Nv3Oo=XWIaX6Xw*UIwE5@VwjrGVWg#vio~j5nbO5z0$}w8dSeAf-
zls^!=lR6*l$3JVXMe-?}Cvj4tuEEmlIM6P8V=352dcSR<$_NR?fZjOl{&H{Y`e;6m
zRbNOtRy09Wv@`k9E}h+C<m(3lZdvn}7(9*yCOz(Eq$P)j!zsx8rAZypCeqWp`ivT@
zA(_Gs?c1!7%1m^Z?|U#s1<0Rpf{&TV=qotjbsC%ThJ&bu#yzwPsLVRQNB5B*9H2Pw
zo#$$+MIx|ikqbXFG?1*4Kq2fntR!OJV^GP%J}jp3I^}OEcMvyAca5P=fVY4DNC<{Q
z+MLLfz-2c>XmL9VU-N(bAQMMer?k_8{a?VkAOU*=Q=Ksl<km4vkmY!G9uVN%l#?hd
z23DInM#)(GSz-PXpFCjD3rx`4pY5iLKa|eIqf<q#+)V3&<i9gU{<0k3+!;=t`+2|_
z8`w+kc)Y;IX}5@(!EGlVg-J(<46l2jk|S&&kmTC*V;*J>f%h%+a+&8w4;-iU6fBp`
zoFvfwI_(T$fpFW1iy))2MV&x=Hs(n2CWVB~1C7+<bNG9_1Ws$!dv+O{4^008+Qk6d
z;rAF0gh-*nz<wUN9G%-%`(qh+a}!-XBKNTh?TiPu+MqQY5sTg!P{)Bpiipkz({U^p
zv;x<IDN?Sm|Al}MjTFwLKsvU|$N%N&{!{5kvIOb#?hyEHr>1{JDas1T@sN$eLo!It
z@1iw%fQkKIpFLuT1hDlQ?fdp83($mpJuiQ_RsQpAR<T?lj6fN(EjBthB%xm^bDbzp
zWN@sQzRIkhZ-G=3M{h&e&4=y+Lfn_d*E;;zOnP7ep&pF$DQhkMSLILPZ#v=RI$gH<
zKTNQB7AyXkD3C#E^S&N(z&iRbow{<+CJ_As-fC;rK+;JVdONEyMQ5&3r|7y5DAGhB
zpPMdl+JALmhDw@qlg?#RYup`7I&En~!~|%fJ9xD!#cfV>${IPWtJMR%oiD<~31rpr
zM^hf2g{y<<0?{DoTvOx!f=${)feJKuF}Q5dI4p*LmKq%r={4wM@VVqWY2a=n=eMR2
zU`+s25kXMNK5YTMbe#)YtHuzH&Q1201dS{pt77{TdF{-lF0wM+%W>sW6zls4%bfv6
z7uGfU&FFv@B49g8#uF2HUGAo|rUdr|UXgM_`-=v;>bk;a0M#+UZV*iX{rmthADijf
z^^w+xNXXQ_B7`EKK|f>(8dH;s1;b7?E&@K`bT~r?pnxi~?aSRcg)x7w-XsFV5LP8c
zZN9X<`!)s%GAt8pN5E@bB880n(NCZ_HUVmpO{P|wz?USw{VqW8@(At}Jq@!rq(E&#
zID@IytkQKB;Ko|-mg788c1%|i!d@fqvLs@<IN0_E+@qt4TKoH!eLthlPuT$IZs$Pn
zkN82P!GZoXD0h+hVK3zx6@{I7SPYtht>2r})c!|zI1susv=G0QfIQ(rm#E}Ok7e@J
zd4LVAbtZq?=~>&^|9a~x1zCe$QdZ-xH)Z=rtL;7=U$86x?_B`31TKJFiPDqVxnUm{
zYoZLmBU~NMA~ia!^<V4^H^5cmEjRs2<ivasHhD`(R0Ns25A2;%v&}rSVirGEZv=|W
z>2mW9k3}s>6LBuYq;COH=vYxnJcT^18^8(&yu9bHD%@HE%1NI79Ssc)7<>L6BLA_0
z@UZTu%aLJ7IGZ({FNoixgI&)u1ihS(mm0@=o310JyFQ`j2KV>%Nio|_vgKosBP7V=
zfK~O+CjPTCp!-!Vk`MA<fS8oUX!;;{6i^d>dP47I$fpyL=3+g#v%YIVf;_FEEp!Hm
z0F@Ie;BmSRfR@!nZqORTEEi;Pk)ykkX;ilT%nqiCzGBcQ*Y@QBwy&CpXA55nUyN28
z2ockw-rkd?Mhc6OG;HSsD{|3J==v#|lZgM6V7(M!B30~Uu>qn*5Fjm5NR@8)zTReL
zk?E6+*8`RaK@gJTXSMT85O&}eT?f$7YPLLOyNOuDUvR()SHRttV5YL-zo|jdR~^pR
z!~<@-)wt%5G{>r5sf}hcU;SaubS7jPA&p#=;?2~#-AapwZcaw=kxM+nH%1+FAS35)
z5#ww3IUr>&hK4}IU+){f+~$R0JzbolUPHw1+BIFGLWGGXWDBWGR)|=3i5fJPK_nVB
zNGXGx%%H_}>T}F@bah4R|5D#yDv+=YkQ#|j|K}&&l$F%+1TvVO(03smmSYQkMsDz2
z(7#0kTkn80rT*}y>6c3ME$aHSBS0Xl!<8=oc|E)+scNC@{#vI<wP8EGU+r0|>qM@Y
z=j%&5OpGDQCrEf~A>paD6m5iv(j9Kl-gr)d{YpzN)t@9P1y-LMTZo;2uJ@t;(=f~+
z7<gn^kY(I9=`Z;d`o)X6EAffFZuFqXTd>0F7M)sguGM6Iw{y%3S1BOl>%fl}ews%I
zaOEA%{+pH<I@Bq04yQG`au$CWpv7{plD=cU&ktO(aWWpdkP+cS3~nNgW<>$`!gU(x
z;qOcm4_60?w4cZaQ%IH%Sg-y-mJ%wW3uu!mc@8)Gq7wln3AEu&&@2BRDh;$xjM~6H
z;RMas7%6mY0kkl@*8KV#`{@gf2WQn)bBNr3<vfAlDNSlURk$v&4Uml3D`S~j$y+}*
z>sN~Gx+%8*G8ho;PSaAGf4iV>pmb+b7_^l6YYW^KfKCaGOg!|Bl@1TYYk8repBBE%
zlqR~`OVe1!QQlwd^cKj(uOA*B%J$Bgx{<<m_})Y7rlr!Uy#-d=DK~M+y>zxhE0u84
z=jOQ1uCImrKYNQo9Y|xN;I?0;^+;05;QpZV%VM@QBY5)ZFb{BiIH;<L#?&7Z&yWF)
zkc(0+RqFgopi;p8c;I@mEw??8*lTM@RQ9h{eWSZV-(iR%`N`A^f52m>q}M1fwD$1z
zIp247kL?P0eWtN~K&MfT0r2gwo6$+TOPdy8BQh=rOOKKVFOVE^4-)MThI6_*Gl4Lz
zL908Q$9EC|_YvlYVgbmcVf`Tavp=3k93)M;K<@bb;D*VpudT~NDCqgo{<`>5`hV8{
z9f459U;p$)r`efGJObHnxI9GBAl;p2Kr!XC6a_7}lW6LMQMVqUFm`g22Nozol#U!C
zA@3Ontb_mArx@zGz#M>aleH!iGnG1O{1zn?2LJnX7Kr{CO-!2u2n`h=7N|TAP<xo5
z*Tq0;_VeQ{h23KP%hdO3bzsRIZDCj2yePmN03(JaCnxs?f6AA9mnrP0d#dWFje6A;
zLIe#gTR%W5+o#F3CcST4&*Az5oA&=!XtAhhJ}hq(O8v_mAo>)j^y;(_a8lpi8_mFh
zL&Ow~#%4M@+mN_~bNvr3&Y*)(gU2=l47Vo-|MTT8R9=_e_vxJ0!A(wEJ(F0rP;XoE
zAU*>jarSU*&q8{@)nWzreoq289;kQP0Q3+@(A%NPX{=-bSfJ~NZy!0$lxrYCy15Qm
zUxP^IQnXqO*!`~wQI=B#yh+JNfq^-~?=rKp>g<-dpYF~jHWlF$9)7WtPR!f!B0|j9
zbDZ;oyj7R<m%rbc^zuExt6zzQyw8a1iL)Sfq6RU!V#mWFh={C_pWR<t4zln%tO~}E
zh#1<-UD+A5WP;JoqT)mYDj$jU45nUC*JUuleA_&;u7+>0av{BBbaeEy()ZIu=`MJL
z<K<?I5BQv2Ku=TQad-cQC&vsB^_oI9@B0g?J7f?Ms%s^Yh`(=+dcoxvAnULpFpM4e
z&SEm2Sg0pgETq<A6u+5wvD!c&6oo`MEmjrfUoV9MTmPddAh-Nz=D%7vV$$^rxI62X
zcmQN7fzQPPqQL<D%nbRW?86=p0$`2Ctk0_wdm+&dG{)JRZlRFfQOBEO9WSM6Xk~`W
z6^lCkX6J==A0A=92Tb)6jaxkm(z1dt^VLi#6zSFDH#mQQfF*EPDl4XQzC$4vOt?j2
zzEOj61Fs4#@s!%kAn?)x@m*%4&K4N37|)Icn({#+b2`<+uBZFUol_c+sA#+@Wb8~p
zR|wYIcIAWQuMEO+haZ7Pg(idq*(^qIb#5R&HC-ILT|+G3URnR1YBuhBmqg&$>4*+&
zqD4epCeSbk!$5=AI#X|nrj7x4V&{E7NNu{V>2T2k2>2}Z^?^K~lCf{rxSCk|vUq^X
zLD}CW$`_h)J~uc8+HB%}g`y74eV;(d8~|TK0q(iQ@D;0eJe(=pX*>h?s@7}}-E1W7
z(?b9oa%h~}eb#pf)$L_Mytk_kqKlyH03q*K(LmC!ss(^(k~H<-tL3RwKv;AO0L~`g
zdv-u_FiHI;W3^=hGn?w)vtjWg%|@xHsb6`0l*)BGnxoZks=a!LrVd&+daef4`;M(l
ziUM^8)$`@4*>@h7)%cAkk<B>?EVXjx*RN<4-sPEX4?y7FVH4${opC`zJwE|=5Zvs%
zljQ2uHjhj!sEw~n`)a`%S|Fa&s#pQIqhA)>LHFSp#5DZHfyUq&Hs@dO?+G8&A4@Q7
z_t&*nenW-?>kASN^S3yM0;%ZTtS_$FZ^Q%Re#DR*p^-~Msv5e~Z6!v!D*$cXZ*Qhd
zZSh?-Tv~Hei_?~L`|J|5F#n=3T5ba`)8)ZJGq8>~e+P}HDgckm17wDH6@!<vV`-(u
zjUg>Y6-9!=m;|z^@|YIUwJ-I9-JCK^KLVL(qEHS)z4W74N2fCnW*OQ6L-_nZ%;|Fa
z7wRk@A!kUaSuMv1Z5L|0VQ4svB&~jo+lA;Jqd+;M5HvbDlNB4bd*94|TA_#yL_u4(
z&jn0v3*7no7N@)(Fgl=_mdzIAyE$3f%a--#1hmsfY8@0$Qu?VNeEFKkc7fIB#G>{P
z$BDaSrqlgIw-&#II7XVZ77eJzAR(BbCE=>`qq(Z+3auJ>^Nz&Bm<^zZ3_%`hfEarm
z<Q!IDx9#YY4J<S)ly0~C@xmm=I6Xm$=`v7DU9+=Xq7V55=fcoerU^nm?zhm}j)lF{
zOpwE_HwW$e13!?yVrHuy@OWQYmE9r9o4(ucNUAoQ%$LIDu_yd#F&YI7a}($lpjJMt
zWJ`BpK&u;{HQHk8(q4E7dz|W<)+0Z1K<|Eg@HkyQ-_vxv#)krl16{troUa}}Q|uNX
z`df5A#wArR6*IU;)zahs?ahcl=Y_nYUL!!ZUb;lr0XwEOXsItjahOe|{gfZ5nIqtS
zJmJcWh$+DTat<PCa@+`3Phze7m!hQO`#`r|3sCvsJCz;(?~FPV_14q5+~3|Dfe)rj
zc22_p9NnBOkb&^vvVjyLvpsk&9}Phf)+!PpAk{yS81V2Y#F9XZ#%eV|I!LAHZ(>-Z
z3ylp$e{C(s544!pZRdCATY)Em_DU4%Zy-4XpJv7&vHdH4ri3$`gT`0nXtE%}mpjQ_
z8NxH`ya=JF*i81tHgOylLaP7uoa=UB$Taap0^v1AokQ_L^k?@yXn@pWvKZ>qKmMRG
z=?Q&VCNOAsK=pjQVfqY;W5RkE&lZxn+Mo1Oi)AiAL%@XfM9KddKc@)%su+NmMZXm7
zm5_0m^>eQVR|$$J^!*Xg1yu9#!VvvbttsWd4d7Pw;s?Ev3RKt8PuTK-tI}L*uz{a?
zAmVeDq-Te&=V)O=L2&F=r@NN>iv##4%E5yqL=c4_#eBN}7m?(82P~G0T-bmO8Gx@-
zc+6N&(i(<$XB&0?&)xuV14^eC|8IBwEkr~=B)DQPk&41~1EHpJ;J4Pal9Q4`V)CKg
zKCno8y?5LUz(U4jFZ3borLSR&uS=zmXH%Zt$>4Rlt|NAF!$0Er@JQKN)V8}hdOfb6
zsWM#tRL-1zDRQpmB&a%*dG33;JF5C)yv@}!18|hGt`Z3npL}~S09kLqYb;nAmSdS>
zQJ8ew+B3Lbg}CE<%3BMg2GJwqzU7q%ZskM6??qm5hjF(?cev(kMuuyPp5~?)8<d6x
z;vBjHp^8VfZ9%FA9oB18+9Q1gOByC_W=r(8-s)~;o^E3W=@MK4h_LkBF53e^>Sd~V
zAM?ajohqlUCWTRX{hxTFuo%K`!E3D^Uo-CkLj?^aQtVtGW#_`6L)kAWXK*u}u62U_
zPZ+G45X6VFfHu6G)^7#@mKVAWoHP_K9>x|h?TVO2qmumw9BqVO*sue*PtmVl;0FU)
zG5W5Yh&giow@BBDs^b(}i90dl%%*(BWqTR?>|?aQqt<>|#;fxt>g7li$}`T6yRH@9
z7svF-T+0zL&F)NlYIO4BgNiq260PYg7dH}L`8?=jWum-sO!yCqJw<j~IXyCWMt)4&
zyB9l>MT86Pjb$Z|quppR9hV+y#3sscadR1FSJd+BMwxnrMYGj-Y{t5=+tT;*`zH0b
z%?)|etngw_|E_9J=BR1->2J4GaWU6#x5%>9>;XCV&?buA;?}~s{N!0x>PoCZ2=05z
zp_D_vuNtz96gs&AT32-COkT3JPQwmgI;?yB3IB&HI?pkRwb9bEy=srR*sVO%JJ0h?
zuZLLL_q_gCnlh$@z0xsEVM@bi`jK^pSS4}&59*ocR!l3p_(!#K%+pmVjo-0x&6Pl(
zk8_G@>tYZyhAXCGwp8!aq(6btW3<MpU%mtUPkt@ni+qmvjZEV052M$2gb;dmbo40I
z94z5;pF!BDH^<A|k`(K2+CW8KLuFHx4MpC$$)zHl2?)n5`7->K|5Ziu%xff*-d52R
zub&|7{s)M5_ri;3Oy93^N#Dyfqu3;8yYbt+ep~a)kLXKlofM<l0D3@$zdn~9d0P#_
zcueA%r?$rnqVcr=i6>f<^BwpzAxYfKfD63FoObvu0XOh0{9qDQ@(%PC3LP(X3`sl;
zJW8%;w!UmCJzTa>KoI6&(oiaiaOKP}$)ly9779;49*OUvn5E#)N^j$GiY;}%@D$u>
z;G1?Y*jG#L{nSz&7VBn=*6JO~m|f1L1=S(T^3hx)y~w5iQbcOBOb{)=ezhZ{$?5af
z#dQX+>oM%b?qIZi&*Ya#bSiahzxQUaP1k9e`O<r1$LLI>*3BwK4cOxd`fZ+QfZ<<A
z*7Tz2pSqL!>vBQ|1-&8Coo<Y#3lL_)YP9`Jj-v%rU?gmv%&Cb%6O;NvhyEuvmJm;k
zLi(HRrU5%_PsC%<#NH#ERC`3}EwXU(%;kWmnc+2~n#^^KX*yyB#a7Jc;X4H${Y38O
z<T~%$8zIRT!MYjlWTqmWs76icjf-D8w;@F>zW6nVS~8|a>Y4oPN?{rTg5@4cDg6a(
zEWJjZexOE@we#I)n#>kYA{1f~spq#JCaUyqMVHRK0bY#30naFWXG}z%Z9bDWC1w_+
z-6?Ge=X>jBz)`t5%jH1W+Q1t|aP-j}_eQzR8ZVomPdkDw+r-BDw=7RwU%g$;;jMzs
z+)3)ioFiM+yyt0FZJj+9x7L_G;iX#4JRXqz+XCbs@oIR=L)`>C_PLPTlD9KTQ>}wB
zP)`wlKgjOREXOm$U;Q3qGL=7}u>|jfU;RnN(>`trWUiGSciyqQJkxZ_toUoyi1>p*
z4k8Jxx}Nv?Cir1q8b5MAX>MtUk5g%^;%d6FK`yi~%{dEB$6l8`^%pd_+E`ciss^Y(
z&kgA*n0nb9%uV^~&ABr6b2Em;nyWA98>UzsXAFTj5o@se@oM<`1D?-Cp6-kg1n)7S
zOHSK+;}c9K%9VWTv82`d`+Rq2H(wzg-mB?xEt6xI=X#)j8lN3K{%#Dz3rI{bZ+&*N
z2|>{B7AqjjH#`iKW71TcWViY=<#E_wG`A-{PP6KzOCt|=f2oM+iza=^0S`3vo$>Np
zCUebZ+nxbAjzs<AvotOr<y=2lSkp8=K~5SPnzLJbdUe^$zh>o_oTGaKd#;R&Wok4k
zV<WThbrhxrAYxC11QE}Tp0I4}1NHcC7b;mx(}{ir0v7N4WAI|Vu@v&%R^(;VX*-16
zKHXA%C3ggGV2tS25l6rLjXM2I)@nI!D4ut%MzA$*y2)X(KarQl??yjxzS`nuL#dp(
zkfS#{YXkd)0;StJp2>&t%VIQX{<13wX2s*jol!;S7ek<rZgI7Wev+do@T%4x$lRY-
zE-`9$m4%$x>HA(0yTl2BIx-yC{4dmJDzwHs_;E*k3a}o!eS=|aJk_03m^7ftjF$g2
zIY*)i{M6irpVMD3ZDkSmm53j`J?AUOn6qCXBz-#G8_&V17b%+fT~=%MWr<wWXyVuj
z*)_*D`i+P!Xgn=~&IM;PCt!};sKfWSo#_0o9jyIoTVxC#hqvltvR=JHDl;y-+1Hn6
zpCF5!p%h3z>hWGK!?<OsJDT40r*xMELIO>iu)zq^AE(5)4kFnwi})V6Kl}|g8<Saj
zX*~G?ndtq$K=y~ia89tL3l5V3f}d$*gQ*l!<@o7e3DK?9E7Z%#aRz{4l32CEJh{ZQ
z*n^}3$J#TvY^c|KG;0irNHsxzKDU3O!d%$e<=M^TDzWb4e6HbE5_1CXkWc1X^5s|h
zafRG{Cx~QGkD4|2;p^TZA+GFb)&aapu5<jAcY@95<NnW6Q&O@Ck6{DbjvL}z65;PJ
zm`@c1+gFq<n@@D6keAl{cPpo>_@wnyyPVA+MFx`|cdRTaDXaUj_Q9<f-Rr<FCHQok
z`RcH246qV{A4;Ke7JeZ@o+Df6(iDKu&@e%Ne_0j)zmWUc=v|rnfX_&MPZ&~#RV#mz
zB{u+Hp4W$1t659T2mHt7AuSx;h*0?0oQG`C0?66lo))7zSLZ^u!K}{P0~d?beg>i^
z2({O|u73DXHVBTESk3zfo>@rRQNlyIcx%OO;2GjAYL(swC^%F;eXBLif?;b|XrdsL
zV4LQ}o?+P^TqJW%wg^`*yyEcn;%jIYd~ZMl@P?D(_lfL(Fk)y^Ip65bDZy#LwzMn<
zGi;iGyD3=R5G|wG44NFFI-fOR8Fm%B9pD;f>zP<*-GJV2-Z`@T8q#MUm<@Z|XMM4O
z-4D$WaHdt?O-3{ZSo6g>pw4aWq)gbb6LOxQIjLVr6VZVkD{4%dAn4@`bjNz!O3}VA
zIqc?9SA(WCLz46%^ct0cwIafsHBQgaFx)G}A>^oR5>DS(w6K|vcouxoA_W?TzR|&0
zT3U`|@QjP0&Goud!rr||yfOh$e}=Qgp}Xs#&5y!7%1+vL#!?DldbGQTXAxeM<#^(i
z7AuA4H))D~^9XMeh{_^=SLx<kwXwVuj|AMjFeoqkF_8H^-5?G8EuylbVC_(KrtEQv
zA>o+Xg@KUaP-J|1GlujR^N!KUSu9=gT-l&DZN7k(wi^IAtm!i1ZsB3Gg?!%fj~9}Q
zhy}wgwEHx=#vPGFhrLYrwU)C3wy<swue%rV4x_*Np;%yx7lDx96`@chH$;j4;_>>3
zCQEnPDX*RWl1~8U;T9C3HC}6_NCNu7D5Tmdy8e7mboX#d^>EYh+#s{ql0>Bt+@V>2
z_a2HrYcK9ClNb^dKF6PE9F-1*#g@Ga&8nWCa4xkiPW)9Pw-il##BIILS~<_!(aiQp
zusz|(kw-9ypGT+Qw13<FU@D(+(-sre+t-7Vp}Bt_v;~SUQ2l|vutW_*9)`5@t7mnj
z{rHdR`(<aO`Nus_T9!xc)3#~m=?_pw5s?5vRbTOGYy4=c;rcAdVIYwb@$EOc@VigF
zKj3z!8%c0mkC@#J{USjQo%znYyjE(oXql$cX9=$-d*D02UA%1t><?S?B}vvSa5Da{
zqm7QoV|t-TM9@-deWF#z=Lu|LW+>*fgC#X0A3gh}D*rYq>4uhjb`rOG9EEGCkbItN
zGV!k(D&C2R&507fcP^0iAD=%}nisfp8Kif$o+hmO+<&Il%AC=-)xAYyMe+0&Rzs)5
zLa2D$C^TBDn!>#jw^aW(X86@1J?8I?^UWQGD_l<W)>{&EJrjvM#|}Z<$r||0zSxr;
z$EBP7_!JwS_4{j%Q?G!UzX%SiEljxO104p71^Q*5FJ>xSHKd7_t_i<7e)??uvzGD$
z?S|$ZWNW?XCXqsh(oTLL!QSS!4gMokKgHI!WVLKHb^KoXl~$A8U+US`#h*uP(k*P4
zICIXI(^SF$g9y1knpdzXpDI#NsPLUWFKa7jCZ-$l*crk)cqKJxai#qzYU_tx4}C)#
z#8)$2q5^{KoyR8jjG<E_s<|BW|17?b+3WZH-qZg?UwKjPLQHofSaZCVOrzp>cV@EO
z<OI{|ev(xGQQc7OjzD8;vl+314S!3qw_mW>x}3OQt#Ze0c;mdr+TXO?I{l!z&d$b~
zo&+nR(0?eA0v>2n-m0FU;zZHc#4+~XQ_p??!;R#K^D^u?{6PE>h(2b%D%k8aWKO*K
zCx-ar=we~kbu04vM)IDOM&r+%BLcol7R-?!cwx@w=Pm|8I^DNmk%S6)mn)kkgxR5|
z-PII7(3tUEOkb1d3HMa7Y}8bkbAi-J)K}u!OP7K#N5Y;jLr=^lqn98u=Rza!9OOSF
z3`(#1#H*z_f;*-0+v9MP%N3vhZ@V4VLW{8=dDJbLnu*?D1}@MG3_RpI$F!MImdiYL
z)n$!gJAIjkpGG7M&Zgj5yV-@=_@x!%aEsf~*?Kojz1`9r=b*`Wqx}kEM|`HB7vyGB
zf^Z__<oTE986P0cvK!NfB=;}kotj(yv6tINCg9cfQjH31eFKA{^WeWjl9UIbd;z|l
z+rI{|dv05129rN4YtWdfpwD7!EEdKwrDjKiUs$Y?coLNkFKgr#A9buY*}lnQAi)e5
zpVd&G{9FyFY5x;Uf#P_&;Zqn%fEIu$8}k_&-lb=)?_0Y^-Oi><8)i!`fO~A~x0t6b
zr2QZMI%B}wcSx5ekQLgOI2>$$y`09Oc9Sr5Z~8nxi{nY@DxvZ74^~SarDNRn(Y?8C
zYBXc4R`cc})BKxr2@~6;#GdrR;Y5q29MEE?2Z}2l;r|0#K&8LW-<0{L-LID`K7LPc
z{)T0mSDyou_S_Q+{d+=*pB#y!hr`5w@1pYDbJ)`%hgqwDz8h|0yHMu0%%2D;Vn>kW
zyZmz<|1{6+0nQ#SH{km6GjfTCf|BOTFXOIQC2x=`fh;?D?>($K{S>FV=QewNL<k!S
z-HXIS-=5vDYwrP7jh~KBpFYLB`SWo3&`v+q;~#}5&;C<#F`858yWY8vF;i!uY5kgh
zR_Jfty%#4=p9``I{q6tmL&G+mC5_kINuzM>=4}idH313Z$494DO;NSsAGmY>0lKwo
ziRjUyV#18MC|A2N_HSA(75<jyZuR=jn6rpZ<9>YvzZ0C{JH@rvuWNfz|1ZT)%H+<n
z8n4pLwKb$yXQatn488mGL446?{}g`Uj2Sr$TeogQ)}%3!CRS*CeEA;lA0L!2Wo-LD
z@gf&eC^>ngu2US`&zY{$7*^@t4Ktsm@=P-oV_wrVW80?rjA`b=Da4>^W{r)+QJ><=
z;lJ=rWa4-!sxDPdo%2Xxz7J17;qs%SSo`8T&fIy4QYA~ESHFQa`+$=T`QVQ|%xja(
z5d}68@P6N)9<B>>!%&!-OH;Tr%`~4e%{yUZc@<XW)H#n7dz@E(R8e6wdE6+RIeQ-Y
zQpQHIXd&_A@mqZOcphIvsl^EMzeP1hj!%u-TG%IPrg?d8D#rG`N;A!C?60*nm9JRB
zABd1N0iuR~f*;41<K54D7u!|k)sbaFh=1RH{3o#d2-lC7vHQ(`XnNxrGH1$&j@^2p
zYSpR&GdFxq`);fAVM7L^L&Iu8PMOc~Z{L3mC(mECi9&zs;?)Q(9BKzug+3k4?mTn?
zYqsvi)mwM*`O8=LOXU8er>M}fCvs&<hq1le;O+a5=rMW*&RxBUzsC1>&!+<Aixt2J
zgC7z;md4ImiX;i*A#50rpW|%@PQb0Z50NihMpYc=MZNbPJrRovjosrI2y0>YRO}eh
z(4|RD3>`cOOP8<kP{y|>g9Z&kf3fp~6eL=Fj=+i)D-b=p_ZIlJY}o=D%4vnZ8_uUs
zpO*J-?#x=aa3MBq*no-^D@q!i-+uee(_{=AHVh+1jKKKu<B>FJQVba~1O*Eg#HCA@
zq~e~2FI-5UHf@?*ROd6<nnZ~b1-$+T4jhP$8#h|Z>g}|dGo~X?l1O%?%xCARAU<>d
zo!u++gGxWWeL?slBc0GTTY3?dg<^eDwKUjq<pmloxqy{TvLJtQ&-a>1LiaOe?;Q*j
zO8msp!(+m~cTj208SL$tOB!G{`RfAizy6Fig`E&Bd}z$ve-CvRp2PY-vLQm4Q0TM%
zx@@C`STsl}d}U4)<vag#6>D2$HQW2liK3uT@+g=l4BLA5?Q1rVj`TYM9^WARUZ1YP
zK@qrj@1A_|nBB#)6;8)B*Mx<3B;0=S35RaKk&5I?k3V3}!3Q{S>osCU3WL1P(a^i3
zv%7+sc?WN~jVn(+AiXH-6yuiyZogPP+$9vwbL*tT!lRF{_r@zMsF&V-2AVCqjAE%{
zpkB`S_#(KxW{Y+<i4{p&o9A$ff)V9^tw_2T`Waw-wEm~1c@Ckd?qAv&iNuYMVmgjx
zjWVL;ipywSFfsl*{umkJM?!yL`M&M)3vr{oE5`d7Nn(UUyF!WON~XCws8sikJbCw3
z_N!!?n8;wk+TBt;GY;HG!`unbI8OqZC2IIks50-ow7@T)(W6E4jb|URQYhf()J}_X
z8Qjv6iPhKb;z2=PG;XD$ED_^9Egt$Q<NR8geJ3pTj|*C)iB(ghde*oy->XmG&};oo
zJb3d3sm1ECIjJTdM#CqZl{~r0k|i^j#Yb_A7cWNjisgh=cTBV8rp{f6wHvo$=h|h`
z>T$-Oe_`q0Yw_pQaY&OgInG}aAGBS2W8w7ia88mK14d53>J6Lm`psM9&Yl(HhxS9-
zRBl!PP1<!s)=cSfL##mWJ#-Z3_HV<6tvfM)=}H_rb`r^xI$=ofE+|#BFv`|yEUdua
zVb+2r`1<uL26XS}-oKMVvD2aVKzUxXNgGTaH5di+<ix3S7tnj?Xj~Fj>ZwKFI<{<r
zioch{<{keE%%-Akvqq>}y^<`)zPu9KL-Xa#j&;I9pNi;Wm1<+R%L?hNwV3d6p)B?<
zUA>MLoqA%?j0p%YEJABFYm2UJTVl@QWyq2~9Yzo8ZLw|lPJ0pH82=GCfBGz}J?hr*
zBDcrn9eegm?6$A|%RPbC_*BMMC{rAVj-SM+fxXbKc_Wz~QP8Y#47>InMETOi-OCsQ
zx9{8&^V&zw#r`A5+>@@0xubArL=-bt*<Pi~!X|%W#)ytul`CMuf(7W^ySGWIZB9um
z)uBx*Y~Q*G?eoXQ(2kjeL#<G@socS9`S2wXR-e6y3Z;soSjjS&I&(H+#PHj$8Sg@>
z=heGN`8;1>&MSLMg+jAN_3`D!1GLE<9g9n35lT;`U{2`s0qGK6BJll3BJ8~O3d6g%
z!RX;b@aN*C$e8h$vee58e2{_k4C`E6aAo@vi>3bjdlPmXI3^U`S@GxJ8>LdVlTf~X
z6a3_p7OWD=;SZ=)zBq<-YVKag?By<4v}OyQy?Tv2S<+)<w?B|1q1)m!mAz9It;U`s
zr^F&c3~3cVq;qreTn{Vi--P`qgg=a|88COndc6Jc5fzH%$JkzN<-3o210pbC{tB$$
zc>wQ3o{|kaVaMW0GTBnwJ%s}M*tyGCGGPD`#fyWxLIK{m$8eM{njd?Pp2nMZ;vIj5
z3%&~fTGf9qik5vwW7e?lNSD%C);n|QS{yuiUY1|#vK?FY9zm&sxiNnJa?Bpq19`J$
zfTgx&1>X4c=-CU2TiM_8A#+-{cQ$VnlUYt(x`z3yHp+Q1E%UFy`iEFJa+y66ks`Xi
z1Fze$4;QcBHq9!}o7O1jA%5%}HyVP~Glq$sWMSOX*xrfrSCA&TlY1NkH-xcA=Cmo%
ztY&!+@zze7)hvfJCF@BCxiMp!e;2k^-k3((t9|?SVfpgq#%NFBzyJOlRHl+zrfAWk
z(uWNxdMGa9cR|JK_3PK+?CgwQy?UW;-MTW3J8@_&oj-qm%$zwBZ^drON|h>M(xgf9
zI_0}0^Rs?Zz?Cak4n2GJMB>DWrNj2Yg9l^Vwr$b@H5IpXqMa>UHk2(}7Tj@1#jX;j
zJ$(33j)m=X6T5l2JBn0&l`B`4id%CIP+7;FJ9iK*S~RR&xl-;1qSI8?r^MPAIhNeH
zbE8e0Hq!q`p+beEwK=)jW5tSvKEn8;X3d%yGiD6du3amw!iy9sB5|Q*IDHlzJa`cF
zDME~C?Qc%3jfvkzR7lgZoxX+s{PR!gSe#gKUiR$S15yKVobB4RlS*$=omHz=4a0{I
zm$YY`+v?S;qkH%6vaF)#GuIbm9OrP&nl*C1nl^1}O!6{3b?Q`!(fs-IW!_SyN@4Zt
z)$&>~r_d*cty{O2ShE1<jCgbG>^raHW5<rU@AmSt10qA?#*O71apy#$M1HqkpEGkB
zrj&_evFRp4G0SiLZpEEt9x86R7Co<S26;Z`;C&pw^9FOpIjLs)Y`Bh-;`^IMa{PAh
zSuz<Sityj39%F~lmlR7K6VvzI#e}M95L4(}=8E$CuBW25Vx~B<ommI&VYASceE9ku
zRpy>CtEj#C{G-{jwCHZN;*#vMtI*%<x&9Iv6GlNVq1Rb-^r6%@#EcjQy-OuSQPCHQ
zwSF^wtJp0-rSYxjpM)+YJlYjWf~r~K$#IzL=WC&FXt(Mz4hp?e?8srzxoA?9%@7Cd
zPg@b+zK9d%9=ea~Pd_4^(ASKwk_L%?>XxX#pmjR+OY~(>CVgylE9NZU&!|TtHk&TE
z?Ovc|`-qCEq-DFcG>*YLa{As|tULD<gUhA#PImV^e;$5-7!gCuIc4N;sbXT|1wr*E
z@HVH=pT7UD&;=z#<Da`Pq6nou=Xs~FxaYM-{Y`&i_}MI<+w~nm;Lu|ECF~bEw&X&A
zzw*=*p?6D<av9yO+hoGvbGy)Y8FfSMo%<PRo+}pqm^EE1^i6@W;klbu&W080svHx6
zO`A5Mbh_9G6Uy=>by+Cy2MOi;-)EkpfLJyB{^Nh*CUF_3ghIGgfkY@PZk$x4FBP;p
zl#hajrP}=SQrTZZD6pTu|0Zt258VO2E3Cp5pFUR;FFg9-p2om=vGwS|>n}2HnNZHt
zN}LL2Dzwi%cn?~Zo0C~W@!fahO(cvK4lRUrIu{+v&pr#@=qaY-`26iZq11klRx7XI
zo_L^UP83-_44R4s1=7=usE`7w;zUHd)z_qw-&`HCVhew0HNXWAE*j(-cMMba-pBX<
z{wH4~nP%?4X;oYZ`1tid<V_mQJ&6i>%GIYHabp}8At(u=h7$`ax6onrRh$*uraBZ(
z;;zi+BFi9=Ho01AL>4a$O85}m)h@eunQ+@Z@j@($a9-m@2`g~;rTAC*)sZg(WyFpP
zT3y&e){75^s^v;qtjZ}-LhL_s9Oo}zm3gRiK6dIX_8d4Y&uOiF<@znDQ0~-s2qw>2
zAQaLyF|vP8+!a>g#VXg4%6(=&f9VRk4H%BUR;@)*;dqxy|9Y)DB92&07}~q5+zws2
zK})H~uTqXQAYqU>LptQkm0h-_{=|<PN0iSmPs^4piiGjvN~`q3<!j)fu*~Yzx*1{$
z>z?XOT4U4p-6&nGu>7Fx)ORqRJr^IXD^_EPSd^$*z6|o_%!U-sPKYHGkku-dm35vu
ze-W;C@2Td5CFmzP=8qpg3x)e<8Lrc!J-ACCN7hVM+jg(SLV$|%K10VKY1RU0(7H43
zJ$PvLT2<GtxoSwEGe#^bDEaZ@2X49E6(4h<@#^(k8J$8bmMBpaMVGIe|83K0cYQH0
z^=ni?vBGXRtFl~-2`7{vBPY&4`8rMUN0;6>bK#OomWwsk(%mzf1gcjk173^%A*AZa
znJqKI-8+m!Z8M-r?s!_6Z$qF-9oI-;dZ0~uM1FV-ImK^+AH4o8dS3k(;lP?#j*5z#
zA_elHTB3K@*f@oF?}=ql=KJrs9l3Tcn3PE;K6C}<b(}xRU0=4cvNXkY=gytTpCb#R
zMe$6#qVm$x+C?bN$In}i(uMQjy?7Qk={W+OhfR^n`0N={W8B>3pwgH!St$9LmMDH~
zv~O5lT5A`n(@rY>Ne5G>(;yr>dkG!IqClmR1u<*+I?NXf1r)Zma^o(H5Y#YrE0#n;
zK}k1r={l_1v|FYd|K!Y+7Ri$&M6Br1P_=XsnPjc)M%BuKR4`qKPm_5&4w;0@H(XJ>
zLJ3qVQAk>zXG)U-dBkD@$9P;Qs(Dnx1%Pu`ZpzaKPo81jj{RsoV4SpCj};@T)p4sj
z%pLh&vFFHX)ayJLsf*Xf7-0#|#S-<tVeV+I9**I_s+m|jYow>-vPJXZ%*F*$na}zT
zoH&Pl$IhZyUXSl}o(hVH>#p}CRZqWmjpcl(JX{RuIc6s2jqHUekvtl|+;F`w7f4$5
z8;dHfdSmdEKgIcbRn+W}brjB%O)9$0X7EAe=FOX>5|}<!ED7&#R7lezdcc4IxOMB6
zRFHEqfC}>}RjNn|EYhY>$zHEsJxO80`tIJni=|7Kf{OVD4H`&=DV4Zt0fEY8o>QsH
za?IPTSu@!cT1~Ug;>C+gr9Z8?J9OwE(-J03D3#8i#XBzL<HwJfHER}VIK+9QQk*>I
z>>K~nZ%m68Eo5HPKcwViAE<1nQl5@dxfr07%wCO=G>D`tqL31dV<J5l70+B?;J8S`
zLZv^c!l>l0QKJU9z)#EcIC0{*E7fV8Zq8A?S7V|wo-{Pgn>UA(lM`acj*WBY&Pn<j
zD*KBGnzBz~CkYot_`XA>f6JCFrQ)1N3e%4sJu0bc=&U?v&YUvd(jVsGVhk5wxPZgC
zHuoD;apvl|cI_Gp7cT4}DOIXexNzZuhd7>)qV2sfW-{hyUW&@k>xDZYjEgZcgOxvO
zA;}ic4vy#Z`*+BaIEuv<sZ3pW_Nh#N@!=~-F+(LbzpE)OLIF$z9@3yxnsZtjO!gJZ
z*!p63JQcb<)?JhN9t!%L%@?1c>)NYG6fL||=67FvO&a8Ju^_*o%$>IHF8(?51lbZt
z74#|Lg_1Z7D*RO1%1R7>ER|&z`ZgT7{Wh)(!@%;Q{DGUVP;%NysTZk{Egn9LzEZJH
z@v7hZMcJle5hJP4rHmBHbE!*Nbw%h>o?GlE70#|NKcSw$;PuDvXua~X)F-hm8X&U1
zG;t$IWj%E({WiLNFY?-}Lre7Y&jkBJZ26zo>D|`3{a%-yc!Vi??xMOF1F49p6Y2gF
zUu$V{pcej6&$Z<EBa{;Ol$&*0EP%Z7@$Z5Q9eJHSTH+^<6;T?(nl}4HTnADYl<Zgj
z<U%1&p-$(GSPY>gjrr>y&xIcf=aWXZlmtRQHfZxrwEz1u&I(LBh(!`*=qWRp{K+I1
zb@2U*G-feL(x)~;AheC@(bYIO0w<3iLe_Y0KYD9W@WFGmYc52O5L$9~@utzUL{elC
zTXHBz#Y*#*OE1LAU_z`GihnAi|8AO5JZ#)<KBZ@z#;u<>6l?96rtKD%S+T^!i5BYN
zgmroDq*2AH^h?uha)ptKe=4Kr)J`YUxtLIB@=3`(Pab%RaU8svQb_+Zt7ckZNfH9J
z7Mw>?vEn<cRvNjryIPhw@<urGzyq=H;il*|)*&k}`SbeIcdQZ*rX^y5<EChXmj6Qp
zFI-ho7kvqoopsivgxe{t<@cIc{HvQVcR?s2A6g`#h*vKLDxSwX;$cXsz3_tEIq^X}
zbk>}GCU;g$t)2$tyypT&13`PlT@^;*qK>k{Z}8Ve5PQn7^2xp2=AJEnSe!U=5aLbT
zSf#JRr^U;1o*Y>%mMvN+A3}u+g@Z>=AbskTIC}E5;4qCOmHdA+t|u(*kD$0v`jV%8
z{*smG(z-eNbavyiEK(p3(&sB76!`z3eT&Aj4O(^IJh=}s#E0;zbsNPC?`JfsQ$t$V
z*REO-{f3XlqbE<%qG28M88!y_bLT+G-wJu?U!nx@L52V9MN83GShFV=3i+n(yUTta
z-nLfuk1rj$OH{;6Vc}11_qmfsBX`M)Xx6bOj+{6JE(A2JRZZ5LDScYOCHz%b@Hh0(
zj<pkV)GwbqRa!)O)wXFpRJe23z-(bbPr4i`*c;TWf~NIq$=wVE%T^OOpD`^a&q;MN
zutz5kX%&P|4kg@a(7IbcxiBz&%uun5BAMhzUn#IIQ~Wo%NU~`8YMK3mur4$w{YQ+K
zi!EG~nX%w64_Qd(L#)<s-GNjol40rUbz*^J4YsdYBDnc;%f$_A=^iRGoaD-u1?SG6
zmoI6yq<P>660NjIi8Zj~uSE?*(>iD)1h_~!phQBXIQ0w}(x=BA@i9t=yZ#}1Ui|}A
z^K}N~ys`&dHg7`Na^<nAae5RGAJAF?2Z36Gk}GYT@WSG>Dpsxj8|5ohvS~+7o;ZQr
znNoYJ&Q0^jO78Xi*)yU>yZ-p2ZY2!u+(KH&Z`yMh=Y_JnK=w=+HG2uFlq!s+6Ww<3
zmHjO*GM8-x7oh;BVwp<#fgPG6N5(X=U0S`<(!Oo|s_y+@TNzU)N9U%s;8d`xq=up5
z-6(}~XGO*|$?@guH?*!-MeHK{X0`1GRmz}us|FY{bpe|99))d!u4nFuo>ECoOY;5`
z=Std`5(RSN<OMgyh_OA>FnuE|8rv5&$`<oh+ot7Ty?KYf)@>73^&5oJoEGgVFs^qy
z)E2uP-hKFlNEtnj*I53f%Y65p1IBnuVJ>>`KihOFR8<;RL<k=aZ5#Z;+)~-WQx=Ft
zi@0c7{g>7LNAb?spzBa+flt2!rgG!X18MyqNl*pS7-RJ8rIHHh@akFaI|NM2@H}_M
z)HrqOq*Mr-OZRRZ>EH6^&1*ILs8OS&-yT|FE9ECGzv<X|#*7*AorRX|d>5);zrIu$
zQ-Mz9G==e85X+P)leGSh8#gY<+h3wY3Axje4ta+T9V+wE;+huYwEF%o7HhZ=z;p`R
zro#I9^XF1YPK7%c8!A+&04_{Wsn7Nqrjng~;z9(4R{Eq?nJ{63*&TWG(?Pnllqpl9
zW5<qCIZfp-7Z=Fc&i68FW0Z~j{KH*Bw1{V&#Fh&Z+`&bQ^URquOJ5*V`m;Xvxk-~I
zQqfN3KCStwxF!xp@oG%W%Xf7yl+><W+r20$WymAXU5K-1&z3tZPKbM%;*00p&>=bJ
zO9>Z1*f;jyRM-ZM{kV9-ofBM~XxOl!oELJ@lhdBqn-eYuaXyX0#Rb!a11jmc$YLtW
z@PZ2<O6Ws{V|@1PnTqn&k^Me(>Xfg_8Vj+G>`or%&X~?Ht*zIcdnODTzY3ppZxCMS
zFTRO;B*TSM#1Ne6@6b|EpR7LpMDm$$ZIfLX6}stDgdsmh?{t;c)~rj|0b)k2biyBA
zT)hACUDnY!PeRnqnLz$eAq)>giG^=kb*GOXS?ELlho(QRv{_z>wX!Nt+IW%R(lV?3
zJyXZ<k64tLDcWY=QV1*W4y&&SgUioCKk(9Ozo!ZXI`u@eg+7MYUo~Nrd3*XNNdvR4
zWj3?@r2b*ONTV)>`iYu<o|D%2VTC_G3N7}#mvEM0QKh68;t3faPTdo`=S5o;Gr6$~
z@Evjdt~(ecbT3S&a=VehOzM%Gg&s**rmM1+(qsmcKUanFoC_|i#R3W!8cL;&Eq5_+
zzAMZ=E&E|oM4lS^uF|O!Q#v#iN2~ekmpy+6d!-BF+=La^fQz{1ilvSv!z+YAEQNX?
zE|gGe2~4R^lDZ+GtK)8nX*Jz+W-^1BKitv4`$>B74W-5Q2!ytMmAXbpMBu@_yC^7>
zv0h2>)G?$opZU1$YFdpn_*+oBkgA4?@uPR%iW_iLSq+_gL<}3!12=o&lremz&~G6Y
z2<j}lh}`3j3x_E&!G82D<t*DV{vq%C+s{AbZh(HmN?!>#9#E;S#5fKWS0!^C`2gWL
zt@=re!v_$BM>?V1Ahv2LF;|C5;|_rpC!a{}YdWKyT0OOg8+LKARZTf4QDxLIilEXl
z6>bNk(6Zksp~VBwsN`2^PsDBrO0Dd1#Ug_ExDyZdp6hSOMV79`oa8P6ruY2$P-NJg
z(4wE*`dslIv{_hj2&X@p5?Z!ei!ZGBABm4JzY3ifQSlvByu4WwrsWe>?4-oWo;f4F
zefy4{9a>}FUn}Gm%>9RtW3-@ixqSVmeEFpkUrDOu$)vLS+KpQ(N-EN+%%_wuRU8R~
zb$zL7^@Y;Dpj0^b6$)!w;QLfaw-Z$?7W5lF&b=tD>95^%Ydcf=H0a%_4Tg@Jf@o2r
zVCJ~t?)kh55F=BmBNiXj%ehx=8_TZUv=v?YiEY*oA4%)?rL!la((fgu(p}~2->n0l
zzj&oiJsi1X!IG@svI8Ca48}`gr8HygFu5qgynVa0lbrh8O%X0^SfPaYDywjIO6;ks
zZNi@bcUkP+u*w5>zRH*_Qw9v}(+yn(<qkgv9|~u}DRUOVAi>#Bx+1UARXwK1NWwyn
zI|^uFWh=TCT5uN@4wWZVPJvq4++I9Q+tuf`N1$HzI0zNuf6}QMt?6v7)YtQB`#JMY
zjs88a>;tJJ>es1-U2U=;Z9I>bx8CWa&SP5yvWT6On_Hwu(TbI&0+EU|Tgk(R_mMA^
z*9(5pqC}F)e6}0=r{i4G%}{7r9Z^_9lOE^x{YTO=`>A+#b{#QQrt?oI@nU*SSb8^z
zUIq)s^uf|~+a)~=t-q<zH!6i=3Li66=2Q4S>LjRoUc7!I&#nFOs%@6-+N?I#Zr_V#
z>v!PyB6(4-QYmXyT1}%uoL2D^@3d{KtjqAJSh!}h<m|6mt{DCt-N$`_gZaqiKYvs&
zV-XMG1Y+lcZ-jHTK(ux2>}4!oyA{>i_QR?5e_E{S&V$EtC&H|u-Q>H0%Gh<pG#MP+
zv6+hUaI7g7IZ`D{g32Y`@R_@EJvt7VEEh1WIWSHTHx@3Px$U7Y-<(i+?A7mt*3YE4
z;X4$CN^ic8Q%FryzI=HP-$hi8l0rubX@RKBr|>$Zr9YMORN7K$Px=`pw2bF}3Z0#@
zo=%-Q$z1`w-%_|h#4@Bup%R`7ebThhvYv`1E<8}Uh(Jzt3YFU2&B?`ebMI`lYUj8p
z>?3J~XthtFrMp*Slo@~iAq9+j-*$F(ma%j|Zz}9JcN!3XQq7PqhrSHV`DRkeW=msY
z-K5Z=FA<}V!bpvm<KRL87h{yru{ec03>4iAt?gC+RX!fM1AueE-2m(}7Y|5f6jr>a
zE8(sR&aF8ig^#IlhXv=wD5OQ=4g*tByD@mp7=`(bN`7OScf*_qTUhAx9Wq`-Hywec
zwu=az2Zaj!V?sYtNSu=j_{eQR2_uyI{G_K|WQ<VUM-c{-84Xmtxqj}*SGo&Y`;!TQ
z5;<IG)X5Pae;#=#>n$eq6--l$0Wy<?Ki(-zeFuLVl-R;}ka`AHcbq70_XaA1sVlM8
zZ>dX3YobdLI)x2Fr(j8_bD%PxLZ$@@9lgen>UQnrPyTBWC5fO?W*UX84ir+$9CLMh
zHr^8~57e8G5u$})lBgq$HFJ$Qf|6-tp}4?|&SELn(qx98f4GRzPV6wC@v5^}sGwzg
zEin&9eNNY6&hC9OmSs4-FzhrH>UX}G(d{V9dSm|2KXpMr%;e`Bk$#7Z23!c*Cve~*
z2p4uJPB9_~J($}fL2c3JA>~p?-B1sgYjQz}J3TBp#%1}qVj}>bAzJ(g0j<#YA6y--
z*4KhFG;Ct)4<>>VBZAwFoZ&aZ0=@90lMIRnQgpEr9Ya2xWQ6$hR#4nf>2IX*i7sw7
z-WBq;3ql&1oe#X_jWWRH{ByabX@#(SFOen&7S+$-p*mW{E8$isCDjGLdpg5MxN#gR
zPjtiBxF`@`(A5~_=Y7;Y%3K{PafjGFz#DIEu_K^q-ULDc;PED&N~}Qhzjs9`?YUr)
z@26!TZ|H9X#@u>lDzU|lox=9{P!1m^grqCtu8E7{p-24qpscg#g3NbSaM7{uD)Y`^
zeC3oP{6)|zMMEyJ6N3Ng%b@g(Gm^v4nnFK}SQ+NE<X8Ff<%?9pdsR<K!D&Bt@p8$3
zpFML1$-6#y%p}P@|5kj2kn{euP^eQ92<r}`@I%2UF@Gxa85c34*d-wp$A1bA{q6tm
zM~7Ym(QC*k+&r<*BB#HxE@SwW_~_=x(R+i{d0xSfPvyOmaN*ppHNLD`a*(F-8jmku
zc~-8+5WMe7ywkSId-TK^sm$jBL;L29@q3Bi<jw+>&SU9fzdpu{@z%l<XU<2DK_ldX
z$;4p;z+DSUxML!Bwk-Jfz+ps=90~aag;9yB^^h)gDwW0kIC=I0;fOdCdWyOz_r$^j
zcTp6rSWEg1Am=`JJSd@K+)_mg;rQvZGKy)|(p5QgM_MZJnzV)R1Eo^gGH9ADYJe&8
zNh`+RI4QHp$7xM`lGj>2_L^2(ETFvn^gmQ6TN;POVxUr7__e=!Uj2%o-^%;fys}sH
zF;t^UB`m6yEKrpBKKV`hB(L8P5B$=!<XhJ)BWhQ#D(L3UND2;H^oonOVZ$WxvZ^TJ
znPn<bB71(Xdo6tKC^6F2kUl44s^ob6_C2zcZ7j1CFCcdQ70e|aL8~%6MvVAt3q8-3
zvV~!5;#J!$OXZ1r@1R27C-PFz6B>oq=Za6>J8c`wZu<9-bXwa`>>A+if()sW8Iz?I
zdz}g;J;eFtgj7Q`ir}sY`Ul{Dj(_ajWw>(lj{9PbX&ciPuMzpex>Fi4ojwy5{j&v$
zh1Guf<~?MQOE+BMDi(7}H0UIG`n$KR>!GST<x7H#7}sv!^O{PesQ>^#07*naR1j}^
z!nxr!=~t20wLu|kuD(~|ozk3wm3diSDS}m5Q-y0Ju}jt{>N<V-@}-QnR5z(=sHmr9
zJC*fRjw>ZQ70$FoU$khEq_d&bxU%wRe<@THtG=)eTEf$s+nkv8f$w(e{mwgMG}cMQ
zx)O7Wah1k=TyP+VKCRv9D7QzC9&+))v_IY%6LB^w*V&HI#{%o1MZ2+23LiBmre#!L
zd8Fk%pL^V8K<Xnd@^DAAQOcAlV?6aT%>KB#x_ZdR#Se47P`I;$!bKUi2*mer&X2kL
z9xC<Xgx^~VV=uB=O(rE!i2wc4GnmU9Pv{ZSiSK_Z^N$K8b0uNmM;e(uH(pALldQtx
zozys_N}*C%DY99YGT{5ESC>5bMp~31<x@%}J2~{prB6DU$-=2>jqLHwb;U!RS7oX6
zHwqaE%!yb1rXI*tmtv}nHm%p#w((k3`63E`0ZJ4$JSX=)h05{oKlMK`1vfpJ8Q6a=
zNGxxXNm{S}D|`dc&ms*(7dOm=A}M29Op}%V{PRrgw4gGdi!0=?Cj}Ab&nU{bfHBTm
znDb7$XQObD!c=d>^w^Qy7B#p)k?-f@QR#wI{*FSSpH(c#929;8NKctd>>{Z?|Ey$Q
zP?S7W=o4G&hByZ95V<9k{nixvq42|fao5;a9kvYu9)EGRsZ-bDfC$i%C}fCV%5d*M
zK;`f&p+Mi-I;*?YwxVZYc{;$|bnykA3fi5x!g8B(MBE6?2`$Zy3V5!9k|#{bB=Mjk
z&Cd8LspJbdsc(jFca^&W6n_}=D&8>Wr4ttl9es=!&TZ)p+`E!EdN}z2Hs+_tc*R?9
zj5inF+9s<z-42!KT<xwaocmT3&TFYqTOd}|NA0|0uBh0~76F@EW%ZDjT=3;{HJ_{9
zN{;+ofi`C?m_+oSa$l@YE9blPQ$XRvwM`+97SMc{L=}rI+)_oFq<V`lVqJ^O?z=0v
zFhlt!9Ee#H&ZAF12+h^U2My_Syb>yG2MP=9<*QwUC4C`e5mw@~vL7~nilm5fN}Nzy
zg|nW6$4&^|`?Rvk8@F8rhx;3(NaifhP5;m#uIs(~(&C#|-<Pl5K+fV7Fi%kT^y=8!
zG`rRF<VlkVs-9<<H+i&szQrrofZMYvSHup2q2ngY{?j2ZE!k-uZ%#so3hiNG;L_C_
z=CZlRdDXUiMiCg+x0~ecC*4k!hArg|1v<f{-vMPYU#fZooH%{XSj;`#z5O3@r-6Hf
z2+$t@mH8tF^h7t|tk)=1#?yB|%}V7lykB>jv|!oa@|Z4l3S(9ou35PpvSv*09v|lH
zc|qUvNcb_xfp}u)1fA)oa88O5{oOvMnbGy`eZg5AUzYJI-MwNn0dr1v&1Cad)cpDL
z5c<Po{E^qAvcg-1eVOUD*k)OF&UvhDnnB)2tySZbyw>VbX>_DLanF71zx4*u#O|Zg
z!dj<uk;EQ8TvT2j{}mjR3y(b(TaiDA#mb0+k|+U+iN#`DN}K$GSAT^o=FFKbKCIi>
zq-s5{!E;`Dadhj{9u2d{M0P7&xi-0oEL7r?-=t6ST4>Hgv_1ZQjUCM86a~$M!#2**
z8kdd6BwvRkMj&-TNMDyXE$E!s&DuY|pxznWshLbB70$33i=?If()HUx#rUDsvye5N
z+hR7UZ)_o6waqr@jF)se<plN5Kil?7>vLsAZtMds<S8G7&mJY0Z<v3NWx99TR`u}6
z@h@4s4fBLkSz5~H7LJ2SX;ZCC5vj~4M}PJTO^ro7g!iqO?#U6YhYThst2Xb(>dpV+
z#=k4wQ{u#OGl=+MV({UfK?GKB{a0Ao7nZ)2R6LdLUBpfYqudeiC-lQW$G}%_-9?}A
za|B&dO_VO=wp;Z6<EN75CsDlE#$ukrRHTwd#;-!9vZ+$mv`$*olUrR0xzfo|Px>0B
zDJyuDPN$^G+Fiw&J0=r+_Y~Gmet8O==uVk3Mf%+$2R^O!$@flSyius|r$bv(#*qJ=
zlr<F6<4{PQlQpZ`cN6*dSFBiJ*7}`!$$3wM9;NIiWln|+8JOg!G0F-o{Gmf!Dziy3
z#Q&EsUzXPM#GrTY-X2Q%W=tL`5GT&g&eE!$^fbz%oqY2Ya{CiQYcgWQ2uZ0!x*a+;
z=59th!6hvb>!y%KN!2MM41ehCn3O_B;o<-(JE)lFb;u*HCoXQ73aNN_&6BQ42^~6f
z0nV=qeYR~}h*9<X^hjfna6+-`mHZ|Q;z;*G>*$dcQ%V*A^4)(DIu3HwQ|K&~)I8+b
zkM&b&OPvFS{PVQ%HwwSmN%_)N?4oZV9HdfzQh(8T;ivF{mkNEqlr{E&VXyk_rc3eI
z_tW>^dRv!b+Mcy@;zVhlFR^gs>&7H-UQim5eug@zNy0I2Zn5iu4uyFg&D?(<!?xX$
zifpnFSW2TKStVJ8e}bLub6~N+eyp&9A0f(@6ApXH_fKO>>Mm*u<@p)G&u>mRUmbr-
z>LH1aeaYuf#XNlx^2i+q<oP!h>bu-@LGRs@j<~ul_NEaFN2LGhFMJA+a;VSG-3QBs
z?*Iz@7d#dm{1obg+;v0ZTB2|#hBXN*-ib^X5v}FdX+A?hEA)Mam(Fbg0c8nqA=zv+
zoupFkh?T+wVk^{gK`}$`+L^_Tln!$%36JYkypI+9W!;M>#YW)>X5S5A`Sg<q%~&Zd
zt!Q08Mku+-Wj}I<t1($RdEG1)2ZokQF1MsT5;vu2LIJI|V5!Vp=_I!;X@Gc><jOcX
z^?9=<rH*%{VkY5qO!R%L@CU%%2kiSrp-dlBDW$inpaQ&A!NmAOZ1FlE6j^@>rzt&!
z6W@WGZ^?epkzUuLzbuuFmD1a}sc;dXtndrKdI}1DJf|4pFl6g(sr*-q3|-e=7mF6*
z#C9tuxs{A_GJf}6$rZ;31job00&=zG5Q=-F@MR&Hco>q#hz~khcGq3Wo6mWfw(q{o
zUQF=ASrfPY6xMQhKaKtHN=UUsp7(Q?u1MuOAEu--aoM~b9a=V##WST#BbEEJ1ZO?z
zb7F{XsDnmNk`7dBSFK>S3BKs}7&H>m#Wr_x;?p54c|J3war@W~FDyJR2xq)cpFKz1
zI37R7!bfmhOgVh~lvE~@f4^N%!6zwp4v-sv!?vC1Eu77AVSpbMEjsm-)IwC^Pn)+0
zL&i=<4dDd0vS>GgSZI47sAI?vpE^Y{q)U?$q^~Jaq#$S=NSYpNVy$h{a@-xjHu;~9
zYe`)+XYq23m^fXw*IQWAHx&y0vL%X|7M16w4uToUg%K{WBovl8%h#-TFBef*@l_IB
z|Cg>_NBdp_aCC<Y{@s6Aj{T2D_1vkJT6XRQr^Jase*VV7pMg=P&Rr;d4j2{qW2emq
z{SM^Ho*A6)i&w4*r@X&DD!fWJ)^7}Z=UO!8w-lxo*^mMKu%N2v7d1;|Y!T0`p)Soc
zBAvL;dL=%|>s4LcO4@wIRk`JJXjy0RvGN@QgkyEyQ-+pJ=BfC3;TXO7ifgExDHbMG
zNr~-OU!n1`D_GGmy>$BJsem6M`WKCZ27?C*KG+uGUf}kd^=)-}UW4bn@_Ha2@}Awh
zaHPGP!pFDQh+oS2<TvS)yk7Omo`1i_zPHL77x|{|lXMwW8u?Xhb-!2rh!;C1isZ?P
zr7qi%N$}g#@_e*#^16D<9@Hu)=>6ivl38!veJHu>7p>WX!>2BQj&<oU*QXLe+|M39
zeJ&lhrgBbVwQX95lRuvnH}gjHLcv<C(P`*p6v~xZ^2>)87Tp)F-va4^QaN)WHl*nL
z61ihwuW(v>LO97ynbf0!i0ZFb<5v09k$m=a3f!S_4eU938uM0eK)b;c(0$Z&iBrcW
zHPE+hBbD9LF?P&fR1BUH%|uiY)HhQW{w<vs*AnADcJ8uV+#sD%YJnT2b^md4(WQ6m
z2J)Pi`%j*~LMgF}!YJXxg~hNgEsRk(bm}}(I3<$t?>~H!ShX8C0h{KJ!{cY-duG}~
zx$Ce>sUjZAnV#6{dh|{`<fW&=nf&LZbRicyt)xlaQma-i$-hqOo026<N{f0bT*=q&
zmGE~-DjqHj(6MUNs8KOw$PlDSlSb|cpvC#qr%&;J_O1dTitT%!-+RVWOh7CoBve#H
z>{je}c0aqjyE~p;Ji9xe-QC?{As`}Rcl-YT^W9~Z*=2XxWfvvp=rS|6=FXkE^PO|P
z1517=3BY9!fkif~wXuH$;hjExx}d9pwKoJ{Gc4T!AfpZ_A+Rm3({Sa0C@A}bHo?^o
zN|_ZaR*2xXI8KN;$0*x{SOp**5)wiPo{IiJVUZ?H8ZoX=l!$MNa7BW=Zgh@GULWg-
zCxCbeBSwrMfb+1n2RM)OhVhp;>b2`h+xQ-kf(#=t^8ESp#d*fHfi?yY9xT9p(ER9<
zqD6~FN(TQvD3TCd7xYbt$p8y|$<lwxk|lyx37#6b9_rMoBl--=A1H({e&Ftxg5yTo
zfsqNu+y_QBMd2g@28)AN9pEuy-XXXs-m3s^5u~&XgZ+>{bJ;N!ZZJ(|ykuYnXL`(r
zFe}bI7CieB?TmaUz&chj9srAL0K~BDf6i1sFdl>e^aV}=B?5TW<q}n0dD*Deqw05c
z=KDurS_A_I{m_mZ&p7BVV<|P=Mbn1*8P&eE^AX>0%$^4{mO*xmA&f~!^EM9+4dFn;
z;Huww^BIk1&xYI?Q;P9X;Ydjh=4tN-@d+l>aOc<tF0|!pn80OV0T08=X3Wpuh+`U*
zV@e-g0;5T7frUH*C8OWXSUIzll?7m^kgg4!!%@2*P+N|-R-EH5RLN;e%lX>CIt*u#
z=gQa)_K)l8PS{5p6mXCBboS<e#;+L%3KH=rLB~@zn~lJCCLauBetl3rL9r1m>f^W!
zNz{mpq1sAbR3i`>`HQNClCO+FG)I83QVpq?!&>m=x98xh5O9NaFO`L69pev#y#a!|
zeP~4|VM)K5!G6%O)L@`|kDto*672zig~08}i#bsNzR%$oA1Ge3a+G#S+AxQHC+01y
zz393~!V7LNx2=mlVp2^vYP-gd%ij@Y!Q(%Q6#<evlJ;>(+tL!1I<p14-;<Bjn%j}2
z83&MUzz<%CysVV3Yl-74;gUxD@N$9DW7UNxtkJdQKscs>dgH3OP~eNt)MC{wu`7V?
zYg^<J4{#(L;|I4yD};bAzV)+lboe{M+8+6M@GoeToq0&LP9!pAf<6P-@5>5^QZp}!
z90ZmbQQ2A0@5qY&@>NWd0d6CRZLvc6MSn^cD=Yy0iasU9Hmot0&Y42>TXmuwh0BUG
zE4EZx@y|3hC%$wYN(WS_Y$-w4f}pbafr1C%xJ22~pjI`SIe)2`L@oE7CRG#5e4e=(
zgw8}QyYvy<`E8muq^mdm*lS>x_@Piuru*pMt%CsY&s?}n%jR(KF8-KB+Qx0WsB{0J
zg2x{Hn7@2Ad6lY6A0OSOUY*)fjlbFmYjkU-)`3C-z&rXc9fP#Qehlp!`T7L3H~`8=
z^zT7icJ8J5Y_-33^ESbMudj}-bOg|aCq(nky)?4Zvl0Y=J~;WIII(ugBk~cv7I6<G
zIdJqO<zj`5F*1GJaN+gPx_du0PbZGQb8FRycvShu?8DqL2zKUdos^Z5R^~PpsXn}C
zCLDMuKyW-8z|Ftq+H<~_9mIzO^2RWTj(fSwMpmSh%uEYThw?|oC-K~vT*E~u1fOyI
z4XLay@1ICJHjr>K;Ek{VTtTD0w1+YwfB1(66r*b{J*7eA94Hxo;K*gk*-&<0O(bZ)
zkY0+FsXJMzy8L{Y#Q}XNd)+y(=JxH|sczl6@@|;CuAEmA3h(%KUU4DJ^PNM@^4SW-
zxXBlW2^Ez4*Mz~+Dhe8c7JQE8`Oc;}bLVU2#>^((v%m4rkN{+n%J`1#O^tetpk{r>
zh$>c0p)+}4X94a5JO*EWlU}1l9zg&8?f+uxmo+qS(p*;FCs(&)Xn#^(g%SnX%6&WK
zui1v)T-i&8_U+T88v%@-*sr4imwiX|p^^<dQ~!zo(8BQpIJj*Yn!RGP03^XP559d+
z(v)u0g*;rHgjK&p8N;&>QBLXymWK8%FDGBe*p?~mEq+g)y<{qzU9@<^PO4S0Bz0?U
z-m_XhKY|B9BMvG&Z2Cg+LmvRZFP=D1&5eKX^hKsca#nZm*!insy^xn_3ceIp$!PiH
zAv9vfV)CliLbL(y{zW{-SsgN|+yB)oef#z$@aT_VAQ(SbLoZ#rR9L$MGzT|3fL=*D
zcI+tVZLlrw37(Vd63o?Cty(2`*r6zZb+eC;56zxEo9Px)2rKn5W5$Sm0qh6B4nVqe
z>CyzQatN&SVWHo&scKG#v@BV&2r3=~(3OI}0`LVWtWUsOzjf<YnlWRBQ04>tmaOEB
zIYy~n`L%>JX<J%?Cm+fVC^8Tb7oH9+TC||%&6^9r8eRtA>(`d3j!D-zOZ$}Cg15r7
zY12etVgTn5@Y(~Ihe+k6rAXBC2Oa}B7f`@}TOW#n)vH&FrJgD}lMy%>=MM@7#7uy~
z0}7e~1ysthn>TL~JQ-jhf$IZ-is4Ox_yZ6qA`lo^+si~c25G5t=f5x!mfxUg0evk1
zSKt<T+1n4mYhJeahQOkHBy*aNW^fjk_YhzzlWMt(b@6PKy>2j%@6Cpi6WRC?m;%I`
z=hz*Js}zx-eOYn-G5O5AKtZEa-@>90|B0%uY&W6`0CDL(Cbc0gqw04Twn+EmIe+U_
zH>D~9Ty7Xn3#8poUHev7)Ukqb!9Pf5W?`67H#_y-a)&DUs;~<fJT7SLA@Tu;cU<d7
zwHK(K;hBLWG0>0mVJJ^j*6{~)Hcc3S$L|Vz9V|Z=O7l(x)5ZQpbd=VgR56DwD`gyL
z=Fx{V_rxQij2TtcmEa8k0loEprd9eg+?!*3bvF^*w)u)1B41!Zt@5a`pk6+X4D$Pv
z`>2y5AAyC3Rdx~U0vl@J!Pw9(`ing}Ai#zeHqIF(HY5rwe`0I|{>Pd|MQIXa!<%p#
zE?1%mwRL?jfM-*W8W#HeJtZEdsEcUscxXTy2tfdtKrRDtNEZMv3WfiHdv|hvQE79M
zAsd(aC3ZnW+w#CM^7_*He{7KsZ7ewE-Pz4Z>Wh?b>=M*JkKcY3+Dl!1GPWIgIrI(O
zcyca0d45#u7;m^$NeT)><>4s-D*z}{@M6Pc!dUt6<%c?eF*Lt;c||4Q2%L5!2^~9g
zA*!MV^I^4-oh^4lZ=Ew(&Xi0{68-LSfZYFnkcYdgQC7LIR0mgkA#X4GgSl1Z@`ibF
z;bxtTt<ALs`+E864W(ntJbcJN!eB8M5J(Q0GYP<7yN<qX><8SJ@I-)M-JkpA!k{~>
zXpE8Q?pxmW$B&<cVnd#0sdZTD!y5v;_vQ*>C%~ftfsf^J3Z!<U%9qN?mkxP~(xuB+
z^ty1%%Nkm5-MlHezBozE3uZ7>ELwSexB1ih%g?A;f%F`Z={s%Z0N$7#Y-r&jK=N|F
z*97Gbf^K6S(xPB`eijj}Vnt(lR-!IxoSpo_K8SpM^gM>99NHgLi_V17n0-OyKctXQ
zeqo;g?T73Nq(hza2udBKNm66s4WU$+Q`b#OqFRe?2x_+@T{LOP)CJ5~S$QFhZl)|i
z%T}#5qZK0MyvEP-ii^TIOD0-U!%pyznst#xQCYcvqZptlvJp72@rDoz4t*L$$*8j%
zH*7$a^1CsZ?Gtrj6H34%?_S^|dj0M_<#2Iig}<g^Hz4@22>hCbEyN+PiU-)PUCx+Y
zzTtg*XG#lLy+fkfx1l!CmYC~=&w;=f|MK-)HKmTBy5{a%Y8(BBw}QmaFQr>z8Oql3
zL64r0P1;oC%*Ggr5=A}NHD|xz)&KPQi|}knoH*(^+?IXEXv^_yv|-a`b9O7hW6+xb
zXpTxi)dG%iXJ_X~MPXfzV5zW{mh-kpDySm?^9O*mu)0UkR=FVfF08?!L;z@uz^Jg;
zM-bDf1b{gd5%9JHfG*9uwd=KQ+m^0hzfPx5pC$mv0EuBWuPx>rqr9y*98*lp-)+LZ
zaQOIOEnlcmA^xU_{OusGYwR(}>u9fI|2R%i<~TVyseyd$a;9#hZ_%Z#L7xyL848A|
z#sK}oHG%z^OSK~%BZ3W2nlwqoGLX`uvxJf(bnjd>y>E2d(5?f{`8zxZ2Ll4=Y_#Xp
z+U4JW`A(M^7Xi)zzJ9r&Enp22Wv)CA>xSxTmyK#Ws(ypwC6w_USMH}amJ(GzquRH2
zKIn%qhQL^0xFe5CAaJDMa2zR|mPe{@C{HRA)e_zpz^I}l7>E0e*T7O9N|&f4h^bUR
zaeOp=SkT?49|V=Mb_@fwtHOpp@O<MpwPk-)`-4?~n$oRBFz=|!D*2|5K-4n<)7w;P
zB)}2)@BjWY#&b;P$Vj~PQl>T!7b11*iq<WamoK3))hBqd<;$YdSn4YyrQ-pP7l*k5
zjW2$d6!a~L3BbDczO~0&J72yHYkW`-MMnVO8w!A{t4*Fe2*r!8JX6x}V#PISNrA<O
z#G!9&yRcN32XU3x(*+hw`Pf%_J6mZ}EU~Sfrcpr8hW35{;%9Tss;9K><mm2O-Wt3V
zVoE&1sf;Yqeku-xmfy=$Z&Oj_>#7JP9Ii=o<>L9XRIj9Ax5nt~1A&^>G0j+yQufrV
zw8}yU9!^u4yI!Aw$G(q`sw~uehnWs7Dw%uoF^y(`UYEe-^JrI1(ER~>6s1X)NMk|`
zN)QNwK4}a1zw>0)d|u%=K)n+a;%%RTmK`u(X{el6qWE}TLCx^;<tqmBonqqhQ@5$y
zzv`ig_6q0(KQPez{P}aeK|PIa9N@799`})U2aqvaR)<LGQUa7ma<Q{fDe>iL#;$|A
zT`p1WTV7UsUArwP7)((zn6q!GFK8o4lBhursk|vmP##INN2b(^Zdoa|fU)((e3-9E
zb47Js8#j*u@4CdjuDxFe2M5A0I)d*weUvE(VhfO05`ewNfcYp7a9QH!m+Bi@q929;
zVsnmBzF%#n1+3jAj(mAJUF)VEldd+fe_ix9y2_f87F{qO`#_(d&@fhv9WP_sv*`U~
z>`sj=hgfz6*rMN9K>*njOueE@(3T-FRM${FU3pRU8!-&^#W9SkpSt#EC{1`k=)sBv
zj--S32JIQjGqgW>UUb2Hw1IK(*3)nKKBRRiW`MM<y>z$J{FajEtFR%>If6W2*E;4z
zc!wwwGb3R5ZjG6JDy_!D5%}$YzlA$3#lwSB4(X?j0MC&%z`N+~7%x^_lZq%K5F17S
zN;q7T=1OSTQ_7Ii)!b$x)%Sn?nLSU4N^pD)NlG%+mmxsEh;xK;6hX9g0qjy+(z2|3
z8cLTUv9zhNgyp}sd}eDESP02Q@)FWzHT2~>doQt9kz7z-S6O+P=&omAE2t+=pGCL2
zx}I`gN$vfm>s-az^U8e8>C@={C{kYlX09mrubBfFUOhwF<bp%n)>c!QZfffQk1g=H
zH+8p4d1VB2j)0gChkJVJEMR1!rP?uacd@b<lm@1H6~xK|Q92dv5l}tf&7aSUcB0)2
zrJ^zdaccxr_>O|h#MD_+0R5O6*#y?EITrfAa;V@|1v(vFL1X0o`pj-gci8<1QNE+1
z{gKyG)=e7$l?GQ6kTb0XRq2?p(2q_(VxkTLwZiJ&K2t`6&3|T>>>Ic5P!2a&=G!#i
z+r;3`luUyoU?~gz7jNDf<w|L=GdeQgvfC{rMbe1LmW`F+=>e$CBrz7e!t)b9V<jDX
zXMz3)F0j&b_Z4p^Vvt@QBSmTt<%3^;kv7E=<;~^t*Jx0-k&G=qU%q)SONrj9a$bqz
z<9Yq^<qIpyV%{Q2xqqW~-)YyvbNbV#Puj^QZ^x8{zR3WPKySYtiQx*LzkH>e_W~$8
zdkt7c;elsL*%<D2f@|JV7R8ooCoZ1=1m!g(p!orH4+8KaFsmUkrL3WLl)Rri0;YJX
z80wj5$54R;lxGqa!?=x`or1;0waY`X(2v<OX84OeX7{bM9*;%<mK)~ltAGB0v|W2!
z1Wp{y!DI|c0Q2crnR}JCw8>2fP-CuOVm#~EBih9Q8(%SX1R~21tK>|&xXQ3E#0Hpl
z_#qwVz_o~~j%e&XnHodlb~}9Um56M9E#y7(N?UQj+)Nx8*jjWGJa?;@>v=mzhYw>u
zf7ETl0ebcErkQ(e=F|r;XbDSZ)%6$`ugIuJ55ajWlqpHm#tt)B_}+s+DpIi)9p1K%
ziWkanu&gnexMnBJ_=o=GAh$1Gy`sYTz3J}}1NHLOA3b@7MopecH~j91m;)8dmZlCZ
z{t{gGXlv&DCA4Jq24Ue`Aa5S(-mWE;D_xwf-?~HXdqo7cmG=!)I5Q`VlqbcBbxU2>
zDH$}uq{z}@A5K+vF6f=$j70la9+?kRYxfiU_Y)2VYbT3s7pBJ9)n1e8ND9(@lLL$w
za7Zc2oe6qNF1CMgT#pEag47_&Uw!<ZLDc^f7T;K&d-#sI{w*JKK<tyxMRKj!7*bzM
zSt{q1C_bK7yoZt})$NgD%JmUdS>^tXYJ`l-$GL*lq`4yIAMtbgp)oY~gC0?d2A$}@
z^66AGUv7;&?G01PYS-1Z{fJ{7jQeK=9pa#~X;P)2KYfbPz>ZDzp1(=+SJTElhbi>w
z3o2RAi>j9`PW7w$sGDB5;~@RBY75=EAIMzzsi}Io;xwXLD@ta8=ev3*zi35uJ(_<(
zTB@C2Gz>pwxA*Sd6G2*!A3sjTix)SzPgBYoY$!Yv-1-PS3yyd2`q!*kQ}9(s^-KdN
zeg6FUsdVYmv|zyk^{)Q?_g@i+_{fnXRH8%)5mdNNojU6Lgt}s>`#`jlPzO=#xZAA0
zNu}MaX>>Ki9>@VNK2qg>F42~@UQ~#KwVEs9Ta<&z?!Nh4cp2wsIvoUn)r-Y2m;bcG
zL3Hf?D@x6Dpp`fvZO<}}YAi<T&xl?A^y0%;npH1{ltkA;-qC<<_te|c%Es1t)CNmZ
zT6|nu>N_|-j*TUbNdfdNF^WnXf*(2taHbd^S>O8E1VA3uj+_{5+S=Ar4R%FUKC1lb
zM}is94`vWQs|ex<u)gMkYjmVrzDO3h2$IrfwI6+DPK?o2v+_&g8x7m_fL?$4#@jA*
zodGQ5*Yr$B)9bj?zS}RU#j0DhwnZ)~>YRpVv*G|z><5;265a@if_{WwMu)q4>jcH3
z>U&J)1FXag)?7gw_!l2#)%=<K0UJSgqVE`8ogL}ejt!LCBb&Y=CZ&&^I-3D~Uz#*>
zFge<1rZH1yQmH@d)Ai%~A_e>n4tY$K8?@kHz<Fru=%MtMgQN8xI)<*@^rO|>{~7a^
zP~YKWsb|O5tZXPin|AD`TFu+h!7Xdp^B@B?sP(5t-$I`}r76A(sB*b78tHLm!&28Z
zCPXnI2#5uceBP<YWXqH|Nf<02wmZ<`yUNds%ig8!b+(%}dF1Ixg=}g0`TRSL*b_*T
zYh<OMmtSc8m1oqZu*y3Hfh<Sw4Wc$hGE(uZsj044I@;y;lE&>1qNc30Tz2jW2gLnG
zW!xgDgmmuQj6_SE6HBzA>kyRls%un?ON*&<V%*+ME2W%g)5gnCX@D%|ikN@I&*_K8
z(6~5cro*ddQckz58tJ+=Oew3YzC7*A*RRyD`*8a5^&3qY)P<hEdP9B2&!$)G!8Vz<
zwFN-^7~f^os9HH1(xoWvJamd$^czbEC|j#SN!ownJhkjUj!GBGOLNEc<-o(iG;-z=
z%FJ<#y0xgM-CUg8R$Y(gUyzn+C(aF^2_1HIb){3MPKm&}QB^agY*cN*8?1Ki+JxoI
znKKDNiMx02E`kEjm@&g(^>*#ri9o}}iWO6ryL$C1)v8sC`t<2T%a$#pZQHg{-MV$@
z?Afy>#TrmI7>`ygbsva!;?cO6b145b@DKobRBbrfB%^KZy%>uzjlW%AyVDtanRzsb
zhBJr{j{d0(*4LPSg^qSFpc9K}&5DPwOzk;_6%W8-hVH!2Sj=~!i1@}B%`R39Tnc_m
zi%y47nQ)^@Y2P<P-%|)H1p1bBl9I#{-&+YmEcG3n5XLGx#~AqwIy$W=brSpt;3LS4
zIIS@@et5u6U~p&O?H81hBe*y4PET8|J)^-DoJ6qKg$!VM*{7ryg)&mdb+@UYQ!2X7
zR_@2x5+0G%yOy-4>bcX=v-jUbJ7cP4wTNZAFN5mM3T71h#lfdy(5lwL>vZ7GORAR3
zR%1L)K6p*n7_^5%0or58&Cn0D=3*EPVvv5xnaAAML^P$Ao5(BWnwEC*OJzBOSpf7m
zU3p3+IjwPim6hsORuFXCbeHZv{YW`8TIQuFc5Hz^xP~n8L9MzT<K|29+qZ8tWb9<x
zzV`rSW?G#V4eQdH&D&`F(Eb9PpUGf&0S3xDwroO;+jXPjg$vN7t2gMt(GzT0mWBrO
z=tT8uR27!*L&i;}8RLdit{mor`jcnR7Z&!|FC-Py{*<ZSfOhUbM0NhGta0`a9X~}s
ze*B;vYnG_#em3veBdqmNaOL{VR3NV>TlfzW^~#khPFr{FC*MWOY3ZCP)V+Nxb$j2x
zf2R^v>(ZZ<%F)n1-PGxEX~j|;`WDc2=^Qo$IQplaJN4RfpW3av%hWr+Q&kUZ4wU;$
zV^V7f>(Yk7bd^PrY*PYdCrk1>E&D4wHC=g|%K2UseJ{s6{C&%)g7Loj{GImSeMQxB
z+K9Z0*{!L2$;>qKcnHls6-s|3`i(}i=aX0Fn0wAy;+$Bb4P7TzIj_1##kjPXIw!{M
z-Lz85c{Xjl43!4RVy>t-md$M?cGUm&?K=&dzL0hvK24c3*iy4vm1y0LLo}v$TS~zm
zW5cE|qN#(sGN|uNO?!`~Vr;2@`PN-Jbm}5mr%6Tq+Wkd!DwU?N=PxzNnsanV&R(WV
zH}48%LLsl5qN!`Q@6*!FyJ=L<Hk3S>DsZ~AkA26_QTbv81Xz!>qTab^`IbF&i|K*@
z=pQ(Bft<5sq+N?AOI7Ibnai|b-S&6^=!^NV1w2nw*R}SWUOi+9W(9NfL4yVfklWtg
zo?5nSNvl_{ril|LQi>EQMBc1fvqaF`h7B7E>-LKmFVg<~`^m<}M&#G8U!R^ld7@F)
zoB=#^=n!4Fa6u>&iWDg#n!0l33N2c+h{lf}FM|H+YJdLx`E>5wIYAL5&tI`(1tm+C
zOw>)3D3P$<U$}4~ZQZ&x!2|l1x}VUFu6eTjJj9oEU@_zO1<*b}#$wnbxM98wwE5~&
z8dA}jv6vLJ;M60^3oNE^X6mr^max8uMLR6vfyHzy?Z8-!9X)&ZRkSnqPdAHTF>{VT
zWGp6=*e}K!6wI|2TwyHc1!FPkG=~4A|0}k#e=igY`5jc=0RZV&UkIgv6`bimo;&<=
zLgj({hm4X`Y6MIVz+9T}{t}{C3g1bHV`t1UD}cT+qokCHas==Zq!qg`$~r%zz@Y8^
zwEpr_eqd)1T+H1!-xZV<ol0aDyTA149Y4TT!5D8o`9RxlJZDOsR03>Ya4LklY`jC|
zvRTuIFF)vX;2V0&0Fo{-^bv@cwp_yah4k_3e^gjT-Q&z!X9#%utC1$4AMoNcTg?8C
zy!mA(iHmJA`oz5c4_|#Ie||}5yPl!F+srcu0hqs?k8Ag>=OPOFB31y{GS{?62E+W0
z(!PG?64#KSu{5?Xs_QZ4xr>eMk16H44j4|$)@~Bu`SX{rsAbnaVodsWX(yUFd-1a1
z(8tEr8@Fh~)}4Y<r#;izfX1e2$DUM$E!p3H_(;c2ou$`r-|5`Fp>fu}W)ZaW!^e+Q
zqHsZVTW-!yVjZ0R+JJtw3T5fW$$bLM$9|5VIZvSA>D;=RsPl?ld_mbG$sg?2Ze_*S
z>^*-eNxhuG$8?=(d)6=ZRj0+Z6-#mGYduOVHVg<P`-ZJdRyNH+$rJz1?w`L2su}?L
zu>J+WA2L$)k%*nj*RkEHcxLg#dXDr~k@ArqT`9my%p3g+&@FzRWlWWXgX$(xH-%@S
zc3=9Hb)XJRId-3|_dU2Tx+d1@1}v~*iF0C!Hgp||a$a?fig9U{IwzKD$55}8^K58z
zOv}@L_nB5(RXRR#==T^igI4^zm)bR`PS0Pzp*902iup*d)(ynL9zTDT-n>=OvjJ4!
zwC4x|=DDeD!|F79`6girU#VmvdjIhg$;)bwt10u5`#~X;BynOY;2A+@;Ob~kA3uGj
zz>p{8;p(K({Int6)YLiP@}IhBoyad!STzSJkw1@6L`aojjPM}%u~1k@X>oZe=ED|<
z<zXrA@VGpd=zHhRooVUPrPQ%wM|%GJIW=$IoG>>1`t=is28x>3uV0IGfa>ektrMU=
z%1@s@T~OClty-1dv*)V3tf(OSW6C`N`ktgolTx8Vh14y$ySvke4<9HXAb@h_R4KmH
znY;q7|6aX%(XL&)sBhoC@+>M>t{hF7GDT4KfMy5Bi~FKNg$f$w65@uX?kBXv{}Y4^
zVyc;Sm%^xdp-l9WE!(?oye*Esa|xAk4y@W!F_;bkNWSg*Gb+RwNi#0L;A9AOUhl_P
zj17I@cv`0e-iYrNT?1#zJrnl!iwcWj!$(Q7GPVK%exeQNKVYmyg~dd8$Y-@pF0hzj
z#$vMZ_r{)1`9%Hd^Dc{}{0}^e-+lf;y*A&Wr;G!^xTzx_zjWO4vhi`<Qut1M92-j<
zlLF{lVic7&*gF5WSjVCN?EP1^Fb$&~tl8~V#zCB=tPFHdWfv=bl8pncY-*KT@S%A#
zJw$^gH-sg7QnvKeC!nwyy*E&2wrjd%bd+gy<N_@%#D+n2NdQt=u{A1OxU>Z?ku=QJ
zCh_GXA8uz51dIgWpC@w!*pGA^=e}nju}Ymw*b7Z~n>Dw^HvWvL>}(NBcy)soTCtP{
z7a!%gCgYlfJ$+6~*KDN0y}GEa(cQ8-(P-wP*C#o%yVJpc*9h*|!uh<YLfxix>+XH}
zgD3a;WUB8%>OXS4&g|@NOyhK7pGF?|_#w8o)?$f2>De7U1qc0l_WXr9O<J+BPD}W`
zVgMkcM*;C6kc_S7y0&d0isbglP6v;jq>#`jlsRJt+O_{MJ$?3q(x%qrM11@19gUqb
zo4T}aPFXU?U8@F5S?CMJfHcS!T*AV4{DFspYNvnMEW&DN*scJ|ozY5w`4-$&w3<Ty
zJbChKtYCsHSAZH@&(w1G7Q;C)SCsQ=+z7_C^C~|4f5I2La{n54{X|s?G<^W(%$4|J
zp%1HiSh5f7(3A#sY9@xoje+yAzRUFv5y13)%chF?NI_2zs?xd-`S}O&d?Z2<t6h1<
zlJzuj@;vQiv7L<r?q1%$L?h3i0e$P#DaAgd3{d(&UNHTRwm8{m6iJKL@1Sm@rVE7x
zJPC@hr9PyZPf@YNpLc8s(PhL`0!RPMVLjC8ae2j3mi1y@Z2?RFxIC6&VPUjn$r1r{
z4;wa2914K*5hF(EJ*-^0atYub=b>=n!c?|wS-N%W7E@JfdXmT+0GJNC8(o5cyQnW0
z0Q%O})*AN&D19JMI%vyBHqdF%paHdS-=0dAEUBF>fcv6Fi_(M%6T}Y%5EB4g`Uzbu
zbsvcPG64)Gu$Z-%o(Qh_UOrCZhyk2WITWP#=mi$jCJ$fF|Kk{gsi^+qYmCKwAj4P;
z$B@|lKxdn-Y?yScM*)pIQ?VG7Ps82<ax6wN#(c@K5Tg>*hoS%mQ&3FIIP#FTv*Kc3
zmwX9=L*G*PPE_NZQ1cZ)KcSv^3m^A?|My?4V=RX@l)b#DgiD0>wWMnrgMfY^w$7Ix
z)>)WO4f2XDS2Lv4|1naQdD7%oqw=AZtpDX{;KYY~*8+kP@zYKY@=J9-vn9TE+28SG
zWU*$d9klly2PzZfTZbzQTCTn&K)LqCGK+pLI~Pi$_5{$MOzopD@QbXepSo%(4Xz4E
z(HxVXdLYr<*&Y7}0{kvps+hV31d25T=!1hE6GMn&78O7FOAcC_DTDq%pq0y&ru6AF
zJwc=$re-P~Z2>PQ<oov@8d-@F{Z8p@RdgpB8Q~i#Q&`b*->G!=;uWE&sM)kFT|2&)
zhV|`EJ~bLqR__wz;^aV37y$5ha!^@oNNp`zxt_j$`%3+~b&!(c<Wc~(76A6}A))E-
zW4YlNUB#J3?G2>*%Wl!{ydQ7-6f~=jo8@*84_m^yz&SBjl=Eut2wJwjrOvBmdu6ya
z<^DB1KE~wY>1v7lIlfrv-wS-i^O66^r$~fRO4*|M4FdWA%GL9c%<09jzkL0cGNso&
z-#~$iY{745n^xy3O2u;sZTa;13(t8Y&qYuU0d4gO_A_gEPddmHJ6rZ06UQ;6b8~g^
z?T1cMud%ZP7eC6iu3t@^7nfHoWmzxg)fP~A#pSWQbLS4vgTD(4cIi0EmoIM+&<6-E
z-4jkuPNLe27cVGtW|b03YDd1TQl$!|Pp`Uv<mp&Xt!4Ovv`?QtiGE>S5>VwpY;7ao
zd)>OJ=qtyL9V->rSdJVyQqcFbZQGV`e7kq=7K#jCUtj9drHe+{gtlR+`#{{631IO2
z*|<=J#Y8;g!5m-+&<E>)8jDFAj>WJ6V+Qup)h7nA81^#Ome1i>%nwtr81`nCO9U1Z
zVW26`m#^3Kv?CaKR>`TN@>zcNu?Uj8j+GrPnQAD$i>2_L_&!e2Jw64{kM77R^)yBx
zHd^RQ^f)Ql@?8=e)~*^7a5hU4j+slFdY5i&)sLUCB7ydx%PHcl`uM;L1$@xi>>_`>
zq~jo!u*Cn!8r|gK{d>jQU?>2f<wkz+YZc%s^#S?Wxh?SI-+%Ur`j>O0E+ti9DR}D2
z%(}!bZO^G4gZ!~cfQ?BnadY~GkMg)^q5n3V9w(($3iaWqvx%^ZS0~C=k|&R_@{v9o
zWZ61bbGW+_e#VLf+6O)ircy;!F1>KSef{Pw!9ze>E?&JZtl;u{c~bqF)x;0Hwi9Oj
zLw@W%;O&`<E*;)U8@BCY*W@=;s#qZ!H+2s8$wl21KAL7NSW49^l_zVC4G@<F*R&;o
zA`4*0dWR7_^P5}cV)xM>l$aGnu==&6Aa;je*hg~ZXO@+Y$+ac-^)uVD#5u7<8~Tn}
zIj{Og#&EhN&WY+hWZC;jU$0~0Ji{&g8G8-pjD_@k`o_VWbW5Cb>>ux8OSNUr@t9lp
z&Aa!ai4;tSCyA}Ke!v`MpqQ7$*!f6y7e~U+Sh2H7D+1wuU>YAQ=JAI>X#fKQco4`1
z^OQrUFHsIxC$i6wj#|~LN-gVFp*&Sv(7z1oOMw2&r5k9#q`6eNRAHJturqV&$6b#=
z`SD|2Jcq)KKD6^p*|t<WrtCAem3_l>k&u)rBMe{CrBeZP`udtYdBpdjvG)Y5@bNQN
z0RAHoE+}@a*mD*Fr9tAvi3uJAa(VlfgLX2vzJr5<nnLI1&6|Rn2iE!k?>BAQL|$HA
z1de@3y?XTo*FIteBy>REQul#)PsmS9JY82~F|v8TgoEfi<YY|!Damg}zD(U1^T5wo
zv9(eu2Ohoz7Ne>qVKLd6K1W+%xCu*qeOQckdF^e)0gy`_uk<u{$34bqRBQ_Jy!g0|
z(`fATFFqcBOdpd1=*RR3Cy@4IqlG>qmP5`xd@BOrV*LsSS+TSrDD8v;K{}fS*Zr;z
zUK)9z!$GjdTg=%n$r<M7hh$G5G4Ye~a)f_OgX153(vUOEA&0gTb98iYR$t{nEqD<@
zTa4woptmC6E_ksa@|P$0^P#LT#zjn2{)vBJ#B2!)i@0i`?_jS|u$(-1K?K0X(4W6_
z&EjWZ4hNqdI&O;2cC)!U)9JncYUH^&JF$CkQo43Cg0|`S>2sotE<k_XmhH4|%MJ><
zdP?1nq!hph_Vl@n>C4xz)IZ$v9<)7|uHT>*jq9n)pT2OJf<qqDoWDn@)8guir7ZMg
zikd|}uJVbrJl)*&M{mARmHF3cV@nSz!qzvjK@b<E!}@!4wx@Trxb|VIV~OXLa$YTS
zHTa$gp`Ev0YBZKFO9~pr6mzDCrOvDBJ%rYNo>=n!UCZ%@7Mu>H>m0)(w#djmf%Jku
zs%O@A)94c72$c1`O1;b2Q+4(}jE*3v9nP7#5??Iz9Wtpbgil?%LAl(sieb5MEn*H}
z?l?*H|5>$#M$A|&ZE7stowLyKb-o&DxwEPIa{bN&Dx5c`$OAvWyHgehFf~7GfB*e=
z>dX{6HOrNtzXx;@WdX?9q)o;2Ja5JJ>TUZ4m_N36JL=i0fyTb#!-l0SiY?VnTpm+L
zM@P~96DLkk?%cV>I)chtT4v9lO@jvy*4b>fY}o|3F3<Gz^c0&{uU;ik<v@-eJxVAG
zP+u-dk|d%2{rk(41el*DO&U=bpxeurFGWyad7+FMGg45H?r&Xrx%j?rsrw+ISm@h_
zTi64aKmc6GWsYBHX<2+ageL3{(AlgDbKURm<gJkhgGdz?qf%TW^Gx7tQ3ibY+M>o{
znB%`dcnrkT0k73qjCOhXHmG)nY`;&Pm;%{{V*x-OzxgaY7c!>OPme2aBd)Kj@SXqS
z`eDS{{%g*BVr@5-jwvHx`UoJHBS7zY{E-h3I|+l|!*)I}y`d;-<Y$n*cQ{ACcG0v<
ziK8vBvagcUhE|*prQ%uBkVD#}f&vG9mH60tZ@EYIX_7MM99vLuP~Ozc2|Z%h1DagZ
zje=O4y#C5lYEwjIK~&evj&|L8LF4uXQd188yX<Tz(*%8^GHz*wdsiwB`a6$<%D_7T
zFR*`@^B=BHpdV5sc8!2XHdo5#nw6$8$Gsg>*(BwOch4b~2egY6@D{W;x{lt)I`Va4
zWtE(2SF1$xmaQV65=EIB$DaC)7%%z?j(<qIo&yL9i6OnaQmv|$XxcxEXyVMd)TnlK
zrrfzKKzzq6nMIugM^Dm_v6HA+q5K37gx*8PPzJknY{_3$w1q#1kDn4BBBhHJQKu#J
z6#x;ei&MnR5x|GVh2VFz<lGZ_&KAhv&W?`YT;QCTE6RB_cLY`SPX@f@_go^XkUhey
zDN|~d6_zP`u<Us?|N1bc@A0whH}==x#fq9skKWT_p5K&_&0Rv+JMpOhYua|>g$RT#
zQ5r^P94tk-6P<m>#yUA%oCtvWjHT--9nTArCP_?v#?P^MR{_comQmlxcVLr7WyB&+
zt5UiMrAd{7`c3pD-%-6q0Na&Y_fhAjwbjKs4xLO;NDS=QRDk(wcO0Nfr3#T#=8Ph6
z@7XK2s7up2qWsRor#OC3O3IkdhSu#ksBR-!vZPeKY;pCrvZ9QDhldAc&z@ZbjJ30~
z6M<;EckgcTfdME#V#ElY?b_I=06}@?pMU;IY15{qo;`cg!i5XT-`}5>En7zNzO`x7
zhM<rbI&`R-7hh@v3JP%UgEJpes#Gajzka<4;#;j+H9B_e7#%uvh&puWAXQP8$_N-W
z0`BRoglF#D6TvDhh7B5r?YM7LT~i7Hi|JF=NoSR`9EVU_R9K8HEkE~|io2v`Eanfv
z!4EHD7+ykpa{yZyhxYVwr1IHq1Qs)Vmp@IZ<xYW&#cbd>jctlX&2#_?=gW`YQr~~?
z(z;f8L{QvG2LlDuf_Cs-?S3lT$_N-U0+Fw9V`?d7lo8Mxf!GL0fHLBr`q`)#gX8Vi
z-WELaRXupZ&Q@oCuxoBqBH=Xh&v9LF8X}eBXNrug=0brS%(lfURbZox%nP@+MJ{1s
zk9Lmn2WpORYkklFEo<UIO;_Bca&s<=-IU9gmiqDw22wGH4RtS-g=QWLrnwB<L*qQM
zN>=h>Y8}YDhB;{Jp@-CFt)GZ0&w35&<KsX@Ws%)Q!C3GIV{SQsOn&BK!ND2P1#n)j
znm?1;bn8bAT6d<DDN+a^ec9TL>Hu2u@sY2id1+f(s?s9w7?d~onIf|$j3obnAOa6Q
z1i@_gY*<dPfP(;}&y^#aSTCGEFa1@wrU2~$+(TTQooE^J>BCy6Nxhm}{sfh;{g>eV
zM8MmnbEb$OyP`1vImR9Ud2(jgaN(Eo;@Z+u*EKAIq60nez5SBbTzpFB8N`MHv42?y
zDjg1r!Bc2HTj8E~@J4v@6y`W01IsxGi}h6(!szh**HqLg6|KA&COm>V7R^N8m`8W+
z$xwRt<vZ2#Oef0YoHSqM$L*%2vkaJ@4ty*2HN27&<>vt7x@0K>>T51NrI!q<=i@R%
zD>zbSR%Aj9^;uV2rlkGLbYqv<tI1qrT$~G>6LUp5uaT9A(HT<&kndK~o*H@EMW?#D
zj%ClQ>DLE9>3OFf)6SbO=*c_W8!2d9)vRj%*v(g;(MJB=;CmNzI5pTJy<zw<cieDS
zqy_ND=25<Pb8&wEa#?8k`7k<g=aq=NGl1`JDEzNIe$R0;g2Z^O<a_-A&q<1BO-*B}
zI8j<1^n~x27I4(QAn~lgJfl($R>qdjBIvoI>a%)qqp|Yk`~T5g9uLSLiFpDelWS}(
zw}Sf)3w?71PghIa&v^ez^G9?0^egLw!e;s8q11lx-}G0{k(81Z`!y?+pcPy83XeUh
zZ_<22S{h2%E`zdpJ>11lsu$G~^nIJ=jHOx~2T{RVs`&`0a)xzjsjhnP^hL_$<}9{5
zH0JpR^X6A;(@&ibpnhoA7GnF{)mwr$ztw>8>U^xEXI}sK0O+f(NAu^ymTE_1kmJ|}
z=ICqJuBDbOTT<P+btzS<R8*%<9l<Z3ESU-f#34!7xwMoHQ`(l6^0HDHEWN$G#ZR7Q
zXdSdZ+qZA0s#U8Jto6}8)<=yRrLMMr|9-VsfjUvo%FLNF#j_BBf1xCRAeir%F=O=9
zNI1zDTlx8m?vaajqFXr*)e~4uLl5e?=?=ABql)QQHK#4DyA-B=pUa<4@^u`h_O?jV
zk?5H@CE;g^jIZv-!2Bm_zT$?cCa@R=+$C=OWA|TC_HdR2C;*l<%SBC>U8l0ME{gKy
z*x<5nS?4Hs37XNlJO&f?1yIT9=SBGxoHnywc5(_=9z@wl94~-5ktOb%I6e$9*8TtS
z<kkFHaEx_NX*jNofKTy4)XwcAm6Qdb{ndTP2j01dZ^(@;^^-Fn8zPItGPCxg>vU&G
zVFvhry*>-p_Ta46#n+}iemEY0ZyDL*P*)IM@YI1X--BO1+8@1W3w*eEaUt5fwT_Q<
z?fw0V?bA#pH2p?mI(}C5vW#vYo<D!iuBPr3=64~wP5hF2;Eg|d_5x*hb7dY(<{IZo
z`U&3lsnL*L`d<|Q|1as7Y^M(bAJRt-z$JnAvaCp}p<z!cfO+?At*t4mBLhGDF(j84
zFJDpc<52R-<snawckA}Ki_qmOR|qc4=17{9R&=UMPD-Q0+C*QwpxyGGa*=B0vY|Z8
zlMgT(maf<O<rmMzQnRmc5L}{W1v1bFj?*#q$Riq7(UDpg&PabBd`L5ng;4sGiK#XN
z_W;=e6eed-t)5pp3S+Qwx8F-z+uVanFc6r1+&KpMe^52{K+4Y`B*6NIufEW!?s@4w
zw_Rf96*c9~+>?)KG=qZ`SUFh)3Ihh^Q1@6@FX2I?ug|I2TS;3CO=ApAAQzl|Oe6LL
z(fuI>DLPao;L`M*e1$%8kVbPvIj{N2-1CYU4K8EOQk^{M=mTFzh+9$4Jq=%{ri1&I
zI<Hswo^MjK0_|&HiNBKo06+jqL_t*LYSQJg<m?j~vp<Lil+8l#zx+q5SefR*3jXCy
za`3(Tj;b%bLFL`k3Qq(m|6#q3dko63k~1%pKLhi)|GfC)7Igu(JF@jzb$$lyU~j$`
zAHLJg0R;r$-*|-|doTWvIu*~vRDAy<U%oF3Ii#f3%{<sh;E|eg2lp8EdynTBjq}+F
z!vU17mD`r4)p8|6ebzqs=-#hrr^dqT!b%QfA(tm_zfwMa&P}i760LKNwyG|BK=apb
zrve4cqZn#fzaCXCkd5kB^^tcm&aHp{{)0|kx=z`$I*=8El~x?D2ixVF^`x+K+XcY?
zD-R0&yLQK)GIPM)^tKURIlB4;s=o)pj|I(+qkRTpjjyX*d`#QC_b6>Xc9k}5`PZz&
zL!eghr(>B=1#@&*vxDCrT=n>28y4;0*Ozna|B5j}|3JHAZ*Q+RsIDpf`ts!qoj-qG
z%w1hvTuiANZ_7e4^7r3=6PD)4xN+m?$^Q8?u)LEw&BXe;0*eV^EQVtZN?6QU-mbmi
zD%~AcES6v~pBRhz&jc*yJ!3IRbz(75osfr&!yq6p7#TpDoKR%afj~-KswE8{K0Kiu
z@=ra^$X~`kb@$57lo5!z5r9^BJOf5a68%pEg@twfFs7}6HGOOV^Ks;k)+(NVV@3hT
z<H>-WJ^>)&t;Z4;b?j0m5kGzPmGoaf0sy=Rvqd|oc18~D&Dw0DUV}#oYkFAg#|Fvj
zq#4-9P@giUw<G-YiH&ty!cP(7Mj$#CJxAC|_cKpa8|1a4f^2<XlL5F12OhG5?;BZL
z{T~B;DvN!9{Sa8!1EdF_FUdLv?4cMf=$MjPuD(MpnA&Slc}E5tf6)Fr=Y^sXfIbAl
zIp@}M6P6g2nEIs1jLX6+sw#v0k^p?2!C+xk4{Ms=K;_i|uOk5dyLjKQ{?*rKZQxf@
z3`lLumvEgu$_j#*623hxVsWCJ*Kd&keN)ftL#AE<l^m2#ML9;prHAim2V1z`2>nPS
zD?6EV30U^LqBYax(7jiJItWU(q<k;nzT0>Em1rc474x{S7Bp}d*B<U6(9P_>`%2uG
zD3qN)od0c|OXMR~&gN^+=^wr)G53KhdG{^%DTHHMNYd9Q3j_Kpw{a`BLW80$fV~~$
za`4`3p$J4f5=T7@2c{f;MDW5f)Mt4+SchT&N=04Lvy?r15N6Sp72|1`BN|cw#&`!}
z*$!Z#KWy4UO2Sk-Bf7N`W7B8c9JZ!+`{iIhh6?@j&gpK{sKJxajXCz+oQx{26jDY&
z=LleqK5*bbLA3*Gbv!RXYXgh=Uk&D?!{}eWeED?tJvxa=lP0C&#TDAA=!~6GM>PV#
zF2?P9AZU;Kmv`oShM024pFCp~78BuJ6V-WBVKJt80?6BRur_{Q@_Nd;G6IH1Kmqg(
zjgFER=SN^PyQz=b6F~Ks++Zuz{|Uf*R$b#h0OI_?D+drKM_}66;naV`II2*m3Dep9
zMuqa_rR7XjqlhvBuxvHer4APGu&}Sc>=u=BNiC>0dN61P1s`N~T{qfr<(Z(Z0A<P9
zpmzk?9Z3*_17JP`aX27l+*A}Pa2ZX;pdNsINnmX+fnr(MD%gRohGEq&7g+7T{rp4l
z*ALkhpw5D^ivROZ1W?aVpVbw3r7T-vr%R#gmpmQoDY@@4DL5B~tc{J`w{l)p*P?v9
zOg*nj`J=HjTbVb`m!6zhQI6p7f3fxX!ZV>FUAo@o`_Q*;+4E}Z9ZS@<NJeV3+>Z)Q
zy+lxSmF4?waCwy?wK7xeWoFB;Nr!?d@YQF!7V<$HhTPZIo2~Pudl2!1Y^)NgX#wr{
zSpb0-gCxbARmwPQ7YkR+fge^b*B`6)bDXKU&atut@`&%_?>x@I+z&&2mbZiTUZw5n
z<;U;Z$zt0hgX&qMGc$%0Ek(H#!`=L;%}*ZGg$7QVOO;#qVHyHfF6Hr{#S;hq)IKOX
zQ$`>*jR5Fq`t<1|C}^N4DqOfQtz5Y>Hg!?yi!uVSY6QkqbET2HA5i_pSHrQGR5ZJO
z_E^<Nr8CM1#N`oC0R6ap_F_@rfBWBWI(g21br-=SzqyU4ptVWNA0yDLT3YTs{Yc<P
z%fuhqikLS7(A<M6!CV2Dfon1@N`btdbZF~Z`t<2D^UWuUL{k$J$ModsGYWh5oH^??
zX$)dwUkR~IT+<T3_Op*4F*snYyP7%lVXZ%SX8?_13*2)(y#-HXNw$iIlyprkc=ciB
zEEj<G+G5DV$yzHA_&25s0|&NTUNb$IB%n8uq%PB?Aa;nMK3g+DDHkXkB!z=KJ-)9S
z8YA$3CbaVk<q^C_B*7TqdIA6qZg^7weW{IKwFGtHnr1m@E7M~gVaxcnOkD-~pk;sM
zpq)&a({9~ezUGq&#aoABnP}jS2O3>Xsddk$;Aem#k>9fjKnwY*#Yz7r{GNsT{Ez=s
z_pCK53L4~1C+0t}{@2%M?R|s4KR(NEJlDirCO%01+}UaW^6B)6X~hz=^)IaAV^Y3;
z{YKaCKA>#Q4jfZQv1GKwSgWo_^DjtCwPT56`ju^flI!TvqdZ6dBw|8oQ%PE)Z+?D$
z!r~v)If}%t5fI;({ERnO&`x}LPee5qUiPVIOIxoP#$rD3TqKc##YA;7eo;Q4onQ4G
z{EH6xr|wn(^ndDb$KRd(_kaJHY&rTfM<Y7?$j0`s^M*TAfTOFA{nN$RDlt+HJh@}{
z1u_8eibCIhA!l~$>cRm6tLIJ^X@3K@-Dj}i8Tkz^qR!iL^(if8YsNdQpuh)yWey<R
zubgAFxTV#Nnz#Zh!~e7qMHAaVaag?KnvBbmj03;@ELpO81I?VjghFqrg7y5Y{TkX&
z-{E8F+4GkoSgyPrI70K5u0q55cBgtZs>-us?K-Y$OIUJ8=Kw)L@`&k35C{{$d(S>n
zg@3LIZfyj4dc_?3dpdYh?u=HV`9lmKnj^5Vmjp3F9y7OfHaksV1~fmAg8bY>pCF*&
znZo@2_H5B_sLx&bnI@MdXG3Yi!C-lkdfmn558Bn<Q@tIZRt$|%LOidy4i4OXO}^}L
zk~%!b#KYGrD{}-oi%T;7=p(*2Qqq)~uHw1z4|COz-T#mk6Q2a9Jv<N&c2Wgy0|ih2
z?dBcbeZFU*5P_Tycq^9ftT4b_KoaSm1r=AW@Mp#K&<{fCV3#5Z8R|1=e&o_%=}o#4
zZt*Wq$2z<$5ZE{_%YPg+T<$Rumybp2d$K=%j={>q&@d{|pcCy~I+aQk$Rq9kmn^3)
zT1)@#Kd!OwLhPY1X<%oKwAj9(x*j!^>e1=5rP_(kzGGt@c@SJn_pfo|##F3WF`7Ml
zwxyenR(nujG-%L(0s{j%Aa=yJ3Ap4pZQ7*KW|1OAX!`W&8fo!qLwsNIGu~Vgo|)#~
zmgaWjr`C^2ET+?XKgMEG(>M-v8yf`14?eRlP@l4njKf4w&z^elnx-5Iq}v=5(cR99
z%D7uo^Fo=_^VVBW-qY-(4+#Nzp`d8Nc`XZP))^ZyHXNt1eE_uctG<Jc?PFXnrvUnK
z`RpluS4KcK0)v^8Zp+nY)WF-0gA=Ev^;g2E6Du3?vX_GNv5w-~Z#<_J=bvc29H0c~
zvhgkzXLtK4wcO|df8hUpFo-gwN=ogEskn7zyHZw_5r~ly`1b7^ZP~e(&R)DivllL-
ziaw>(`+4~2G3`HcoHlIRNw@F$(`OFWmC!_3^q6?i2y{7WZxC~o|4vnyt|R2lS7CkY
z&OF<&o`(Q<&&U)!Yc7S+d8X6w%w$EFj7CQ`F<*BD=J<AC(0&|)kg&=q&$J<OK~V2v
z&Zz{|4=nQWL-5^g{A@yy&DWoP=<Tz-9jxc)xFNln!b6v&)$$h5RmPk&*gRUo;zT*G
zk!e;;J+C(Gxd9IpP&f@_B@q-;%g%+7w|xqeV%6Y!u<Ut7Yo<v6dlc+qOZ?GQoGBLv
zxesNfSZcn1Gw?l%-(%*#M;g3aM)JK4dH?0Rpcyi4PpZ<mPeJ1YPljbYHvpKAfNgTY
zJw0aMLsl#V(I6h*u=n4Drw{J$Ar%~%b6<a-<?Ud-k#~Bg%KECEEVlErBI8#H+e=#H
ze*RVcSN3b!5jcA83fZPjO;yVjRX1#(F|vV-I#XFuMj)Pz00RD=K7Cp!7Q({9A^~!T
z4<Dv<>C#d4>ebapo+V2b^}4d6j6gygfk$jea+EQZ?QAr8pAD$Ka11j^LO8F{(rY{~
zO+}-ahUgu8J`UgI&vD4!)7*w0Ohx#OnlHb>G>M6YV&VzU^Cs>O5V4e66;ZGlX&jYh
z<PlH+{m5gYq{NXC*z5O#)?5St^_Ib-KSbc$(&4V$w^;Kz|6~ZAU`;3<jD<N!9wNyj
zh+7~>81KSBchdqH>ECP51ZQ)_9M)oc=uUsS9rjW1%}%fFMz#z*%{}o*Shyk><F*^m
zgymIJ4p!HpxNZ<>DO(Dzx%h-$e)vZD*aCHEMJLLfMrF|e>(zy)LO8(POQx0hjf2aj
zquxFag8IVTVFxJRrYO7L^Kn87Fc0=#zq7RJ!V?-0p7xNrZu@P$C#W^faPUh>j<P1$
zk;iad%iL0u7`qkcLlr<@JtCH0`NE*|)PEMy=Iy)bF;A8Y=l7;r6Gl>|4C%$rR&U%&
zD>rPRD>rTlI-CZzs!@wZ^+aCdcHJmXE)Vh#2&8?7kCRt!4;nqN56xS;igxZlM7A7k
zcFf>D0u;Y^^*Z$%F`imAu1^bBtf9LP0?4OCF`6-MI9aDv@uAC)?(MsGqWq!brzja)
z<J72JURcD#wK%Fi%iA$^9U%Qw--V(Y_A8f5d<+YguOWc`2~~tekBJBG!~EN&nNFw2
z7Ju;r=RE@BmUmA}KiG^EeEIFy-xaNTFtFaeWM-OmJcLHEw-2a0pb#)5DmTiA5@@3X
zw+G4&@OO``>db(>l{&~KKr_gs>MqoF-5u(*@t)XD&d<Y9m7OU&*9V=5zCMe>vOl2u
zz%M!?^+j5m+6Gh}masTc&a0-v2j{`m^NK*Nuoz!=`Kbs3i#BpHAN!2ju2?lmEPGx}
zSB$k`4*gCW?o#c=H$}fQrB1>e`i=z8i*`jb^8Nab3QfBtlx6if*mW7#G{Q@zFvqXZ
z=03#z8Pz?Ed#}u#t72cUzMoc8<!y*NRjLcXe>#t4pRN8>Y2I~F)|ov8=5ugfh@n1<
z0<u3fJ#A%MmR*hRy^LNYbPoL;hfdMbO}psy<s0N|pOJ=hAl&lB3y2eR{f<A)T(*&p
zoVi5lcy3WVzbB3A*@pf|nuHc_*h%|OoMVf6FJ9k4NfIZf9xdzBH}<BRzGOYUegA<P
zR4YrPdbAO3)#)@?w0q*hHNwv+a|W8&zatg$$|>@-|IAynm6mPVP0wDvCRoUi@6(>L
zWKzwM4E0&N9aFdO1wNww?f;_YwJV!iRVg1wM_}8wZM1OVLOOZ!BxTK-l@R#0V#SK$
zXs%tmMpLIwrGp0#lARqJPL(W4<HwH|7V8TZETBDm_E724rD^{B`5Z_(iQw7Ce0B2V
z$@KQ^TY}a7xN+k|TYvufCsnFciH;pRCg^*dot;J7ux{5SzP`S+Xwf2i_Usv<%%n+^
z$iYG7-Ds%Ky4o@&ZOxiBbne_a_MDCYCFS++-n~nM2M?xJty;;GmGuNP0(<?Q(`v>U
z&Ii3A2c8r3FXt?*&M^c4s^>9=a>D<Wz*7VkQz5Ggi(!w$&TLiRtY9X_VxH3d=O3w3
z4qIwkFe44wexI?J50pD|N}5q8TR0XI%yD}@b8IZeVi+vPytx@$uXiY}TY<FnOo+f@
zUI2^9!dOh@tn2|SVKE<R!O2IA#k>$K1Yk1gQ`X5MSj@GMcWgBISf^3c3IEr^@M2Yq
z++se5G1__aIr%YW1zCRfF&i{}rz7192r5IQfi5z>dHmtH6tV0GB(iLCN^8mp{G1VZ
z!w)=IpV#DvaDS$S0WjZU)h)U<pb#Zb{5!Q><0l@>9oV}3!`J_4>fwh1^lx1x6Mf?C
z%a7jE)sTBs*UL`4SU@!cAik7qT58Ba^`H#+``|<RyM`;jWIhSS0W4aZ^NZ$U@H-m4
zH&D1oHOiMk<A4GFBM;VWb@JHJN`A>zn|GCtb<ZbWLIBedK(bF+2fD=V%sw8>mT-Si
zy}YWA6M18XG@WB~oL|`WC$`<#HXEZ!nxwJq#<ra_jcwbuZKqKi+qT~O|2)q-Yt6U$
zaAvK!&z!Tby?@)E%%(y?V_`=_hUKxE0_?3mm_njZ3fq3yRm)HNj-Gu%YZT0<(*)tG
z1mv)oun6Q(!w!m%xREU+c9oX=Zpn-xNjG=bZp=mq{gP7HMv%qd@{oS%F7cENm8L+H
zD8hg08YOOL?B77w$PZ~pfDVA)d{IbN30Zn;W~$Zd$hgm}9aYcEtIz=5WPw2sm+pEd
zC4Z@tpYlfPeYU8{>$J^we?A{Qo=78|a5z~4D1dX-E;5|RsLQ%Z(2S#()k00SKuGQ-
z;pa>Vs%wZo$E_XPqFGd^`#_=<u8VT<cM-FPclvZ%P1T}SRQ$J+pMQSuJU&U`emw}U
zO-;B-D!M=erSP=2sNQkw(=f}&=B)&i+>8~geHGUuV+uP2FI3Hov_`uTx}H3lfR~iI
zspIq>{XT^<tqp-7y=nw!;G$e{_xA7Y2k5cR50=4q--Xa)rzh;#-T({H*H#0{I3F1h
z3KV-q{Xn8SI5<PIxRi^x@UvbdJ=%6RBWRe<4xH6UM>Wwd_xUpX@>1zX+?+#5)yj4c
zt$cq6GHmO%!Aji<ad8xuAe@DkIBSOY5re?rD8PhzoQeC#iFdS@bq8?*hkE|85#Omf
z{;m*mT2uI{hPv^SE}P<C4FNxOywL{kNkgNZ3`0wEI16uVOG<!6QsHiV3H|ACRPSla
zWtqn&t4fU{-j}&9tg20qUnj13q9`$TWECc3rHV0VZ#o}Q?d%Gl+q=|XPCGIQ&e;_X
zshVh0523YDGIO2O-%Gc1_RU}W5*F(}d8!{_-c<|01w0-CshBmyOyRvEHpe2CEgt9C
zRXU7lPu7d`K&KVW$CBKxHNO^D@hVYqc|3@;$8gYWn6-Mcq&ZJS)m@pe?jm1aGI>o+
zC~IhM(d?gvya}i<k-cKF_D|oq%XP|Yd9uG}si5x+!FKw-r+?wJ{Dy?jRaj>{g5GPb
zkuBu&l2}<;Sqj9Xf=OCof4zRZdwu!pNCEcia;XX*sK1NlvfKCp^xouA$)?W&6@2u`
zr9kM49Uch@Ys+u7-c+1r9oP*QfX%HGFs^bK^n$>a0G?j<#b%E@9-FzZIza0Sd3iXO
zQ2@_9_qaU>7w&Aa+r$>`=oSR#3)iyLStQCAV2f$&X3?ea@L9mrQF`jvKH#LKrJ?yw
zLkR>|C@<oGirVnqKQyJ*gMZObH70q5%C{{>LW8CoLjs<$>Fs9kMZ^K8-p@a>tT~qo
zWC%VcXU5}sBz?HE#id<a3q5Z~gvP9HrA)>^3SHBgH#nIEReyXc{mcFgNg|jZF>GOv
zRXc)}8x1lf_JHt*IM)HeP=KqU^(s%#0l)7>xIvscX$R7uYiW6aY*cLN#Jq4#1f{ac
zseQr51<5cR&IHp_v^qkBUJ7>VVYT7K<50Z8uIgA!{+7*hU-vl3{^hWn-%E<viGifo
zyU~m53MmWTUA8J!2?ktih-|v-cxxTU`H2bNmn}QwV}M5m%@h(b${+j){|g}{^HOv#
zS~Jb(r0tK~6xvZ?XW1EN4nsc@0aHOuNLfe?$Xp?mAobXW%C8$!OlP1w%45yp0MBa_
z6xC%#fKO*7qRyqP3A?@>V``pE84!1}GyHXM#1@~4xIc4I0R(JR_QEqfUw)`Q7r6Nj
z5ae67Sud#=4U2CJ#@snsW1$p*PPPw?G9p2R77>NGvvMQN^nkkFvA-}XTD!?5uF4E)
z@(<BRUVwXZKV{g?C^DAoS{SwK{VfRkc0}NXDdqQDd>mx;>wEJCj_o_f%5Sx}RQ!N-
zm}muFe5;e$9cJlt4q7bc#GK|!0UB4_@m`mxaV+!qov(q=xvovp;Jsn@J|oGp-_mrx
z*3Zomo{E3mrZB@{zgHa0kl4-W>>cd<0LVXiTd^Z25V&(aFc2^2L$2^2vqwUX{Oe{2
zer>t(xt*E<e#PkpM44#?R8zTQ1<(n3z;Uk6GrZhITzl<!g=oHm9_eyfOg%Xs8-m>5
zjLf(F8{axt$j>cJT^Gb6Lv%Wa8+<w#7StJbAEVNlmlKLcXEEbM)EWBfvLjFH!D1^I
zX)l^tXSis^ow}5p?IQDNgAe0K5?prkH|Ov?!Be`L!l7(59{Zb4Snt1;YLs;8%-nx!
zcEkosn2lIID7{{biYqnQ;3D3=+v>yRAQCr+Nd!SYQm#zt-&S=A%1Ygs%9l-Pn(TCf
zBp$70E(N^{vn1sf?*D#(&q{hg{?%nQElurEVb;#%r{l^4W`+t?zbT6jU~^xBJ?P8S
zLgE`O*X-l;ikuB^D(5qs*;)kL-_a`7)pMG_e56=&ZFQRC87m-k0UvL&C?@XBH~%6&
zWPrE`bz5J*LuIq!m010<UIKV*=4mRd2V>AyI^39j-k#$A3$6f>Oq{KcT#+D1Tof%l
z92^|Ox*VW))1eo%-(tN=2J*3hL?+-7{_41gdY^}Q^slV}bhBBvd<uYuj)C4vVExa5
zPzdvYmv>M=q2Z_O+y_8xGvDeBP;&A{@y2=$W-}ZJMIs~2Z37@|3MBp``uhK^ng%*I
za+dPw7U-~aHa(U_fQ8K^7V@24s?sj|k;W#P79gnC?VVOEn^D}79gpTGu{GM2f;b5g
zF(w?c!ct3x&lbIjBmC<eq5z&6dJI@HwfA&xzfm^39$KIM?8Hc%=Q9r+!?5Wfob=52
zv+9)UZ!Z=q&`d0(#$-HI8zlCvpEzDb+GARBwnrt{5ky3qFVC+!Ty3pcjgVtd1g+pX
z|GAOXQ@ORXb2KZd{el>pe~t_<m!J`WxHiR*zbkLEZre1g8fq%p(inzQWdERb$Y+#e
z4Bybw%V6%io#h+VFR=I1n9YR%KIS2QY@S7><~990b-8FObF1;{#CkYuNK9d&fUK)0
z$5yPT?MRf_7a7}HRi7_(18e_?iU@<P(4!q(yQ~`nbT9j`Pz0YE5DO99oL516D9i3U
z<d`A`6Tu7jxv(Hch1>7o6Wn8tNc$0N%Wp=$;iFL4j-bgL_Lei$gN77H=$qU%slGnh
zFKliMfsX-HqhYknbPZl<W(T9m-%Qu5;P@BXDJY$ePcR^dL%93-A6WeFH<G!s`-je0
z7Ook1?o&T(PAynQ8O<YS-5|#^HP(~TS(;+WSh>O1|27qlBc&@Zc~h2JdJvcERC7cT
zugRaGVjpX+p!*FDTI>-&#WLwB&Jm1OLBa3eq7y@u8kkv$9Y4Xa5XT1y36s>mr4;_S
zhCMxDqf|Nl025Ja*3fHr6=Ogl;*+AD(L+5A=!=2vGlwAh?}4}C7{nKQbuKt$*0TWO
z(=(ZLMJdxiC|TvALD>cI_kNlDprQp{N#U<GykV5n)KrGd1cPY?yT;<^Pp~Sqy1{J4
z4SC~LuL^DX&t;<FfgHJF%Gr_{5pkL~k0e_6mx5Z2zA)<uh*Y4L-1RsL7ioI11LMC-
zTKlwGd9CAK^hK(U($f-Pc6ci(kS?q#p*(E+(HPpV|7Cl|u_yc`^X?ATVLZp3er@dj
zEW7nRl)Bfw7!bb5$rG%gzmgt`4{?viOj|=RRVu2D$H6+4N}#~rfM7gTDXK(HgS6Ky
zcozm>SQz(~>k0x)addQW4z7>}U^%qPRA8;@r3wK+hUAA>7_zGNWQ|tSZ(zbt#jq#B
z3DKSjbi$U!aLcB1fKN(-V}ej;PlT0<e@KlbP~$A}kp+C_Ry(5rWED88Pp?N6%wVN7
z7!)E@ewZyvkUU6A*w5o|ns81CQ0~W#HH)ONq;mx8xp{6H`}_*UDOp4V-9$3DY=YUq
zw*-f+fqtaL9zVbPR`Ea4@x*SDx%}F_2n;{3Rr~C>23Ev@7&V&4k9UH`?Rg9l5eNcE
z>|+DSSQ%er69hC3M0zm)!|$MMVZ)wyWQ`v@e?jw&b*V<03$PWwW&Ft^0A4_aPpnz;
zpFADIK|!yPfdU*Nj6||vB(<tx?LYQ^AeD_|5r$A&sRDZ1cnYITIv94&&Z4FtbN?be
zZe!{lx8UeS1=*t;+H8o3s2_?=rr*<VYi2nftO2d%DB88Z6j1HpcKJs-3WAexdo5b^
zNl}Tw`=eL!g%3S(P%m^TOY|{EbmIGbuTIF@o16WKwV2ej{@(DLtL%d#T`gaxM+H9q
zKyR&<QA*)lf_;I-Q{}7xT|KggeRxB!m=XC5b~f6^(YXCvNg@<dhJ3Yd8kct`cV<ir
z;Z3Pk53)i?Z6VGE0k-{LRA?$FM959pdIix#?Cj#PQ3~KBJp?ed1gb;gHJmlHK1$<X
zhyomoA{H)OCd}7+{#sn-8UIHfC|Bx6w9##g1yZ15|M`gsE~-`Xzl)swDMNBO7Wz@p
zr=n%MPrG`u5xb%#`6JQb9s2o3T)czikJtM_l*T&{uMB!BKgGRI=dJ`o*lMDNxivFN
z`149ci)zx!$#rg!7Y-n5df#S=l=3}H<p2#)VHwwOT?f)51n_|I7Twhb+G5BViP=h#
ztA9%+8c}8MBi}m4Jt(&wus5ga4@iBj`$Av)X5Yj&IaPy&D(&vLr}}^BYv(>8{G1X0
zg@aG)-=fxy)bf<Wh)W!<h3dp(QxlrAp0m=~v-cy$$$lSEtKI>pPpjuJHI2i4+`GNY
zk8^?bUCyJt5p|ten;~bb=QBn?bw98zpt<lzo*!}s6Tw}|iTD2VaLU!!@Wh9gsuM#0
z^T!{z<WjpQ`?M+AYQU8UvLOk9ALfc(E!V`~Er23dtIroDtNS~ELZJ2U=3u!thwCL2
zG3d)fvGUPaemq@{0d!_KkY}?s-6f{~`_`Z3gCzMRIyT18YJmSlF{{58$|-21GPeJx
z0yYSD_U-$8qlNN#R)$9_(%tHk+Cn~#+S8M!bZ)RPBAAqYOs@Xx<0Ul>e`8rtq2oB)
zM<i(H#oFeusNZNs%MaCQGiKw~fBK%0fh7!Gc_uuFafk=87Pn0#odbSNBk~(L&HhA+
zfkEs9&affysBl}qGr^g=P?Sw9a;~zMZYaZkM*tBMyll<qc<kkEoh%psb7*KVVxYR|
zkhNS`q<p%+J_#NHH~)PVBzZeofO5RCfb!NArxA(Y>kYfZoNP!rR!nbLd?8Jn(<`xc
zI8%$U0-_JYXD0{A+(^D%|1zk77{?HT6;guQoFD@+=&{tYY%a{noO}8m*$x)c@CpPr
zefU*k`B(7msr2HH7elqm+(AQq9i@1I>Vw(GY)+LVNj6ye!pw(5=kE%+2L;-69+K5^
ze{rA?_JgyEre+b3W&)F)0W8{a^}JO*Vkyt2bqTlp{rXx;q>$cYX{`MiAZ9F(`jt|x
zvAlnjdNzEmpFI@y$>490xbHM<WEQXX>={F<pyChb;wPrufC8%dE*LNlrww#R2gkw*
zNmsay>Dk(n#j;D(242_0ZlSMVOQ$dAl<91RV$N)xZ{>oqg4ms|#M+y!MDz%paU7pt
zWVL!sVN7dfBtu`$lK*HXrhKTB{!L`k>81kg%&)FC@(^Cm^WT-Ze9vK>4?s;&1WJx1
ze=B<$$!qs4faz*IONk!@@=JJc_sw~B5qY*E`6J|fM9ii+6}?0hmbVJY)n&Cd^H2ww
zeljLUWRB0)wjkkCa@hh=oRHJ&%pOUu%S|c6d87>=S8{~RA=ZY8%#1*uRgBOt;Getq
zWwi4UG3t`jRh6ifoJ1)uijjY3-@|9{@`~=ZW??v)2-5yTeY`wE-eNmV!EwvW^(P_s
zn!|0#Fqg0|?9TiD*|-75(&&&vc4y{8d$$!d@vWFEv;vKG7nYgc8f8?2mAY#4fF!8T
zG>AwjFHS%!V$<f6gifbKfN|TuWh~)Sgn6^GNE!byULE<5lk24emIB5)Sm*6lUc066
zl`*`>O(Z~XT(r<|*y-(z1ya@GIkm`Q!K~dAqX9uXUrwAhU859r7(Xl5M9F_LsM$@Z
z|NCEy_zi{Vv%q9lkM%sNbb_?YIduI9h{DNp>Qt>DzaGu(zZHRP1PKuf0k824Zf4Rg
z#?~wSPIr+MCS7z3p3(IVw+xrct=<&K<fp#AzC?C2Ia(<Ov{W2$c#=Cf@GZZ=Uub(r
zN8dwV7)mZgx>g&^B0$X=%h)vZsBcS{B3^uM7npU>TUj75B$UAoL8E9Y^hBX+M@Rs@
zxH9YyKfME^(EfH7(Rkurn)?QY-O$p&ZOlrSE_l_SL0-F*V$52S<|nA#+E=((UC=bX
z1>*oV#|-Fg$WRgntUrtw(vXNV9EJ4^SY?xS&9H-FzZ1Mfq%HXMe~dawHy#^hWV*Vs
zP5rW%_*2q&|DxQ!*aMr6YBWG17+fOpHgyZL?(-7g+wA`&6)S07L`;krG1y1*A7T(V
zKn{nhS@2LvZgIQ*3E}jhE6xe1P=WPZ@!c;W4J`5=P>zEVAz^0|Za6SX&27D&OcXU+
zbTEE_qsv0;>}j6tL%G5idGur(iqX01WgI4s%6%E@^Zzdu9H>k@w_?RQB6I_UyvZTs
zj`esEwsEt`U)qrcwLs@xX`A{fz2bV>S67lcU}v9lD=^my<6A`#<-~0aMwI<f$m77)
zv@_A?y+9#D0b=a5rwBM`jSq`zEMs8g2Y)1noX2h+%;B=-*`GQWA$=3cqH>`dAGEE*
zV_yNRZX<WU76e&p%r>`o2RAmmy_o;?6f7i(6nqvlq_NKle1g(joUuGioXFM9sH2L@
z)#)T5BxF&4ht!F^%qP^E8R3g;52KvQhrU$Kq1rWC>co^M(iZ^Jp$JSC8?wTNILKt!
zl*LcVT;q*Jwi4JyFV1hKX;e1kz1+R-kVv|dm_JQdFeWh5d^v)_-X%FI^D*SF66PDp
zV@eP5@tq4JQb1Q=JA5%pw4I1e5l*CSllXKvssP4H_3+vA2*$GEfeEIgCGWUN$6_-}
zebMONM(gin(C?cLM!L4=y@FJPJK6;C0vhmv-OP{l+jRSj6&?P|)j7E#9|U3j?x<EV
z$#?yQW_;#g7{mjHc---X@gr>smnnkN<dk;XcFC)kaOo@OronfDKfXrBf|08KnJFw*
zPHQCE&<TznXno!vq8wR9qbc|L21XLcrqv+p1bmL6Jr14Ky{^+de^|K^<dzGHkEMyc
z<|*U+@=wowkCXA#wxUxVjZm{$muzb=Ey}c==5Y9+TWfuV<t)`o)~kc+>HeSV_IBr2
z3;V-^|C}9EvL+Ljd>M_6pMxGJoUYjaN&cv~ZXatlDzgd377qMi;yaswu?HdA!+)X4
zY8gAg7AzaWWCVTZ@=_kqiLgbi<+J&vfO04d=fpX@s3dyz-?<&2a}iw#1h?6bQ7Fu>
zf`Wp9CnF0B${PC+D=?3~yTeg%*l#Z{40bT5;t|+bzd`5D)_+%0KeEis%p|=Y&hh|}
zA6~`>a`^uK9t+HiQarfYht}fS3=&X;0xXmpV2Hr``e)HK5QSZ38K}x?cVLE)%tP}!
zW1?Xhex?XDG5(atXAYY~I#|9O!W41@4z0@9l|66YLWZt7O`!4Bz`xfdQFrwk<=XpG
z`aRA^_bVtHu<{yVgGGpJ>FCBMvtlSiy}jSAkQ-oY?=2fWKr38Kn!Q17lCmrX2O82F
zOF*=&@21A29q$D#?*MbTd=hq-N(?QJg__}DF46Jp;c+z3r?*A(+F9Lm_7iC-EHHyI
zV`?O}_JQkudpK2DLa9}|)T2X~;vH6;NlPw?!)|3HBr!I%oP)O9f`+OAfwrZ1?BN+E
z+LKiT4EFD-{x}yjTfM<9-Jn7iK~2fILO+sr`4D`wU;9&O;_uF$y4kJqvySOADA%_b
z@<9w0tCL?uC^$z46;=aBzNd4N68cdDZ_#B}*Q4J4Qwih~MOt?x(el7}T;sYvS|TAf
z$JG09E%w^Lso;1rm&eIz`8P@DHU*QoV+wzXRAB{?kUNdm9sV0QkuTuL$<NF~)~fTe
z&5oy&o9S$Gn(xQ+i`W6NQe0x(8IDzsEmj{2)2Vcaj=~)E*v=8u*_O`a`=7}Vto`|K
z3MKmsxM@rv$p(ShBK`aD-wj_;{^~0sYTz?({>*52yuqXIx#oS3XS277q$Yi^XJ(A#
zo6NblwX+6Je_Zg!5U1VLtmh=DG4YwCUqtdPjWZ>CUWSIsgN?s2umH5x^tb+w>!0-@
zn<*XzZ{~Y9fpH<;2#kfI0v0*IWU2b2<(+_BORdg{tY=%frE}y+%GKPhBKNhXJWkjf
zkEXI@(HRTwS4=}f&?>VK@lVWLlr<ck1=;=@c?)k@U9E@CFYU(R2ie{~Iy`LaLX)ts
zn1TuRPEn<)Ha%^32>Fwi9hi3rkx+Uc#M#B+d3`=tsZZ+m1!b#_$q?~pk%w=sXZg~#
zHwiQDE^ghvG%^2@<98SHtG!n*B3N3?h(bU=hJ~i4)3}=9{J7HVgGOo>+(X(I{{!Vz
zgXi(;-mbOdo7{q@zC_Pv<_i!{O3!n{cVKukUQ~TxMa=zb!Lc$QWj&+N!yxRt{9b#^
z=L0#x6=0f>{RzbapF=t$&A4zZ*Q^>9utEGeT!r-ejx*99BBc3v$L|+>_1v5%tbytG
zc;xd{E7}G(z0=_TW~TeB?}6zYGvITF6EZH6E1GwHUJ^hnbz`I_CdppaHQx>f8|KPO
zj4%egkZuac!2B#Qo!gG7@&+G}wP8gOeBJFgmMhfd2~+`AN^lb&jTxvkNDBrYOf1mb
z1Fvx|R|G<&0b+{?`!aVf;QO-B(l)7oZYvUgc{heq;dfH^6Q})gN*d?8dpwu@o%3_q
zCvnNbkMA;#iaowQ+jaiJ-3oHTot>TLz}{n|`;74^3CeuVoj>kAj(!4*C>!En+izG6
zHfAj=TTj<EA_4Y}&{$A|<oQF1#Ix$H-MM0XkF<)$k8f}WtX?bh2Bg~46}73(i%p5+
zPl~n|9Bs*bO`1Llyw$fDNdZHVKhp_0v6YwN01p6WB{I#{)Fw0GZ>ukXB!rFp?XCB5
zW|gCtY2KPjiiFInTLzFQp(KL4XP&EYq157<XmdTC7_x%6G{EQdLxNsKuX6z#VyFa3
zT;2~#c|0&<NsvMQ_Hkz&BcA$&+Xs)v5=BXEKa9%{J?LRd0~mqT#Wm%<K4UlnlUTJ7
zc6glP#7Ze-JY5}Q$-`Id#T954W*l?*mX--Ix70PD6|OK*t2k}_FKtabjk%lf{jGmM
z2>09d*|g$posbl$n%m&&j7>YaKLgZf!azDpW<{Kk=Q@Y-M$wk*@}~4Mqqif@1U7MS
zhR>{^lTfD%D&ORH@&C`q;`bJ(BI^+3OTKmfjv#rwR`vHs*KY6IJzdMCf3AV!mO&f0
z+{H3aXE-@Dx^KLH&Zl;2&k7=(4_Pw(;Z*KmI0jsVpZ{Gv;?YLjRYIZdJzanO$MAVa
zB8LU~couoai`bw-PSH!U2GQR(5i|BQwnBB{lxgrf7PRF?ZzUXM-AUWW&of;pcg!*L
z>VY7c*sVS|xQmMm*H!|1o_~2^gctU$1EgUc82{D-A>~EE`vN=GyVU)d8`9P*f0`w%
zo19Gi)T-#HFmv-PulRH{G~37tnb%V~BJO-mm*4FyVePDgC3w-RysPLR{}4%HpJQ!K
zB?O2DrYmD^ydO~s7z{SKcPR^;YUBHRT*3<9qJQois6MpK(bUjanV9OHi!UE(%)IgT
zcqsWO|7No?g@~LI$hw7Jy2sW}NUuG5N!;cZ^nH){-C(3b_K2Bo2I;$5rFbNBf9?~-
zbbIpn#*URRg|K%rJ+r9I&zu}9G(aI@vF>&%;5j^f-1QG#Aw|jJnL_`FeFa?*O_K^o
zyMEg%IO-`~NN%9&5e`j30WV>wfHkw_?G3|nDg)Er`SWp=n#0-+WOKe%wjkR(p>vSH
z;5ez_9p;dZ8hoiA;b%LqtM=AHlPFDdcX;FWH1@ZKBT8QEC?S4R{i-@HgRuGr<{T^C
zQ?cZ$vIj2|(l_faGsN1)uW-_QZA$$=L(zIcOMMv!(UiTL{3cHsWY*zCF*)xfk!G7W
z7Gx;yFL<~Y(2Gr`5Q@-prk`i4dWGBib>2jGbC0=P(Y*x=f1P)tG!2onTOa~QBe0&$
zi4!@Md9&JULxq|Q1^IS-zSlfE8`%!`rDU$8pESrBZ%0NRFN905Le2q}QySA7lSDz;
zaeK$#|18sA=N`B@ik`kzb40d8zghsve2TtJOnSA*<^A*~uwmc5QM5Q1{`W{cPU&wN
zB?X9>F9i}&^K8s|o$jeqeZYLo@)pgX?Gtc4LLL!nY|3nc?SZXHl4*1RcN4{}`2msY
zu&^+5l;`3XhCk{%YaMQO{gOFUKT?H(lMFdAF>&CKue1Le+GUEj(2DfduJsJuJF6H@
zLw*#_skR^-E>Puyae8FL<u=ZIqkud+K%$WiAqU+aCuux!T-(W#+bZYn7?jEDaJ8~p
z<S(K-*-aGEXbRTR(4tKmI{o&|ZWR^$1-p%@*m&%2F_Lgs@4$`h$U><>Ptrnqfd76Z
z>}$>+3T~vT<d8tza#%qyHcF>C)^h)wZl7ys=NO;#KDb(em*P7O#_f?7;owdf?fH~a
zjbVpP$Nc=pf~(XIK>4I%lsxBjrkSNY^L*#JThPV38OH8MscY=8`i+w4w)zBS7>Pa{
zC2m@;1%qZ7sjebpmKD@H(%_<_AhDJ04BhXe+G&4O9NRhL9v*9dR^<Hn!>|grkYA?S
zwroJl7{44xs2c%`F?B{E*_{Uew4o1vjvp;GrWn&iIvfU@Na4F(+k=jQ#PGKGEmLUs
zXxzLw4Ywhjj)1zO)}0eFYgiM1JOXUwyEPhvW`*`V&X(dd&qJ(&-}X#86nCaYO4_kl
zhg~;j3D}-D#eNUJEc`&;q2dPQFB*7*B4!%#z}bD2=p!lv1R|ar=WoB6Zol+0hdj?8
z^O8*TXE6G2)CIG&D0~de6aGHO8ws?4+5GhC*!!c^quozUHetX%m85^tY1f4b|1Q7z
zfrb5|>p>2Fm*vL?B89B*R<PF7zx#W^EI0HE65-R(7zI`JcI8CBzd8(UaB%>|!u@^{
z+);kI{9@-czY|#5ryi5QeTDg7r|&zar+=2bhL;|Y{)9FBs5H%t%N6E^rY*=pp77Kj
zg_1xNVtJOm?&GDQ*51V@SN7pj8<}@*w23n9De@+-&phyLd=FLmE&nZAfZs+Wz!s5B
zLWoe>T6=V8NIYW9?l;QC003#hV>BwCk)Wqd4CoG7zR}NP`bL9)?K{?U-HiUkbLu4O
zcczdlvvMDLvuA-;7BOb4hGYDf{f8ivRlmY~hSK5yY!Qj}tt916ku~tUk)%x1WKW*u
z*0QoJGgXEAMh@5eg`-U7%$&o>rBqQC++k-h57XKZI{~Yk?rRMs;KfS2;5M=H5Z2B#
z8Cmxyf`~-Kv?tN%@82=L$%ApusS>M(mRaifnR;Rh)Ab_STMv1$r_hX@4BLW3uy&$c
zL?8v_{L7bjCwv7(`dfVPI+A!<rcqr!2Nb<p73m7Uk(@Q9iykLBWMaG^1p-4y{78$n
zo_|-)AI`3&c^@e4<^idT_6gF(?$qtw_0R*kRS!-dKLc+U0p(e$(7yf^l!l%6g_h68
z>C{gXTzYHiw>E13Kt12Qg-8*E`3@f>Br)pqFMN2=gl4T-P9m4zo=E*-t_A#+I|`Z7
zF$~CpKJN`zs#A!2WE8V1B){xJBk)3zlNg-}kYAETpV43YB;TqG7*fWP;b}Y6v*}Z+
zg3qafS65d<ns!T<L#n`OPz69*m&PsjH|3fV{81{0%#Vgt3GrP_lX0Ptpty=8#TMOY
zFec|-$5A668lZ?Njidab<&H58)txu<R*yWPDIW>}auuLqq+k7^c^}pyP;{aT=S!XW
z97rsOQ_9>WEyd;SZM)Gng31p^<D<nLftyIL8JYdB(?C(v3ckzhGkaR^3Qd)b1yMoE
zb&o#gaI*NHn`DnZ3b8=3wi`#Rs2))p4JY4e#0dq1=OkkN_u+4qjWGk*JvB${+Vi{0
zIZ1CMTXCQ57s#e&5kA7xplFj>IX`v+u?GF?In~lWefG-mZ@ot;rL25JQ&Xs9FOc+j
z%E@dm`{#Hmi4Q9P08evazNd20djKEv>$65RC!{8g!;NT`deoxwP3Sm8Xn5QLy&4@v
z@-TY*+uC<i{5pL?Yo2xX+A)eGR%caLJs0k1#HKJ=9^qqQ&7Opwf)I8J%qw1+zd?lU
zWK|{y6h>3Af|@$R!+u7;WK%i3^(nz2NnkUeKR-&Ir>PE=>4b2;OTn#}i{#L;Fjce1
zRy{^~N6z7i=`+DNZEO1mB*wnK1>*z`FMAcWQ8A4BJ#Ld{bvTM8PHiQp|5GZO4L`JR
z*`!79MPGkF!L)&RB}VpE{oRe;>Tyd&DV;cKG@9o1c(E$30l38q*v+P4H~gWbGq|ec
z_|4>WVKFHAuCKPm0ttluQ3LfsNh)yP(Q8$4*z47}!jq<!8y+=rZr}BPPpEy*K5Mul
z<12wGl0l!Tb74h9yL)Lkpm12N98=MaoN^OgG|Z*DE11@iIQh;{#}ZkIT-|k-h@Ayr
zY!&2Ow@OwZWY3ELg`1fh`SZu=!GsSjgt7YPWVEMWO>Eqo<~7Uw-yH8J^EW?lY-XWX
z=-yE9&YX#3#-7aO)I%}Ke+jdg0-b(|_+xs9V8pBqb@vh6omcAVPT?<(ekYz5EMhoy
z5aY_$*9dPGpWBpppVN(h@;Ymlf(_5g2Y&_tTa-L;6<$o&c=N>#eiDP$ub*<b7By51
zJ6^BZLVupx7)_^Bv)8z+egOEEg>aPfzX{Ff8<G{=HU$MG(@MuX;-%sRo{P1vE)h>c
ze%JB7#JuTp*<Hd8&xv49EjAk(y{Lo&9?8S98M1C~;O47mWsCJvJ`wH`5~M{JRx&Tq
zUOiy+Y|Fa7R9AyCwL3&=&$;3y#0cWseYVfA$$d|PDE|yo=^2&+IGGh$=(UR6snqBK
zf7f3|3(;BSqJjq4ld|0UH-8ECIqU}qq1+{i^V5r!pM3;`+12{@hx^+Hh|S{5i#n{%
zyw!S6n-+=cZ&#<8;1C{QWr(P+^zm_N*!%6kTRF;X<=P6o)70*FB+E@iXK;=mN!s<W
zplq<nTjT%B9t&3_!4UNKkLgLut4C((oK<KZhY>jVuroV-V?t~pFJk<9yJ8R#i*poo
z#^?2_d1Qfi%=b6I2d*y_1WCR2w{=PddJEQ1??6USsPaYJaJA=xVOP&|VM2+F*?dC4
zxJEDHUUE^_SD^#Uxa=F4ysOFI2|cpLUzb__Fe)Vr!sCfdNKF<BS>^S{`X1g2%S&2<
z5Y$t^XjXqA9xl%}v@s1E5*e7ZAhjbG-#17oHn9yIJ$VqPv-lPDKIr#HJ7z4G8G)sY
z(&o1j+w_no?O=RWs<o^5ujdAsXTps%KiG_MJOZg?Wq@)!w!WUV)?z1be@7n!0L3D!
z^}3{cy8*~)&r0wSU~sxWniD%^?o~kmPG+cb{wPHJ->)b^l6_lH`vTsG0Ng{u;C+DA
zwnDAyN1fvue^EC~V9bJa-%E5&+fyob7Gm+Kn6#F@=Jpo>?thq<?6X6WhGuP3DS8En
zXAPyAeH?dQ;!`gOvf0S)vRSKTmu(+IP2FIX$K$`&eJdlL>FjTdJZGog+HQ&ZdNGEx
z>>1wJTOy*MW>vXz-;7^%%UU1W>%;_?Oa%%PK5B+Kw0lR{To;DS@l{z+<1YK)UH@8W
zl)?Lyq|uFBlu}}cMhgV~I@Nt-J6_H3^bGvNXRZvxyxsbUZ$)&%)ZVW#%?g(sYQM#{
zI|{M=)z{2SH}V+l?*-gpinvA|723(=PZ>dt_fMY#G)*f*oVOQtLemlNzcoT;W(+ys
z1B*eQtFn&s-P+8h?{rc3A&bTH>RWJhX4jS*Y)Z!;lIp%NEtP1|p3F}=SY)v~>fUxr
zyAQf$^2^bSwra*}b&n+P|MK{<{xz)92=?fN%K9Bke4If(EjKg5E`98Nq{~h$+vew1
zlmvqlOTR(lIZjmYm}r!}uFd24L+e#@yt=d4bbJIBFHMvuNquqcN21<|S!Jc<!j`;1
z!t}W8@x~cfwh=)FxqWhKKgsdp4k`$YS-b_1*l@8N%Cu1zBb{|(z;k;0t4D7lH&)35
zTX%jp;<n+h_#t&)Jn+Nj-A5S35p-LXbk^~GNbmdf$oJ2g5Na}Jayz+5qpv0|Q6u<O
z*L&hqp?PF|nJdyKot6|UJbu#KL5BZ6ge`R`kN|u|aMpnDt{HG4?_KOPGA58SgkkA^
zTWcnQw5p{IBSz&awye{Kv--a3(RB9F`}lKhE)0|JKwjlg(hsAdC+jKFDF&xkowz3s
z@DpPD)ri!#rfhny;0DS0QN1?*n)x7+V{1enH-<n^N%Ap^T*dijtl1pofpJ00hZH)X
zyBNoX@U4&dGrb{Bq=-O&Uuw<XP!I|;Zqe#i*ERY~v<LqQiPddJV?n-%#BI=1L)TY>
z+%seRLi-`be$ngt#imd4x8LhSRx=^bwAw4T#YH)zN|SS(ebepTvy9~*Gt2<uw{8F2
z-VtgG--$?!{y(*%NgR&_=iG%#-k4vK$A^O<b9KT(K0*-E6WWvr;GcP2P85I^bBpQB
zub2b`{}~Xl#=U9oWzi0L0HYu#F79^-XvAE=_&lFTt6b`Nf8u<yNSoIihOX~OVf{OW
z{gC!v0M$EF6IE{Gk!YW-mWc{N*RaBEwmAB^G1|+p&yV$pkx>AHAy?eNb4Cfh9e4Av
zGQvYfMh-?9N^NDLY}ZmhRQ~Oo&F|u(N^eh(mu+D!w_`Y6A`Q{Z<U4?kpBr<mPutM>
zWoUZ=vLWN3*oQkuxg)6?i2tD~o(Qbq4S&J&QTdSNWs}Hj#@nZywG658565K+>IuN&
zW%zN(H8r%9Ol(ei`IfzvB_m9SdANHVRBnRYP|zO9C?2!nB`%}wX>Ph?crUv_Dl5T=
zV8!BEQx~$e()G9+@dBfEGgu;LO{gAKS8^_^-lp4&P-$^0r6=TS36{d@9!9KtNbEjC
z94kVPhl88F+ECXoE2q#qBb4W{8rHyP1$X{A(0N~1RJC7_%*`AyvFu$o;%~@NC@){G
zSrk}$6%Z4%Y`&(Zk?UrD&tA3B8l`5tA}Y9Sr_jycc6<okRHl1NWtcjn-5&oy{8cU^
zg?aWsBwVk{=<CT!d)jxgw4#b7WmxUbfwc3%?Vw<AJ?%cZ0M&nf2K&*3Yx>4h%Jn>q
znr+dp?w8yZvwjnjVe;7>-(`Nv<*Uhl`9;+2{`|YW;a?z;AX!Z2>iMG8h^f^v0*FL9
z*CZ@Kfu}m@XgS_?(c_(uj_W*LVULqB_BWl03|-}6Knt_FK>wJ{hcjGAzVCFnr1gEK
zN(BLA&zMVtQVGEI{wonh377b*z_JxTShDXj?Jk4rekfa=MtJ)O=eL&_M|E<?3aU-Z
zOY*3zF~UWY4S%p|Yu?T9eC^=K|2?xBQAyKX6U7B-m5tecCv~?hnNt5B2*4x7THoH$
zilW%R*a^<{Q!jylBaS<Fd}_YezwdJ@GIX`^iZgjSq5jr#^&L!uf#&X7K;{908r%3p
zlZsIGaH{^k(<Ah{mw4FPcbZ%8L?**;j64)lyb-}p#f3Rbj^^!@p?&&Fkx3<~XCPi4
zA#*-$y8ftTDyE#dmFlGJw#+}|jN~t>6pL=?P2zX4oTxZoqBr<XxyN+_2dBr#K4}T`
z0))NjnW>INuRZ6NAjTD|#Dz?q_p57+KNlRr8tpB1IuT13Ji--jXC?I9mx>!Ynq^W^
zzZZ)QHTLfVIsYDQjUhDpKA2ajWZZvYs@(UpP?9Oko4k?rqS^anMVCAiH2iqVfO)kB
z?+Wkz@sd=`tjpZmF^I59OZdj+_GAo>HFYZTDwo-yEU(%e#~1(*u#}gpb>Xnjnk+3Y
z{-O{I!k)Tnw7Z?H!~r=4@z1wglss!77Ez|z@UtIYujkU~aJmR6%9{g>dQF;u2N`qe
zCX3ev-GZh8;C;fw!R4V-O3&_%C4~I%?+x$@&jD!c(dFe7bgD%S$d0x)&Bo~K+uMQ!
zYWd&Y4`-{&cPzedPm8S%#=Q1h{wp-K$}OtR$YcS@6E$#8lW3e`NFfFxABpl9(^dz4
z?WvEATZ3Y}+mFAu-qena%74ODL>P*<%?>r*EJjxj(CO#B*{Tl87WOImoay;34S8eI
zKxIi6mz)q8B84_0Vj^01y+qzK*g+`;Hd*IIq`8SYk!kAV+(sj@PB)l-+)1>RI)EWj
zU`295+jhwG*$FB~tH3>VohSJJvK@l=15Vr;bYQA^c4dW&j#QMWyCt|(dNV)NnrPTL
zwUb3x`D^PXq&~x|WVWxWjl&zkF`5_BGeq$gid1+V_TFj+L=xj-<XyJBzY6T>&YXHr
z>c}Z<$*hdIKUa;Y-SBwc)8Dp_E?p3&KF(tkse~jUy_iU+8CUI7RbES+&<<V=Aoou*
z94#%Na0GI*sWp#@6?8#-OTS|#UEW+!H6GKq+s2|@-LXRQRoJ-=CY}tT%4uoMj)yXI
zz=fa+;zBox@tRj0V^>FR2fErSxNQH>-8~ZUi<U6!<l|vT%&tZVE>=tGm#a3BMG?;H
z)0}lY;7emoB*GW`LP7NWF1XCWQ$O#{tluT);Y-u`nck5K9ZewLRH2#=;OZ9gu`0?`
zqC)q)EQKh4PRjcz-DFp*G-JwUv;>FBr%E7Msuy5c>wa$feJbwXnBHX|VYr3k#4r)B
z%3k1c^0bo9=%aE`{ZlH*|M#^%g!Xb2uC=7<U4VsysBc6eG@uazv7$g83U6)q%ns@p
zdLRL`0|76pfXX)%+;3;MJj)yY96VdC-{l_nEUp13BP|hbMpb<|rSSEm(*k5=sByQe
zqWEp0Ia<#!oncJDq?eNWr$dbI#nGO40J9JROkuK7<0*HrJ{_fR&7ucV(wV<BUWgsc
z)`nN1T<$HIxvgGBEjT{$znLr5py!zP-BrA`z~@+{+j0+N%?JL8;abl=a+%;`0=d&B
z<^zcUekx&6FfAC=Z#)7Cq1^d<E+9WB9s+|y<M#sj+P9iXeCU=ZC_Kq7Cw3vq3;lQ3
zA4B)A49r#xcJMiDfeCiTCMVgfY9k7sh<e;Kw|rBlh7-l>3fyPe$<!Hq&rkySoRy6!
zm@i&2IRy=Z+>nveZ57;Q1G#W24c-o4?!6>=d*NS+$UpPphY*4l2VtOvNOQCwonO44
zhY82C7!$r{^ugbYC}S&B!BduMW{o$Ryd%>scfvfMX-R~yqQnZ)wtajb<oTptn!%7&
zF)vdTK4~#y5_GxNf;WH&jA|I+sS?$HLmsO25var@c4CW(^p=;G2d{n~Poi&ve+S$c
z2otw}$zlO`m&s;usdTv7d=c{PR9JDl_?JefT8=4!29QpC^6pY%cIS&^5~<~h;y7GR
zR8BT~dKCNz@_qUOp;S8E9sc73Q_2x|o@kZV?78h+tapG@!5H5@JW$c>&A~L<Mc1N%
z#?v`f(hv^2Nem7LZiIMal{SNL-?E&yMuPjw25C3_rUmka%3C9v@Z?^|HtI}x^CFH%
zr9CPY+Y`rxiF@8;lu@AIeWWGTh9{1}oLMWq;uii$UTWl3U8n3lzLEdYc27HPS2$EU
zJ_gsGa5I>AFzp0Wpj6j@kfH1$KU#Qd{}NOeLQRa$#ZXClWYpP#g|2$S)LN+wG@*-i
z!6T%RK-lxk;3<OfGL6mP-YHzJl&I%SwG2rq<BGYMdhBFPrb!iS*1dn-n`Ems+st~s
zs4X#3*-(=L+-~n?Lje>O8WH_StE*(SS25SscB&t&LZeG5?uK@W%vfFTH(9wlUo^gA
z?8}rWDrD~y&I%w371YdfFW*!aplG1IVHoos;z;IP#g<NM+)vC1AQ#NSIE%;84r6VG
zd0{UKW==Lmyq&#~3;>$-PR*7p7+yda0T*j~Wd~(0Z0#u_7Tj?4eB4j%;V_Q@f#Xt4
zK38i?offL9)yAyudMwRM9tZP<Xo3EdPJeZxwRS7!KNX5eOP>OgG`Rj>dkp_Ftu=k3
zy4Y-v?-&Mf+aD<Ut;c`O*Q+YW8ZG!{y4G(4deoDPztPX83UM->kzSTNOUTF8t*G7U
z{<a;+VX@?WEmEW;j-ZS;JkKo|8)p)tUm^~3bAW_08+lERsn`zt6YUP=v9S!Qq-eeO
zaX>8)T)uxc*&R>mL&^jFqZ>0tzfb0~FqVIF=_-_-^bC=Mb@!Vkc7ReR%_|g(c*+wb
zMq>N3Z>BSw8ukp54><O#<}gq~0ih)#q%TxGm`T$|PB%rbr&Ie}D4#rCk@x!5@c?Uv
z^lwk&iu*(YmCAE?N<T*(k~V>H!}#Lx7l!A6RXY^;?F3eTDe!MN--+ahm@Bu7#=?kZ
z8(;FIDa5=H*V%Kr-(^mj!+yz*sK5d8c5y;+uL_*+N~7BE6ge2I--FSF5nt5kB~M)m
z_wGly_522lr@Ug6E-KS#{4UPGi;HcGlo~zT4>YxVr194Fe1ctx1;ny4mK$1ewge3O
zq%9Si7^>2_bQTAwL^jfGEdGA%F180$5A_tjaW4u_GvFQHL>GnDpC!94jid}uYRYZM
zlqmH-wNo3NFO?GFM%=S(CN#$@d0U@{37i;rtrkjQXmxouxyp7s>`LSXGb`kRY*y<-
zPL^xrzXyr7C>BUG`o4RvwK-9P1#KwxrxWnpmi8D9Kr13;KmvJz+^Gz1<qp?#XAP0(
z%^H27EDqXuJ~!KT2eIW;$#HhqkR$;v93UDbj*D}3=b#0$G4?KoMRoy(BQgNQ466j`
zYNOt{?11n2qpR9+36t1tM)|~xBpXAHDl+t0c`iLd{%EBuMTN74p0NE^33rHNJV)!u
zh>e}<y?6W~@p9ygSV4rxG`yc~NOzmz3_m)BGpkoS_Tq_Ad7WL4sTpyI;*l7Rvim$<
z0Yl6k69dC$6!x!PMBLO~c;dz1@_fR+xpgk}7MoUvqj0;jnVuiLo;i375roU4=K3+B
znb^vSZo_^W7~9$EV}j@kST7H8R*k!sb_nn2mA%8$R~s2`sD>dQOvBVcO44gnP{Xg#
zuHbO=zC6hxqR^=5OXvn>=+_gsPw$USszTSp0nXOC8KVJ(Om*fz$p=RMC?fM!Z<A_7
zebl`Z+39s2AVnl{G&diJyB;z5^&yvF=F`>rD!eK*m`3&o_6IQDu{Ix^5x8#O{BeF7
zw4l?@$2}LcPbGEdz)illQY}C{UYP$CzA$>}i>iA!CrqJHfeHQbmeFzvyL8@zL{x@l
zF`3GGJlzGd(SZjN)7eTT&Pn;a%M{^N>flk?Y|5FupkRf4$0AKyzeLt6Zg?4haKCnR
z#f|m6+fv;RcNJ>GS9V#;`kx=Ue>kiGHL<Fd+J<scs3tt0AI!2)qxjnnHw!7V!9;^<
z9lxzy9*X$K5_0x-ya&iw+tyegB#uYTUxY_Va?rnaS(rDbe0a}km|d{e`6d0dG#$F6
z<2Zg0ZIjU8>~zFN@o1Mqf#HGCfYImqLuf9iwa%9f7OHqgo_3Vb(8I#V?Y_f!*}Y*c
z$(&Vtg7~Xf+BoKz<i^PKd@j=$;{ea32?gkuMGng)r5nGiUrf3^<S$0&cDlU-BmZOH
z2SDfdgxoTG*{PL_txlf-Tue~qNsJO@I-JN=MI6VLqv`tN4I}q|GTau6H*rr^o6)eC
zW5bB}JuPE?k_DIi?|5aL?W-s>VYI}5@tWJinG)dE3`g%u(*@QCVL6){05G#}XN2IS
z9SjyC4x5ws+&(@&?q2S#{tFQU@Bx{>GCZjvV%k3CnhnMFJQX?7?(q~$_^tw;cfSOE
zUcLci$^tgi@mU~CTng|CI4l=w8sL30iTwHf+fc&V6)yY$-?#JB`ue&8PJv0f(k6si
z<1BOB*D@ZwBcIz4g`>+(BC;(GN=i<a4Q>{qH=#pfxGuQbqXA=o*j7SZs>scfIj=2=
zmr49PUsB?Mf6B8%la}%HKiU&xc@=eNj}(^m_Su!txa8J+IfFwlZwmU)Fy`oO$jWoL
zlUNeV)2NG%F5x!@{S$Py&RUqWr}KslWX%%iFEti^I@47Es`G;NUx&(v=GidOO1+Il
zYn%H-3BM^(7<Y9i;i4Z#-|U{;Jywr##9JA5MN($6Bdc~8h#4|3q%7I4eh)F@^GHyk
z5%7Mg2Rgtl$TzCg;z6GaM$azWxNdm6Z!#=Gy}AQ%A&rO;bD6gyy-7h>mO(@67Fc3e
zJh#(T6k?MZP0UeBy3bhr1goONK8&i3vHmlw&1OQ(wRkT<LV$4zYha~wL7OKAs_{n}
zx_Fs<OZ1}gR?ZD+I%1Fev)1VKpHqd@z(-<H9MRsw$H^>qM3lx7icF+ou(!`=<&p4a
zd%Eqnefmb%TUWM{Srp&5U~jwiJoq9RAm^a*U_6lby@<Eb=~ARb>4@GmkC_X)e^P6Z
zI&&9E`RNgaso~BgH+7GDMV{ezU0CyP7|#{~o2U0sej3>sY$$L_?g`f4Nlz@8yl1=U
z%Na8iu?%dL!T-#&3Jy9gxsq*qgsi)IHg4zKGHpSINCB5A^N`fTPv3(vpl*2PYibkp
zdYuBsuf_%{Vg!!{qij4wCDUTKuJU6x(PZFO-WBRsq@!)~Vy=^Dj>e>4RT_6^rCtW5
z+vIb4W@nn{!t?A_-6FL<1mhj+hJ24jJ&s!*WXm3p@USEH_cK~LKS(^Mh9S|r%p!fm
zwHB*H0mb9H;yKp&_AV@>Svv}mP=T?TPpRYS=Fw@3UB0$h2~#1t8PnWgK(e-}<byk)
zLt!zCOYkx*{ZU-$uU9Ek0wg7>03VbhZb3j7Mx9@2v`DDk3RCq40;?f^IysY{?IjSN
zrD{11=vpl8fI=h>508*I{<Q%~TInp_rYkcjQ$ZTBP{jT<A_cMJ*;rDsQuDcdqXlv_
zG1p`U&3HPME#)2}qGXjAbkoEaBA$Yf?%5&tA|r@#>PqS{2^ia+Trlf6l)#U|N&Ulb
ztpulrxe)=C43q|AGDUTAzox9LG)iSueqn8d8kljk<027tsVphh#60UlFFh7DP?<s`
zQ|u_yO2U`(9+I8++>6rO<sGG9$O$2?dZb+7txU$IINjAMy%0(go0;6ps~$-v?-git
zq~oZ+BA0{F-{I8m{tbr#hx#Yy6CK3$!jF+!w-HIq{8f*CFqS`*23v?^Z`{)z{qyNw
z$Yw4QvF|F9@ciQsOBfeY25jy8a>+;}r@rq2<-3WExB8)X=foF%d~O`~j}Lmz=ht-H
zf%)V%yiMM_({N+0_3<MjO|RZ@Z~lL^6&3#1HEkE--vynGhxJaXj{X1J(}w<~b5$`$
z2#N;uSWih9it9I<u#U&d?cN$51QsQ7I#C6XqFB9*$mb}we25`&1?=8l0$=i*J?`sv
zfrLj6Wx`KGg>ntUF9I4));N;BlCs$%c<knD@h987W@GSQxNOV$rSPZ6T02S}LwTR)
zf2P8AnozrYIEMMAC~(N0Bw`M#`5hc>7f`EIXnOIsI|(E)D1FCtldZ1XczxL2Un4+~
z9K(UvhUzX<tBgA5z01zH4Bu4b`SWjD<4&V-#z?Zb!>$vvXDEb?qEM*fCw4OblCw$S
zqmO4_#7i0y=Tj0;ps<c59P`ZG#<9Dz{1fpC?~G?L#^KehK1e>(;J3AskzC?dUNFJk
zAr|!#^f@BzfTuAZ%62g4k5Ik-?T#9AJh56~2LvHb`*EE({hR!WPEmqrmKGkv74IS0
z@mLO~_sQs5Kjs|0r{PrBJyOC)KQXDDAqt+9kO5sQuF~$$SD|u_NBH}y*z6O}lk79z
zsbmk`rOl#-9-PG&5e?@9)+{52x01{5xX)UnLgFw=ACm-5cvxc}7ia{ZN+I8$^zVMn
z?g;dr8pkDOP(tPpLK!qN!;Q|q6J+S8F|jy{PJ0|d15s1SL-wWgXbh`D-fs-~X$w}F
z_{j`M;^cm(B=f-J>gR>ycq&&V_|J`UMh>IlO?3J!E96H+XYX~LEDkRTekYSWLBe#4
z41bZq5UUp0_p8xrMFR_VLVl8^0eEW_TGQjp(w)Fe0I(G~nZa!roFs_3k$b#Qf^g<8
zy!h*KaCZa}(MJsOiKwk!5D13j&%}#2#6eX&0Xg>HCmIPcxU8bqR7bZEzCwvdp|Get
zhtn3ArJ*-SMW)l4n!R_u^+ebYxxKx@1tp+8{@T6(WhkM~w}vZn|4ty&uHD%{+jdN1
zw}3Vv?UU0R_6cA&IvV`z_WvIMcR+~0C+-grx|jwGeGGz<AO3sxb}uLt(?~n~15yTI
zImPDRBe}93{te~2m`seF|McA#n${@P@_s#As2kM9u#&vojPr8#I^kUm+auxA>tsYi
z1bQPl7T)%2ezwlnoUlorKm!N(mDf{H_7B~0Pw0Cp=F}J#Jz{6M|FR0oEwSeN8VTK2
zq)DapJBoBB3Ni`R7)HUtsY8l535XqQX=~L^%|clZyynHHA5jHX3P{$O60{AVYTVdK
z+?tC|={!47$;;*O!YL;}^uU?iB;FLL;gLcIS_kE@X)tJqA5A$JNKFf-rhR;@Z0x@K
z)U%vsQ77jGXr8_mGzwFBx)hdvq@}pID_(Tpu_ayUEJ=B%_N7koa@ms65LS@M=^XnB
za8K?MCS6`e9tIvi6f6sy<qDUhKshgag_YXD)!dK#1axp}#tQw3Gw0L#tvhMa^a<MD
zx}{0QN4y^ql-f|v_a8n^<h_0Op8owe)M5dNxpQPESmF<#@E6sqS(S#4oucKFI+ZJx
zXAAcg>o${@M=oKNKYsdL(H`2?CqUW!*}u1bvWTZhp44EvsYGm20Q-N&g1?1>-{Ysx
z)8XT%sdHP6)e|i6hm4&}y?<>_70Q&<VO6wXK1%17hDJ`BPBX`kBy0{$oH<XFsaC0+
zPFY8fC}pAVNQ?jOrPB%LrI7vfqrt4ue@K5;%S`E6X$VDO{Uz7wNRRwlWuCD)`o=3&
zYy@0-_@4eg`<N>5DJzNKi1%2p0@QQUU4oS{6!`eKk1~#zDb<ars}ggf1P`UgE1y^2
zCm)64w_gR_=bt5xX0NgH`HpMvP<^iyG`&GK+IRaEwOf6Q*0%8=IJb=s^QzRGD8*+}
z`#hA6zn@?D<V(l3xA_Fj|7dj0%<P!$JN>yQK&*{M*EFxJ41WEEj`As)?blz36Eol^
z5l-)1c<_#v@<z^c#?+TiKO%p2qwIViyrlN4HP90JafwG{UZG^Mm7KvsA7g|?{**;)
zXycy4G=Fq|?XcZarl5yxogV!3xlq>sF?qhod-wih*!2fn)B~Po<jlXRZspQ6eEJeC
zpVY3&sa&!Av})Tv^2(Wo@~|R&;=)zj-VyC-0BvCo0NQ3q_qOsRZ8$~J#M-b@fgAVk
z)1T88Q;*h-s60FTl~ZA_T!fEAk9kvL;o9xgreO`OBVj4e(XV1Xtbk>`QkMRXEH5bZ
z;q(?b@laMvls0Wz0Z;Sr;Y0es1~h&9_7z!g-@Xm=oq)Ca%a<<&JPn|F1`i%Au@Q>%
ziWMu;^5x4ZZ{EDZDj$mdb?eq?PXlNRa{w0bQ1%<56e%<cKSO!O(y$4EvEg(9l=FZW
z!bcj9udgo+8Z?N0`|USEKjq}&;zG#p-o3kk96~+xtxK0K2A%oJ>qe!pQr{==afA(j
zYv%^AFfdkJbSjuC<Vdd7#iaS6aMtKz;?n91PYCOQyy?Q~V&E?jx)^K<)bMbn34E0k
zHVFVTGx@-M`k8exd)az@thYbuTd$)|A~s0qS1Gf+x=uL74~4l-q79MTtaKl@FMv9+
zBjD0mHAZHZvV;F3=&7b0ywAA#*;xkP2)-LZ6<L-UlE$OU58qP%t#@eMFZqOl>0~Yo
zW7OK7k#E3c={OCQQ%b*Mh_{+Xq$>2)yFd{Mh@ECdIGJYBkg>%kv9i6f4C<yDJ75Fs
zLuYmv3-aKj-o04?KkL{78nrioQu2>E9NmH{=S)ud`Qb8ye?aFPf5<<e$tX*Tcp?w~
zS{6=2zO0Ctc7&a<@dpIOe-AzDc%;LE{?}6J5LASuBk>rYK5gmfMMV>6c60UfEREl0
zXOH`Mqu@LLs>Y6KzLpX%w&q*WErDrLqFv4XX#0z~Zts+j0rtkrBPZSUzb_ouN-g}>
zr2{qZ)RVFmE-N4a+BR!M)hd=_>-i}R!Ba#>MzyL|q#_0Kvh&+H?09woRcB>)_AHqt
zR<wZOZkqnmjaqirDC!f$k5ALa{%H_$r(j-B!bd5Z&h22)I%48<>i4I{8VQQ~QG+xu
zL(gBlq+R<C>5N@!8*?~**g)#iZ!me5sv?qIT;kD`Q9}d-k5tYXu1ek4j=K1t%O|JW
zFx*DFqG`l*B<9V|27mi9G}<d+Rc~sJfO9?nl;*?+oP9bQDKRHX@UVy6;G7H4DXj{J
zwWc-K-csJ2NyZRKQ|o6FMa%I<%`SFSzvBE;q0l#1KeF>3bxxF!%SkpKz^RFC9X$z<
zf*{`~A85^`XEeBSMzg0cu0DQG|6F()#$W<`wqAQqml;~eS>=c4m8~S!ACABL5DIo!
zqTdVT4cMw>w8QSzt_ij3Gm3ImX(84ht?O5#s-=t4@EJ>}bm6=VOcXk}8tmj2itmXF
zR?*ZYYpGh9qLeLD=xNfqBl}YG-$n>LdNhzAK0m_;br)@G8~hO{;E{vyi70x`(1g2>
zoF+K^HAm7rgf%Bv%<+t<e9iqBI%nZ+S1CT>Z9gLFd3t&ZMf=^mcZK6xsm(rp`cTuR
zO(}Ec%)&w+&TDJbs3DZ~fZ%~_B!ZJ$z|z3^?Zk-_ML9UVm1xnTMS^#;W||Wq2@~==
z>5LhKw)LrC!GeU3QdGWtd2upe`0(M{4GTEvl}-VKgoMzJ9XoW!E^iz4;G|c=@5k6D
zPo69<66M26eGdd5CzVzsR=7`OWqV_WgOR8jLsbGV|JOA)M3T_OFnr9cqd_!kj~}@*
zL?*1XK~>lnVt!U`!{O^3*403norN9hO4PE58(rsh#<atM;)8T3{(F=SYecEh#e_a{
zBpym!rLVR8nfHPX6#$P@E06A}L`&MkoDZp<h5B!~!%)$hQw1*k-ZP?RRsp{Q9`l=d
z&;;Hvkn$7q`_xSK5deq7&P<S{{CCos$Kr^6R~7n>I4e%rC$ZBU0sZb>J_F$+Cs+$^
z>X4VPD;$gULFssh6ZA`QH;T&%z`vI)>;X};t4n@ibr_$2L}39cCtxp6te1^<dovQ0
z8R-~rR^lfwXOHlR8d(S*OKDrDd`63I`!#IdpsJbJS@0WjF{91LE_?9k(nfjPGXw9^
zO1w~O?nm2S%<VZ=iT9<$yQQG^TC7lhx_99yUA=LOp?5M;B3|%=0s<ch-2L+hcWBGP
zS?wwBE%YGx5x*?Qr{u{(Eu2v(3)eq)U_0IMV+%~iw)M!FjS@5VWkgU)*A%pR;Y{H;
z>pVO5bz`_4KnY1yu2gYq(YPLEbq`IG@^Ee5s4ms1UP<&F@f=y*DPe*H2s#xhbzi?@
zXai}1=_mqt%5XS1DqEbLJ4xjJW0MqqJ^+(rim+JS!%i<j^%h?<Rn#H#J~H;St>lRF
ziZS^!v=RKb^xb-op1uD@UhIgzOR2O5aZL4P;P?D9J2DQvCy)<_rdxbZI0Gp$uXql9
z{PNA=h8w{18i$?C7G&cFiL&x{7RT9x*!mvuNTyu9b9g0Nus0k{Y$p>;Y)x$2wk8wX
zwrv|b<|GrF6Wg{Xw%?s|&$;(|zQ6PAPW7s)A68edmbhlV+B;LHX1Obno3{VWYMCbL
zmEaKMnkSc2D9-mC556+anyl{KE@mGp+eic7n>vG?eSM`VqI&%vYl=G@)$Ba<1=w`W
zFD2}BQ)-e4ltlR2+~OLyLx%_tdmbsb=kRB)7JXi=o*zUV2WTNJ%LV)e_-h6Aq3Yu^
z_nNRtbQ!rB^VKq%%o*%sJ?4S2olt;sIcJ~_Yut>Jy0UlL)@^YhlfjC_pb-NF1%>#N
zU8z}*dbukoF`C9iqTA{usSP!9UT-=H8~F~=t|yzr)53-X+I_Cm?iPQ()<S}}*IPh{
zXi5@c_X9-jzR6}C)!Wt4ah44s*tbHhic*_xT&$gKS3FPaw?)uo^jCID^O>TcM7Zf_
zphJe{lcR3m!1#{B=MX?|;#l%p8TlkK1t|K+k33c&&Q`-B(OeS7s?W@~1M!51_Y!VR
zle}|O_!vna&Lsl3|D~`)VLQ9AEk`z6Pu6*EA}A@QX!hqTZVKg6lj{QSQ}8?6SZW3l
zqVBSJtqEP$JS!jB*2A~(+GT%v<WZE)ku}YD+u<^evd%)iWDg{lxhqkZE|p-PpOXUU
z=KO>5Oh4JU<C#lkDrw`;Tmqv}(C><&?f6iOV#^%1hq;IB1cq@^^*zity7@pr1VNyv
z?u(X(2$V7l^5NJmS5`)+ga=<yHTs5hG%0D4YyEoZ8P6q6$+waZ9&V>_0zApr$7$V*
zgSr^x6MMIkX*|`a${+7??=oRVp4{wNH#w>FGU{FgyHApOuLmdfZu9+KkV++dqJ`9R
zD1!65*_|x5p2t_6cK&{GfaLK8Fy1PYN|G|tpq;+YNftyT(If@tp~DEJqCK9;AX7jJ
zyL04xSC%9a1kK=)ELCD@Z^yUSW5OJxL%(#}2HQA}GU{pQd-!s8matffb`Ro%Gn^{t
zdV^514briDR5aJV<H|t(anFToSO|$|v_H8r#=juXupBDtM5bW~uGXvO0Q5qRM*Y~=
z`wOZkl4*}2CsL0uA=lkBrYW^4$sr>l6xQ?iZz?I?ApLK3?{Pb1aPH|pUC&QsVGVMH
zgLxb?XiX_+A8bIf+ufo&y+3RySMz83Jv96`i2#orNp_V-@aMa)PdBG?cCN0A;iDv|
z@0~oduz~8;IxR(|JBt`JY6|XZ6GM?DnzT~(1`mO?$9}S~-o0W$U0wVbR7xWJ(iAp-
zV9xJO7mpvWcIE2G7jdGJQO2@SQw9TKldZXhuVmoA&{8C}rV&c{y1TnCRkvc_-y~*5
zb2%2zR4TLo(kl+~RH7a%ftQ&EK-p4-d1D{QWBK%PP+s0-l;tYQ#~dX-M1Kw%nU{Ma
zW|mA}aBO%i8GbP_9#R?BqGJ5rYzXuE(Dl)<F^rI^piJa1<kkLmH+0bunh%9qn*f0s
zHFVxlT;V0MzQK~GRBT6c8_TSBu6tjLXjYi^sM?5T2!-Xg+lWd_XeVF(Qy*q)Pau6q
z^k{1u&FOfA)jmOhfPK+p`&F;^*gA`2kbN`is`=12cfod<hz(Y$q~ZOjkg4-tLLF9k
zlg^piqZjikMg-M|Ew6nKVoV%RN3;KmPZZT6kvQU_BQ2h#{s-&vA<}0;&~nvd3em-f
zr?auGc17Qnv%m5Y6i{^02dSp*=I`L7WTm%!eRX__gvhR=;eG6@&&2-CG>y;`=61Lu
zmb6I?!Ipa>AfIsKic^X=3y*R@>Lsv*ehTo_Rf6B9Iyd-wFHe}pR4I#N;e#Bs*sSkd
zSmjq91whIWKjv494^=k9LLY24kKztQ<jk(}Q?L{W1r$hNz3+#QJ*)sff!2&&soi$3
z(@sCkyPQTBIg|{_<x`z`9G5dCdgOLW7P~5yK59VhO+-(bryiy7kL_+(B$_|@I-#f^
zM;+?$qK^}Qo3HDL3FNGIK~RYUj)OKU2wT%gUFHXtj5KiHV#-tNLaln3=ZYru=t{i`
zgCX#Ggf`U#!G9crUbF`IorRDXO3khnw_@UCa^p71ur1t^+#mVGCZFc5McKYa<A03o
z4$5Kvsk(}~VoTvGZhWh3+qV`f$qU*KtR;1Dfb(wA>HQk=<>gQwzNS0H(pVZDf#GZ6
zorB4!P-tTnphmxIrcU|r2voc6ZJ_z7dN^^qirdF&>krI!6xM`0g5p?O{t!IEex_7~
zeg*FKE__i^uunemz>?vN<o4yN;vgN>S)iVGK59B^$8&1(i!I+r1#j>krMGG;q1~Q#
zbJ*D?&1a^|O*PjG2cwtFU#HeVdr`_cHG5sh!9#50XK6OZCX!cJvLf5Peg-FJs0*Ny
z9&Yw>dyNo<;$UZn;n5r%kWCJV^Evy^VtpG!VB6l~3nu<0AHSYLNO&qA#|YZeitD4M
zIm;jYrG%BWz+_q3s^|T96f0tMM9|=ZMwWApdFmZw*8S}&I!3=)2#J8-*z27jiC7c0
z^^A}b{p}J3q!~~6=jd~x-v0NXSmrpTlshR+a_+#fzw_dhqx{~_q;ge4SJf>0W9f-8
z1iaJ>d(DtvzQB0-yjHKbyDL*ZZ$~Yr^{%H}K;VJwqRmTAUnr|!sc{TK>B>x1$QPLS
z--e$RM5YgM+&QfDDhQ?|{Cow`9a~3^?g<gk2ok~~iTye|ps^B@?_jhPUtOPIk#D=1
z9{G4i_7Wq$PauhEpnrxKdS+Ecsa7X`&qiZAHKqFwRxewQMAPTM%rJ9rc~FH4;A#eY
zs5+C<A7S67b&uv3uKKt#9%C67meM<?GeS?zaPfb(?x2xh2oXOwoIL+(Yq?vpoR9y!
z<keDBd%hTJ3CCp!!q(}+1F{u_T30oC*7}gYu<lbrWJC@lW+$}TdU}xrFu)nnZAX3M
z`-%Rf(`9H+vg&eCn9J={!qNh#ynC-}ej{hA`0y4QJLh=`9na=4<HY50T^Y9c*g}VI
zaO$0$6fdwI5^bKETb-bSGqHUL_QimwQKjWdq<$D$&J=7DHlzhlA{^%(1-GVU^@4fc
zd}_%J6wVRPUeC4;?T%d9`*><yAuLPWwri@(${>-g&i%Awt+F{3>5J95pB*pOZA{j7
zj)H|A(|PM1?$r6JEwU4K@0y?2U1z<f?ye|I<-#Sj$3qo6kNRDovTBu)2w2xEQ_cz!
zwz%+z7J6}i{wx-i<hw<~!{aie`vEGXT!uJ)T9e4da+6GMU45zM#Wst{#CI;A4e|a{
zCtL|LaWYk~7-Ij0DsAP_6uRSB{&!?ukE@+8zil`543@#Q3Aan+bG6oV4W66KrUR6U
zWvFVON@UW7a`}CWA1*ePCv$kq4EsVL#2dw9@KEfxHv2sE0e@0K<R(P8)a>^niDZz}
za(qRs)pSboxzFne$Zk8pTCq7I)_u;><?sbOUG!ZOBV@RQ0q<PmeJ&e?r?DHBqhp+y
zBCLnHGf%I>#+3k+wzty5kjxc}5^%jpJ~n-iXlJGx3I1xjQb-jT5x>vZZy=yKTUq=~
z6YCQxxZk)J+Tv=bLhDoJEB9;%A|5@7awP&%YJWy;YMd&%PGbZpL8!$DSF|J&-W+vM
zsW35I;nbcHteb{*I=F$&Z4Q2g?sI*a{oz9Y@T^QJXIl7LCPKJ9hgd|I`p9*vTJ|DA
zo06pp4SV0_8$03TwJ6q8cg(PU^myzzSLe)(%UndLJ)aARxXYM*K{_xbs(#!+hkY5t
znVFr@Or*l#ZVz_&SOd`XqSRcJfveP*U`)%6lHb8xUE-~(+8rVGNtJolnT%(kUtYuM
z`dayP5?fr>URE8)XT5K4j=V+1B9fh#cKAc==k4J&EgrP7oc-wwJ5dU1aJ)H&YDMdX
zedBVrvYA9VWCiCYY8z9q18;nRi+ojXkLR%q?W*0WzU!N=D|5Hb;wR(I{gB$dhl)5{
zP4o8m>VNvG`X^`3#C<u@Kj9|A$K`w>J4N5kX8yS+XBYgUoHJ1|!HsFz{LH)8t;4pn
z=#Gy_g{)*S0cUgDrNU+!#&>8XtTt7r^E<@!ugFRFY7fepAf>K{2q*G{^xLzksF=zH
zXNi}8$Doa@eVed<pcu2CbIO2|j{!J8tculbh4N9h!X8Gd_f<*?APL*uv#E&*C>rk+
zxZ-K)+gNM*dCMS>+7PC~v61jN;^eZqMB_|5JnoLV{lPy^M8oBDD6*^8?(lFn9I&;K
zSc30<2A^;~97`ubzgO5`&}-M=*a@CpeZFJOOtN!|Y;GlIr%)Ze#(1F@0qqDkMy;<-
zD$y3=diPXZ7$7YUt&G%kerNiv<s+Mi;$^t0>TuG$k~*>GD_a%OJe^m5E>yQ+SBK>N
zcwIy3GV}4<@54%w5<xQedYLXz60J3@V$$rf+Oy#io!ez`dY}l?8-9M^XmH<Kr4JW|
z^1egy3$7U0h*3n4xsYraBI+X*tPQjNKK_R91V2n5tL3m)0ewtu&F3()IjbC+zc#m9
zC?U=61{#4{Uh>$-b)Rbz9iSQw$BJN#ABA?LKN8>jemzO#P#vsQTt)!>>EV8zo6G&2
zm@b_&T#I!YFVg$!8pD~8F)9F?dWfXtxX`PD&B_o;+ur=|)H^zj8ZEW@|4t*96nwc6
zMWI?Y#g<x+tkV>{IxJ;ku~?DN+U#BD%SM1&wEw*pJkz*-qFvDVRokw{dbbBrkahE@
z&sBf+)G0qRVcV>CyF7{yPe?c<7G3<%`YelNg2qntkabuP4GB`QMU(GG7!0SiSf3Rs
zrvfW$&AbCy$%gUl^w@3>%g3CNeT30ZXIB@}V)y%HYG%)QbqBp^aSkL}{W{cfPUqz3
zBw3>x8|{I`VD{~lrZLef^^JlN$C`-NWJglKI|hr~qTyBJ#{^BCqA84#@F_o)Y^zo;
z#jlYu{+lSN6nbgD0)>F2=MMD(hL}NchFkpiGy%~YkEu?yL)!BV(~!H&PNxpa3)IP1
z7@|lV@xTD7^;oH^a7f}NvoIeYAE)a*$qXL%UkZbTO0SPsW%j#+Lc_Kb8U`Ef?zboN
zG}m``fS|vVun>gb%4cN4dTl<?-v}t5W-Oaf#OZe{dM-4uZX|||6i=qP#W*+!A47`S
zM?|CH%wE~rpC6+kZQ4d5$#QW{82tsLk97G+SQ-A3tJXil>W;OeA&ER46-Cx*2&kc}
z*gDe$VZ;=a3-RL#DOUs~cZ>7wJme}B&l@g_OGzOhsrMej^ueQNcx{yPk;Z(D6kD)c
z?16)%$D@>!i4J(j)E<mB;K)w4zMe9>0KtEUX?f47eiGH{2v&ExE^vT|dFK0m(1v;&
z6Ps0zP7Rvsc6D?ezrL5FSl9sL{vL2Zrz>gqC!JfoaVkJeD0w8M!2Bj<>JnzQRPfBA
z;l~+yWA#fcb*d{Oo+e3OMBGoZivdJ@u9=x&IJ1Lrin)pgBT8trx-JwBS%kervD_sO
zF+vJ&sg6pg1iJR|>Bbnng9r|h_Z>%~TfS<Vw`WV!f#=<@8?9ivmMfSao8NQY_3U?u
z?_&?KAZ={)pj1XAbfo|onC}E-XmN#I$je50!CA$V1wv2lrB20<DlE!B2F`J`u`!0z
z(yEbYh1lKj0}wbkH!LL|Naq-9D(-`#>s)5+I=14;;7wd<x5XYgxSlSllE@DAFD03E
zX&khl-;Qr6)cDLQ!L7H^Fiz!k+f%@n{0w&gGO1DBZeQN_Kf4Q&Vrbn*MHJMnroHcb
zqBNg1&J4xE=NE4*Hby9F6OOnkEL7(6jgS%&bhojWOogJYIP8y1pRczKudZraL3X2d
zxn1q7u+8~hqdvZ8zfE4c_NYDIpM5>n&RiW~re3H?ry$z~*OloBH+2dLW8Kf!3#JnO
z<wpEA5L(l4J4PavO^IkMkCSWnjEMPedft^E=&Gj#KX^MJ*-!q3&eJu3ey~lOsHg2b
zV3f<c0_BwSGC7`N%T}K3MM6OJXjQy=I>vH_RPr%&mMiM1l(y(^6n=E9NNvBq{W(+e
zM1ly%&`bZsQEO$JXjNoU{?1=rU9xD1tu1y<rPY0ZBS0;&I+Y4oEm%tB<!Qr_zy?%o
z*ci^PxW-;sGc3)yG&;RRqRBWGp$fff17k9kp2(%*2Tm#oUi3ip*F`r_8V5Ul+to8!
z^O>Ujy(^yrwP02?Dz#dLqNA3)rX(I8S_`dj$tn89dl{m?*|x<EuJfT-s(g;5D||ZZ
z9!qRLW}Zih_C9u)0aDo70%_SmWo*lDw;VlRZ~9zJL)5VKO4GXK7<HMg6oKci$NcW5
zIYU@Mb_08O{bhN$J*!~0{l^;4_n&zP+Kq}&*Zao;P3=jU?CivWY?mF{-yMIP6${aZ
zoQ=NhLgcser#{YoW3#`;`IDGh&({91Abmf4siiH;Up*d{2a<)+q<biD$!<;J>bft-
zL$-Z)$^L$(B&iN&>g|QI>%-~JkSztXdzvFPPggdWl+>-RhSw|DR0Tm$E;Oh{6|xWt
zu3ZVYSgk9~YB7gw$040ckMYIN-!{fz+xK{eY~{Bl76zq)pI-^T>$%QKyL&yR#Sk76
z>!t_H7Iz$Nx^Wy`ar3Uqz=uzSd3jidsTdd|qdx{8bR18phB*?hHbtgeO|YYGJN@5|
zrz|OBNii&f_WbNcVF;z(KUxO}%Mi?Wxx}cCsXvqoL|QMvA{f=up<zS_X_1_J&#3*K
zaK6vXx}HpAcE%ajt#~av<2Jt$%unn}MP?piif+weveY`5*sFe+lU@d05H&d+TVp~I
z;BsKXq~Rp@N^Qk%w$79adJm71P~73Oh0yEYAZSZ47~dwPJ3aT3olKpWYSo)z{<dC7
zBl4n$rt$44uuviQNBYZ%{n#bdt{`otTp~w@SAuYZSb%}$gfAj#{MVf>i7&r9VXbdz
zH6_mVuwCnjNqR)uxYO%{(7NqCYiICTQsWJ({gre_l$rcT<d6NNV<pdnYNH1z#hUSl
zrQ;hM`-XKLcgvf3_vnn9h!>EQBu|%%0<K8gnE@!27acG!r8=4LqqI%S*5^O%9|)A!
zO)i&39Do06Npel3y}AnXDcSP@kP~w!pY6$3zYm?_*}|)(IR>Xx>BuHitA1BsD}QY;
zpPk)}!%~|MKMD4C-~7L!HHvbi#r#2_Oox-o+_-mU-O?f8_1IG|ay*zImCH&CLHH_m
zx>!>J{Jm1E)3Mph{n-zR6f4Fb^mf-fN7IE)S39ED4-Z(GLxp0|4kmRRUYkU&W2wMW
z`fDOj7EC%7Y3K~8HcSgo{y8FM8M|@SCJcX~$QD@pOMm7&Quc*@)`*F$Sft#SS!PvJ
zK-?ak(M=Rm_%=>Kf;VYiM(xT!kdXg_?0D&y(H!+k^xC6e4O5Iy<o+TR=<Ynd<*}eD
zwlX%vSkv#+9HkXf`!deMz^x!eOnU@OeSj)_54d>ZbUZ!(lay#iveM6^FV}}9^;q7}
zDf`qYU-*55I-AZFf@jErjCm_hEmi2{MXBfWvptcNKpAEkmNQ*FlskD^%-`qig|ebg
z7m99DYo=DlSIay@*~-JBB|yH9obDS>Q$dpe@2^BEIl^YGU38vCowQ56ho!V#J4i5N
z%-G7=5;~=OJO6W}^M)aZOVx5uE*~v640_9i9N$|Air`hGr>7^PeGKQciK}Liw9v~#
zWP-*+ze7xRrNQ~@Z}aIW$P{T&I@n?r9|cGb6Gvy58{Y2<GMS`Lom1qL=-A52i?Dk9
z)e6ie%au3oa=VE9`MjR@$ao6Wvca%e#a)MT>KRqaH>)mN9K3F?Q8C$f`6E@nBpT6y
zA#%DB;uAUr;oxnYcaL~2H;2`m`6g#pp`kP?rO2OO{r0#$+Wyn$T;7R*^F1g6la6{1
zR=6mEOj;=z79;2a!x7vc6h!VmnX1rzs7I1f<GDtsWkyGpbrTXMi_belDv1%E)f8d%
z^GtBRZ&@0nKV?~eSIU6JZhCDewh@*(EfjfOGloB^jKl|&Pqaw9*mmUcVKC2bS!gz}
z);5AJyfbyFM^$~qD>1ReN}-VW!+Di^lGa44{ECH0V(qGvCseJL`{u9p+ajDq_7b`p
zb^7K(zifDOajBdI6#qz3pkgf4%1)$GJN;IL-&TrxjHB^U!2S!ocF?jNoz{$N$m+7^
zO|a&ggmkcSjgR+AgJTU~pwO!&G&V<}%H-_0J$4X=81iq1j-wU2M$Ey{e#PS6N%?Y8
z!bq`Xc0x|q)Rv!=`H8AJSToe=LPbetsRH~`#S-IVLOw)t)~>YlBro7WU!@c(xL=x3
zBV|;x7;w0e_h26RtFmTAHJ#y3J6l~6^}a=Hq5TJoLNgw(dj{JVih`BVSF?v`EXeGR
zWz)UsRGI~>4X*;|@A{6L!H9(X##qLDq&h7Q#$4{(upREE1Z*U&V_~K-v~wlG%+$&(
zUA0pTY2x)%S$(|V?mrjTkHYD{hlkmB?xF<@P3RR{%$F~(>5ONv4v;UI-lrEQoT2(T
zdzZ*%2h|NdUhSG+ZuRtz5R^nZ9xh}Zu+2*yjHM4qjx*?YnniwrghWMuqh1yVb5jTF
zS3Fz!ZEf=n=(f=r42fHNkzJ{S<-uZc)Vs)p&mRB$OyluCxII_*Ir#AAO4Fikld|RW
zUStEoCW!m`Rka+Q3ANiCCIVxzAbKS1SaHr$^!HS(ha)Z(Nr17y=jYr4kv`loBM2g4
z40Z@RUm`HsMghB&$Y^#E_yV0q#tNSUEU6k;DFr%>KUs<_OtRg%hV)j}6Wf9$7pc$(
zNHS~qNNU=j(EEN_ueY-9<1uFp!k-9NQ#oNwE9xXchN}Pn#zrWQC?>KfM${@)<aAb@
zFQk`${VdM+?&Hd)(f);Pz0??KLi?C$p;{#|B%f;v_+x&u))?G#KPQ6vU9ZiUd%c}=
z*k%VLX)<Eoav=Uxa(8H|2*;C3e{e5plogZL{VC^H>mXW8-W2E~-B3t=KFu|t{W3^)
zUNDDmc=da7iM|{GlC4HLMOOh*+-P>3{9XzV(YpInwilSb(mZvt`B3LJ#A$yRsk0vI
zOv(W?lcOP(zK<xcc!v7zP+wo5Ty~p;c9RVNN10q<6Q56o*i&jGffRDx=QV_9hBS|Z
zU+(vxu8L9`!mrnR=Ec1Wd&jAwI=MwYM<pcHxu(=M<~i+I`L#(QG)HjUeQm>Tkesum
z<R9!Gh0H!gUP>d0Au<Hg2w&Ou_}(c?sj82B4UQIfW!Hi&T@^m$#xeHKD#u_w_ldno
z!APfjf3KFw_Tj1&W(|g$hVl<8H_9BFD=aQ3G2K}dLZnn~in6++F}XXeE_Pq8DmPkb
zKEGUam1(ozi3CS03)9on@134XueZ6Dz+nN&_YH+aOd*wC7ut>|kIi~T;r{RP7nWEb
z&ung2%*U6sNWBh^@MjLe=_XtKw*m_tA{bij#;6Dk8uG39B!ef*MS=#qb-awiLsJQu
zjcI>z(3DK=9S)ZLE$_kc$q9wLO|yk*s2@idj-*&IPO-H)CVyhudX+w=&afJyvJWuF
z;=;byP3*I2Kg6zL31mwZ8`1c#z7dN+rH7hEMNV|&O?cH*{N#rzh4p?JW?Qi*;PpYB
zuh2_Bac@Os)Gow5?gH63NFDI5to)cDQ2%j?fF-vtkcbEwg3^}~qo=~+e-@+P$LO*H
z!2y%wbxWl^SEFEs?pJ-Zluv?y|A950`(2Ewb7BPg;o{K{yiB^QXHT;{0y`v9nkVDC
zMVQ}~VDCoX(9!-WJ7g$__ut$~>=(l5G>+`&H%_&rGKJ!z?j3dwLD?dH3YkK-s#R#H
z#(PXC^odh8n|vffzNz6visQlPuebMOI8tLVi=<G~A8!wY17RqEb^~vIPf0sRxjeM{
zB<l&aWzT<RDcrT6-LLO&M~81{VMb4-4a;)ShcF;oNDW%tuj?KLTO1V2lmO5%qr{1G
z?(=IUY2fA?$SDVM1S!`59`PLY+uwqLkUKYjmzlYoPtbu0jtz*x%O{1$f*Lyi<|PiW
zC!!AqFpbR01Xm5uGDJ25`+_~SL|+XT{3LyZS#3h0wI9m;vD-d>bbmAn;N@Q4OOc*(
z<H|)$rylyonH7OWBXx#5pO-Xj?J7~uU}PFQw8aTFU;qsfexJ)B-@T$Pi37vqdl&Mq
z)tVEN%TWd|d;k{AY)}rQ4uC2b%^(l%cU(R%j=R&v*jgCq-jLzF;l$qwm-qL@sSM81
zy+tw{nuOdgwEj?RngY?dti;`qrA6%!8+Gz*Yaa=@;AC1yD^=XGncu_ddEG{(OvcG$
zP`wJjBk^pK3Ast!(a!mf;!|_KwM)2!;|Ge(jZts8#(_NU{>I9kI-oHX@gZn}V5zS&
zaeNvGwFS#4!F=IR8RkQ6#(j;EgJ>`8@(f``Z;+3iPOK=Oip{yPNmc9<>T-DtVKI0R
z%$WU_sJjvdUu=##1lXDe^s(@kT99*%rrgm#uXxdO7KmY9!6*+)^ft;pMeQM5zs$Sk
zOiRh7FxgECm+Q3=U@%nI%_)_?2!|j7yq>OS2G81H(k`a6GuX?eoy*k=xP3r{&^dp#
zD)&ObH2lzfTy*2|+G${;M=ElX6|BITAjP}jy>N_P!?vNYJSe^4O^sCzzpVBSHGAkG
z+1qs#g>k{Wk?6kB{B5CnRo9%LHMLEaH&M|iuh`5THh%aFj{+6oAWn9eD$r=PTxz*!
zGxr`F-r)VNQD?vXTBu|LQHIlbQFtMUAL1x}S6B#-WE6u~lVMyf(J(o5;55{eSV|(?
zy0<fK4xrTU^qMP@M0?&()i>)4LHKbaL$xs-kCYd8=?@Aiw|~s7t=sNaSKH1`{<ZtI
zyZ2+EN;`?kNNoPNh)HlcEhhRd2A@0Pupyy@wXQmo&657^Xg>Y^JtGu#KzZI_uo$*A
zP26ZSc-SV5fmL;MU&<#c9BJePSAVF~3EJogseDInk+6$GJ%lafLnvJ8dxTJ*R~e-v
zEEd)9k+bI^nApu6d><-EyYT}ZD2wx;RX{9q@?Q$jXw8hb$BL}Nbclm7eF>sEW##j_
zjBEwduRecXY@^a&mxwP8>tlI~*hJTfy~6B30{MJ)6G6YdxdplY8HD_}MJzIK0Jy_p
zt^T31a;A-Qz1R>Q=e{cv9!jcLf?ZA6EzaU`^;fH0Ejs&#{`lp_BO+EOJr=Bw;$5^V
zgKa8>6+P@LJAE>}nud7#$MLLbMwb=F`dgi%-P?gRN)trproz{W9GAbhoF|v70xNY!
zCB)+E*}4Q4AF42=s0w7r>!gepw`Njtv}9zrGg)DgHoHS^^17{ZG)-Cuqh@oJfVqO4
zyI;?@Pmz8XEm)5VVM`A-eqp{Z1QD5d>pwSby4@lYDLJ+6VwJs)eBbV{I(*(Tb_OC#
zI4+L5FZ;vM8oZtyvh%@0Wc5Fz4naVx<>m^$Bd$G%f{)77K6;_Nd^#M@6n8yeL-~CA
zM!z%dEL%<Sy#kwFFyg<n01EwBJ)Uo=WOJ+A>u-u>GnFFmyObuTDUNU0=aGoWDT3aR
z>G@y&5bMH^fMXEbwu<Z+K@)*d^u#f!9~2VEi9vKc%RF&TYN57bofpQ?i(xjU=&R6t
zBpV`qrE8On)@m%lq*F1|^PH?HpK59fD#ZC%sxz8JX}t7)dE!P`!#GkYoJ`^&0PJS3
zTB{59_x0->LaB_PQ^CIZVdaVtM*ad{Pl7Q@^jXvC!l&xmu_DxRxCV76<fshBR%rEg
zzuwEIQLk1w!^Agx$$68icQx}FfW^pDk4o-xe&CCIcDMO($v<i;t*x9gZCugsGKLy3
zGVW48&vKgl=nNCS%H=Y?<r--=nRO{w>1t6p=*y5LQw94xL#z}v%m)<&uY19IbvfM&
z)?(f@%I`KpPKh#OGElG6U3wKMtJZ9(gl((@i~!fP1yl7=@d)UE8VnzZBEC|v?GNV!
z=Pil1{}hy$N+wQct<37ArM=OGTP)~+*KhFY6Kl(Y2MyZX=<xhfiBXwi*io5ajM`20
zQ{c#I`PV=JEEa=a5zrZEsi^5e$3!kx>(a34wzsz*ov*hM>jAwcbTC2U!!S-?b5s5I
zpR2D=H|B>E*%0`lyk>tN!{>?&y)D{bF*`m`z1fM0R>z{ObLq`_Rz`g8@|Kz?_Kif}
z8})H^7*p;}6nyzm{P-}(lY;0zK9@P?&BbY>c5IU8m1EvwC4*cp=j3QT8ujfZ+_!R$
z!agZYY6usi_5gRTOq5Cgd3G=3xSxq)=p=t7^G_0p-6^K7hx8&JIe_;CNi+;Fxm<6G
zQ`5}n`~hISJKYHXF*xVwg`5tixa#5ovnBdm<1TGB*b8Vdjb-0A#j{FQ`KNF4q`rAb
zgD^VNX|t5XZ3$>`6cZZDWO8}HiQsqWl3s32E1$4fQIkn0r?A<qRNiG9r4FW1z233d
ztZCHh|N0p&2<PbeyXsn%%)l0g$NgHQ`|p-Y9}vF&nmCaY4#f$Ao1gs>h1dQo_oO1;
z;xJumND<%tTl}K*<=|zm)|gv0ozUOW<Sj)7!bQ${#GMSJva$mTNqKvvs<f9!Kf{Y7
zEvK{oLayRQist$D6=xG%ITk|ssihO0U!QJeQci>9WQjheD2LZGMoza=&Rt3s;JaXp
z>2^6S{(K^FMg+>CuhU`(=bl_92XvQ3%%&6NBVL}+TKi>E0?B^VpcF|U1}AJ*Ya+jK
zSpYNR9#e#gSPBN0XcTziE5_nU0?0JiQ1ZGGWEn1nC(d7IQmR(}R*!|^`F(U)F4Awm
zP=5~p;+*z{;x3VhCnc^>sDh<PY)u!6vn0LMZ2tt86aTx>!7Y^Zbo7<>Zb&v)9G~rr
zQccyva(=}L=~@D(@~B8Bgx9jT;Dz-=+4u?wE~jxy`B>~0l_K3OiN@A+T}yYjKn9N?
zUX^Bj#`hL-(wey|N<?pxR1JQ`ziF3h)mIuZeZ+Qp@8L{)ad_M&@F+t#sL|>-A6L_I
zvKrJC;fWHdqnP`5NV$IOBcsFD4YjY`LkixUzHhax`_aj)Nj7fnhA&pux@GPCg!W^p
zV(<Ct;~c_=Q1(er6C59eT~~J{OHOCUrm)cN@<69n2pSlo0Q6z3^TgX9J%LciHD$`j
z{ljP~zsw9rvUzEE5n!f!f>Q@6o#i969Yxm_mV+lklgO`9QWHz4(g_Ikw#$sntY*@p
z_Ji5qNOpq;U8gyi6K>*qXH@wRsm2p3=4{O;0FSyM=nm4^5jGolg|mkYO5r~S%4|R0
zRe(;jEX+d{ShD}*P8Ey`)$R3>N%qCEyn9M6@%tBy>oLJ$R5Amzq+H4vf-LY9?T;%i
zDG#-QtuNRqwKF#oKB6Rl#NCgn=Y@07Q>rBrkB)CrgF5|k`EaOXze@o!cZ{hgHC~EO
zBpd{BIB|B3ty}x`TMH1vf%l*rPZ13G-Yn%vzrPKeU;<e;X(i(4?(=p|qT6X78<5LW
zxB#>TxqRMKx~)zHULqf!ms_B+`>b~($uv~bDYQZ;uzmE{e<ex`0WTQ4{o!Idf)cSP
zBq5jyQn1qMN%6fi80HCUdrv9Fr;CC*at<5UPC>H!RPhXig+8kXkE=>p?bc<sQuNPW
z`}Wek4jW1uts^hiVOS&ndk;7xlI9=pFAP10AaYqctxn;uudfgXuA80SRLaFN1v(-h
z;3GZloQ{voi9XNo!LTM0GUhusH(Om7==_ZbPMm8;S4O)-%z|iQ`~W8%CZ}UleB*t0
zN=s^lr|ajzUI#wtQOTXpIWq?zrHoe#>P-~_w=@f^r{a^$k;u|kKV2zgisHwrX~M{M
z-@)E%E~O;}O)7al`ZBz0ff?_hQm2nCcZjIhk2mzd9G}<EPa2|m<j}8sC_UuxszNm*
z>p&y|)ue&H_~C(_)!AB$qtn?6Ypd@^XLmv3?)0|dS7c;?Om^GiPm|?!yd4Ggoo|L%
z;iD(YiF26nR;jgLI%3ztJcGz*@MQGqG<167{%o~OBA(d99v<OtTq85&*_3h|SzdU*
zh^4HZ2XT3q1h5U;tC~R=-~G{9ul(JBv$d@a&N8$>G*T=L5I6^=tj=*C5J=4t)9P_Y
zZ9JMH8zqZNGKiEg1T+T5--abWB3eUcNSBH9iOaIkQ>#Nph$mOBON|g}fZkN$x5Yek
zOz>Toez6;lWuFSZ<xl_{h*7QE)eno+#;RKM#Ls%ium&0z--8{>P$cMGarbaPaaf7$
zzsvP#jFO3Dk$M~*33ziApEdr^37ZU{l<|JC2p@cYF5O&`K3XWJaLCoyrWiSr{}i?6
z@iMtCz~fKiXAynFUc=+}(c#!&xy@Q+s42r3k&lp8w;V0ZspC>2!1L&IzX_2_q9kH9
zYK+ORlW>^+{)ED~(E0lq)%TeCj?g=PrikXcsCEvqdPYlmq)$~3(=ZDbB8E4sv{zm}
zUnrQkpIG0ie~iEK=_)7mj)kp8;D|;d0las{n@&FU-2G`u79i;Wj#P%&n^N>YrA+Wo
zDW{L}+iigukEO*s>@l9x2=QUQHI8330qEVpH#b*TGq)wb#*HEw;kXmZM6avl_;l_4
zPn|C*@Cut#A|;&juPIYP^<q5G`ozwb$TMsx+h=IEnu_L2ydR#P(y-wM1O^UV7`#4S
zIr~sBLYiA0m9y~akZBU>CPX{2riOUAd%J5)2v0y6avrB99u7YZ&jM=8yp-KbZA;f#
zEf*9zJno>cnN>fFJ9C6R*;H1y-DsyJ^!<mh$}|4lA(5>-K}xJX?C1@kHi_bpW>UD-
ziEujt>BR36Pb!Tu-wkn_G_s_mL_+;T^q&Jz_x|WwtIC5ZOnkoXu~4HgPoq|u7lA=@
zk{rf0U9MW8)aCOAtq!DNj&g{#wY7d?RCRFx3W?ZMtMlp6pLh}z_CI*e@D^SF@w7m)
zfQU3mgcs#Gu)MM8hcZ#-H}^+VXY++a`^!SYER?Oqll^8;p<!VQz`npD%P|X-9hH85
zGd1AaZN6MJW<~V3emO*C3P`8pWG=t!(M1jmr4=LaFVeRd6IDP+A~7ygR|SJ4L$+`*
zW7L*aAv5rWF?hg9S=HF<_TZ<cxgnC>lG`!Ml`4n8(*)xPu=pc16#on#>~p$MRR$DN
zT9%CBC%G$F|4BLpvMn>zxTGhJaHvx*adLqskN_c|-kYHPJcB754G>c~7V1@3&F6fQ
zv~=WX;$GpcPWDX_WMPjZeEVpVx2&snB$UKNV-%>zGWmR7DgO;GkNOrNLR2=25@sxe
zH4+G~;$3g|U$+N{xdha=!e(9`E=Z*MEx5}eddWad6^@JyKEG~_h|i^v{@oaQqEbca
z{|KZU0$#d%b5mdAz-U}IU+7C;FdikV{Z9Yq7q+2J;#V+K!4@PzN~Th2HHFq&o$ITA
zkS2x9IQ9g5nQL)0RgK4nxmM3q(F5~a{Nh!~jPk9_UkjA7lv=gCbW*@Hqg~?Y(<MXX
zSIDCBdOs&wEL2kAQG(jmeV$e&Mq3{2(~xdaf=_vgy?sM$v(_xf1c`@ozR}UPgx6nZ
zlKg*p2ga%klf~&6^2=mgs>Aca1=g6)qlWs{`A{L6GZz8`0Y)F{a&{jcLSjh?AT@ot
z-Itx;W<jV5la}PiOnuArP2<Rj1lb2`G?hV~h=_=MsLVu6AEp%c?TGs!p3MP|WF(Wl
z#&5Z&u9H^tmxz`=aiO7VJ|T{h`f7t17{|sTtvDLM{_XwcVy6fRlh>}bD&aq5`X|D8
z5Ca!GFJEAqSge*9OHd+VTXtEL0WNc4Kqr|_5+xV)ll$lO?eVPn@-O4P-QC@|H)U>s
zRJRfgM}-jKZXlYL7uNPStV}B;LcRv4C5sj_Qh$wUv&s_%5=}4#)cfPv(EIas>Cb&d
zR5B?Mhgjkgca5)Q*w67Kp1)N`Nw;Y2xStFClld_D{jRvY9_<=dY3EWyan=7X(GFu?
zrhd!ulW7+D45=ZwE^IJ5&367k*+?!2;8xz$)zvj0NumOqL8UnMKmEKtQ{c+c+>*KP
zHpnE9N(L!?Z1n`nIj)GGbb0))flz_-ei>2Ue-{QzrPuwZ+t7p6YxJTUN4&zoQzpv;
z8SW~!|Df8b1@rxcjhX>=C@i2eH#e6gQu8eH|JJUy0^<GYkg%k8Y!NHOBEN<Kh4owp
zHI3J^Q6m+MuWz~3dbLp~oWt=z#`LsxMPC(63r6HiFdrniZ=%BAR66aHYW=Qi{9IgM
z4z!Lc9{hC4Kxl3LUhW2p%FY(co~`6CGpZB0W*MdwxDxXFzRNVrXD4$xpNQ1`7(h)6
zghIs20~Qt6sluF3#s5~Czn!E3(A>-w%YNl@*b7_QL&QpysD^;^&bm~UO!ljn+?uaY
zZ}5I`$zZe2sCU*~-5iL-QU+eJRIyO(a}&e56&TMJEvnni0RowP&YEVc<*(eX*K0V#
z#I_#;5ny~skI&R)pU0D&@EWkg@VH$Hv`c`1?E<}a5|J=u+wXmuDAnLj=5e*#zchL|
z*#kt?bvOZqXB_rBVWLP$&n=FJY-%d9|DWV4N`d(;n=e-5Zx?0Mzm)e5FErV1!a}tu
zS|rT%h?v9B>NLlFo)9yfK<2n&ztdD`yV?2Kzbo~ov9mG4XKuI2AZ!b@On9oLl8NL=
zjD|wnXi$hv+uU-1BRbGS?hlNP!%=fFAY6M;tCS+ER;We$-2td?gi8?ldoMP-kYp8v
zfsiNzHUZy4f(4WdbmrH4!w_QzcWwnb-z_x$!<vF}h`qF@iR!T-yfThw#XGs^*4<v1
zJKzl|#S-vl`NX|F-=R%#0Q>0l)<_Ww`Hd7=^_%M+vwEQz!b-C}39HRonjNbAXHfG(
z{2r5CJB3heJeUlo6In#Iwzkw7wRt<%9Nfb5LYvE`j3*EL<&?Kf%AW_=F4Y++1EiA8
zfQm4QLBF$<-~R+!^*{TkB_&Wj=E?@uJV-91dh|6hD&xEq)h~h0dZj^ehZh?;1*l&0
zft&;@%xlVLzg7AWxx(-w2?F~D?C8}R|BrfCj7vp4-Pi4)Eli!jwzf9(lBC9?Pz;(W
zuz~dsg$V91T3JH&uC8D`bY?<v2B4i!lXb%z`{xe7Tw8QDcClPsyxwhXTpq?VABam2
zK8HXWtyhcs&dQWalx_~jNlGc!tVBq<ai3AfLD?-nfXtZ0@9T4Yd6@^abIk+NQ6x*?
ztEN^+b0~itvbBi(p2A7a&xJU0I4l=v%k?_AdIZ_gNV=gJ!J<G&wrrrA3mFM{ehf9A
zvnj9p^7y`9K{+4K3|K>IwY%8^5jvfvYnLr>$6o{nVogz4-w?0T5Gcz+?IKL>|1L|h
z=oeyxRpt{8V4sI=`#j%<haeHse=)Jkxfy5Q<S4*tDnkoU%KB;W140J_)OBV`ozZ5g
zmgsXU8+hsAm>_h8$B};{!GvzEU?dzfMDT7(D^N5+?(FPTs<Q_6tqtn|u$2@ivN+9w
z`HEi{Zq#N9f-*t@JExt30s&S_F_6$k5|1ufXWLwDBki_j&I3U#kJ^wJNPd#qHL4cN
z;BdM8#m?sU%?W<OwDY$<(gr~_4Oal(Pacwhy^h;-fv8G}9KJawt&rfsAP%UQABq74
zHYftJov9(;7BC<Lav1e(7KprT;Q#lGz~^q9U2zf<e)JhIUvxQJnQn8v$eJc?>VaO0
z_G^Q?1^sk{00mzlE3(*a3A8co!V+VF(N4wcygHO^2r;Vt`YCB%h<ktC(cZ38?pwWr
z)0lrzp0({m1zZ6E!rcnpHa0rlRu$L>Bz&$S`1Bg!>im-~fg<7V_24H1Xv-t<R&yL<
z?$?9yuTL?{Wy%yy{dqR)ttzz!f}b;v+A*<E`##VjU;O!+A#?Gk<~DyB!#STWM0a{U
z?f8uCq2P%6eR&o$+?H^JhJb)Two;PM;Yt7Gl=U`O_SEcIcw_?vZ*YOfg?(hd0|d65
zwPcIAQo+w%z#jO?_qUgI=QJMrf0v0}Z0l&c@S5{}lK&&FjiZ}&u)04St-egksM{`U
zl&IG0$)Q}Mju;@HyRrXVtQuM@pG(hr$DrF<q?)XQM6$&QVQ_#w8CNO*grh|Ax``ZK
z=_>7}VxS}w4>lBkV)`Y(KOzr$3gWjL?e%m$jFpT=A*YbT<6c?e=ZFS(>yFyY1!foV
zl|l|r{i0a+twQTJb(Pld2wZj>T6XfY@+dfFuzwd&PGsv~JX57yg$kG1BuX3Z29#<e
z6C4~oJB*%v0D36|BxwD74}K9FwoD0-aEdqvTZdW86vQ!+jC}d<P$CkJ+M<+-w}O>c
zy4vZ@&Fk};28l!<L&)c4XXdHn-cQmE_7-+1Sabx68Tk38B@p%i2o`3CZEf9?ic#SN
zysZrtf9|31gBces8Nkl19K@p$Sa3flc|JE1vbkN$6FJ8Irx&`Rn@Pa{OF*>080(dR
z9jMi0lie_XmXJMHs|@^!Fk=gKD3KppZck6or~eQAy-khL7vmF!&143wy6PFSb0`vI
zNdH(uekuP(?p!V;W;2xS?QKE35pSS}3`nP?JYi~K09K=`{)PqHug~qva8$BCi;L=9
zoxvtFiQdq)DV^Sr_U1#xeo(&cZdZ7#VUcrs)jFIjwT6MWUw&DyvSA%#xCZ`{dteDb
zNBXUp_yj^C^qSc9VY@_}hR8Vg5%u6yw^<(2SVdye;f`l>r-s1e2q6*hC=@)6KsTd<
zDGvB_GEEr`M3BiBlt`z>0M-9s$y${E9=Sa*GBhzR3+yF#ATYOk3Hp;O00jr$yh>7e
zD*mn*m5X?5j~E7mcQBLPd58V3*z?s8p*av1fl&0notr8sKsV-Un)wvsS0B<zMC^58
z0sZ*&B*}~gtPF?W)tr_5oQ?-c6YPZ}<RN{2dlC2{0xFnC;3`(_<GK8~Ku;ouL9LQk
zBA?3zZ&}T;gKXgM({Fe&L&xua9dt5RCJcu~U(^?hgch^FWF86vcl*09!E%Pczm$9m
zDggQ9<6+ZxuF+bXYHg7~wL1R4m8wPXFQp!X(xRPevYxP`IwWHByti_nK`W6gK>d8W
z<>lp7^i)&IVnblLC-AKt?zV>v`mRJSTZ&^^9f)B-|L?PBK=3g0141LumPtjCVmF0O
zyHvm*Or=zT2&he<KbiIfenXPPIFVZpEF-fH#=n$u3Q9pV5;IxIo8vja+ME2}6%zYa
zfD$&+>t!%W?n>+d>F{!Z?3aBGToe%;M{Wz+A44&+z%?zb)@_>sisX4oJQE0FJW$M|
z&|d~eQ~94_fokrP|INTciFv`8NZz2`Kr)Cxi1aQtI*vDeKLEb(PoX0zbQPNGI`Mo!
zhNy{!+HAoDQ&i1EqDQCK(ICjIRxE_{1~LXvH=SsWZEyTr`64(F1+9#PQ6>m)l+5cS
z`vUnOU^mP_l#fYzYigQ}!}0=B%UqMK{`1Ww|5;VcC^J|FQ8$>kH>Kz85m9alV+lf+
z{nPdS>}0oVJSynIRBV#fZ0D@$WDbB>G(tf@00h`-F~LtO!j<YImPe;OJ?!)$-9VF6
z7^SRZG_@NvGm*^=u%7CAe;(Dqj4}G3+!;XNMS#HFIA03oeE>l>^v8HIGgu=rqtPII
zS-y>>yDKHT%^HVRmrn=J@&)4VHd8Ycn7_Y&5{tRg19p^Hp`ZD(`E0Qw&|Dq^wFmC}
z$h-*YCvV0dwF#{1cY5Umg%%1y|D(O{$J+z_6wSlm-i;2=W1ynAE`;D7tT!1iAd^m!
zsM*D!Q7ZxhaNA+GyGlGFs29gj^jG$oJMB@#!|}{iASDZ{maF7%Z5dcB{{p1UJL>*N
z9)6M#8fk>E2oTQ-96x|`jNl)M#V~WW+63^nP(RINeo3TMB!1qF<u97#rGy-e8YU(;
z-I=t;5eNFgTt@cvTa@l&N6OTlpX0!0Ed|=DXPifRC8F*M>!Kg2joIFhmqc%myRi_f
zzQB?#fXDf6whFR{8PI*YShN2FhenN-3a?)*A_HhK2En{pzv0CwKm-H?6f}f%g85YI
zv_OBgPo`1-M15tTdrl_p{Qe(mBMNQ-y~gM1!&)64tb*Gl<!Gx@u_w&k&O`j(&rT;x
zbtIy(6JkiXtQI0bxG7*V8l1`#4Ae<AU#q7K1j$@u8b8tcZ87W$-)_0(stqLi*>uhU
z0LoTLtaSa5eYSSJNpu57^e)@W{h5(r=WbpKs}WWM(2>uOH{~-Dn%9~S_Cq(rf+;Qo
zp#<rd=_Gydvv3G}KCmQl$!~c7Q&M?2U{hnZQ+fV#8NjO`sv#jEm3altI><!%ReS?+
zB$%=}o5~NYaH8scVFDVM3@&GC-8L7x-j`l{gU^QcLF`U>U#QaFAI<<SM8@P;8GYoZ
zilMc{n`JVKaJ|^bzU--WzcDdq-(;e*uL^+{Fzd449b|M7Fs#XP6-Aidc?lB%()#m}
zr>Ey=^du|>wM4!U(7N0)U*i84r2z$)mB|o`A7ds6sKu|J!!o8j=wjXB&_GL|;L`<=
zNuLR?YIRXQo&gjBO_~yLNdVtv8qB73+WOYGo5X;Z->%2vV*y4(Cp)%-6Nx4+6A$;<
zWIC_EDw^!PKuu$re)NAEfT`BSa4@VurqpZa019ntnRGl(hf63aY!wb-ViX{pYTPm#
zkB|mG1Ie5eug>Jhe{I$v1fu1D@RhCT6F|<EaY2*E&#^&+3yI-Dvl>IjAAyJ@ra`Fl
zdz)UjbqXlF%-&z0s#iqM)zJlk!KE<alA8ck6REd_TpDA~FQKeaV<32Uj@I_t<3HUV
zqeWvg1)R)RP~dO%S_PGql<eZ}WylgwRj4!Y{@v{l?~4hOEf|172?k<%Gzn7vyPl!4
zFVF~IPrA{^82=X|t@suY;>ecjjSV}^$F{H^_D55TneAXnwnQKrZ8znXep0KJC!uLa
z<H`2qcDK9Vl<If++&*6I{+TvE2WS%m(Q0%#gVWI@;3MY?9Oe|Tyy+PJaFToitT+c8
z-l{*h>pAjYqqEH}eT)4*dJ5%!qNxH=xMOtnV^BHixf*@_mVFwMEmY`QflZdLTReCs
zC??|>qPs(}?Ah_k|3-=JZ-?Bi=t{Y^*6JL|dNXb?0OrlSWtY&0;%A^B%WN`caOTDO
z4RRYOswC^V8E;1Wp+(Sv!k`K9iN$<IBL5u-MdItN&LqNJu?zEqE!n{WEIp+R6Bbdc
z_Vi%Vsq_GSpC{4txIQ25)jCJ|_hPM=?NCDK7;!$Jv#xFXvjSHfx8r6?ba^yQa;y!`
zX0cs_gLDn7a5TJWLGUu5m7h3~Dl&&k>nZm6w20k)pz+0<Yl{Ce)S`f&^+Dz#xJ>Z#
z<D<m*>*r1~kb0?1#?o$IpB&BGvaH`ShGLa}3xj~2AQAFOjL!qzv?8!_kAcW}&8yj_
zwpKx&wmzIl5zC^i+2&&HasM~HVYb8NOlu^AwZao2MSK?(8fdruGFZ%oF8AYzMPWZ#
z)cfVE2}00$r$4+5D5+;|8>F2Ug1N6w=4pCu*?`th$%qiPRHix-SR%Tq>}(J4ZSmP$
z&ILd;mFltIXm_`+j)b#_?G8tyD8j^S$m4c85&+h6$$hk4bd(<kG_8K8jz{y*Xp@ke
z+io9@T;^hxwzyZTxFGp$#CS8XRTP0>Z}umS=sRf4(9Pkb{M*YzYFfy_0rI~lKt-f`
za#9`^{aaE98(&!P(}8L4Y>7Nz22LiZ*NHNyOd&`Sx+M1DY_(+VIvK@|5=_C(f0%-@
z0|E+4d4D8%UveDCD%=h0fAl!zz6lDNZFT!A4PTCDvd^?H0p&5t%T)6ReHchig&SKb
z%t7?0NW;(;kH#fqot!_$V~YXVQ@K1vW6<NFl8*vfp)u&-ZJ>IY{bfArCOpgGBp3Us
zK0<XeA+H=FQzHkXzn-dwrD{Vbem+0m8&2%+_2PBY*@4W!w~L8{iX!^ZDF|E(rwj#Q
z!K^JVg?T_e5SH!&R`UrdqOT~(DeZ#J5-$OZSsafw_>d3~cL39c;x~$-IaL4BLlBYy
zk96G=oOrx?jh-=7^Y$2=nYvn)R`NHHpkMx4X|aiccJ0Ic&>V37d3oQK(UaI130=?E
zB$s~~7Xp!B)YaJZj^x*G@J;17^=>7Q+CalFpt2KhG_><2fdRWT2YrgjxTK4036(W8
zJ7v}k0ol-GvtALo3<u=|z0_{D%U2N#-pmCiN?Q%((18O;>KArl__N(WHzZ?c1INrY
zRsD95X}fp+urm<3A6q3BjT>XWBvZu!7MVQ_B*$4GkL=lV(Sjkv7vx3^6SgOj5UBTW
zZ&$C;sFUfn6lnIb^q0-f8y5K|eQ)J{`-t8bM^wI&s0Z*FIqw)lG6YU(_>ng-jjA6!
z=R0S#*yBhmH`O50xF=V5{FUM(8^0`<6X!H27(-Yy(!-!s5SXX{W@@C+Da>>R(jOQS
znp|utjAQ-psY-uG8uD3agR?*}XgP6S_iGUg+y6(`TL;C_FK@hoKyY`5;7+g*+}&M5
zaCdhW2$J9gcXxMp2(F8}F7EDkdCz<5p8D0j{I^xi?9P0<pZ+}EBSK*a!i4;b`uJbv
z4c^tSYZc9p1p<VH6;c5%J@4JeK^0xD&3<Ei2rlBF2Pncf4GeZ@2VrC?%p_b03G(xc
zZ3$Wm_S^DLC$k+sPc;})<g&Q@-dquQY{7n++7_?@J?{?WZ#Z5=+!l>^7e5pjE~rcx
z4z)>*x(>L^6WFvuKJ2)hEY-(C`Xr46fcRn0zDPt7*H^HPpP)h+a_6a6%w)qJFV;%H
zU&WJP5|(cT;RsV6C%iY>vQG%qBXwJ7W2u2dDuqSfi>z^MUYwz?V&+f2kED4uTd1&5
zkHUT=m!F}8PWH^B7Y?`jg|oq@olI?BMj}-@rA#IcA{l-}^deJPym`LW*G7a{ex>M$
zU6xHFv)M7&B3}lR@_8yzKa^^K#50^LPY&Znq{YC8%)~ecnWw#}@@v6xGC^WBy9?i8
z)gO)K;Zwgm!|?yCNg@-xG=#vg5@PcM`D>xX-PyFik5_(lvw6(t8C`4ZY83;cv<80U
zh7E%P87&RPk9CACt|yj0CQ4OMb+Bamiu~GHNYkxucHn~1-P|Q&`o1`VLLUgWhMgcB
zCQth%-V=Het^)GUZ+a*;)FgcPYX@{N^->9Ln~$OS)qGkDnKh-If(21|%OVl;YF=@!
zW()UY=5K21k6(f75yRwHsNgdWp%2>kifO9yvsua}+j6m=ax^IwyZR1<t$zv1=w*Xl
z&?H!Ei@~}pjj2Zt`_2KR8#Xqaq@;qDG}|Sp#Nk}GZU@tpv`}hWL_Ym7n~^=AyaoLP
zUglbHIBXRamR4MyTfYmCf3S%{O63GLNvZUUEX9I;0qO3$jB1Gd!BigG*x0x>Wl%;B
zHp;c_(=>|c1jUoee#u0mG92Bv%!K;eMW6reW(m#+e`shhFhrYx6A37IjiXh`o1pJy
z??4fIAA+$v?@ywde1ugeqdA_h)HA0-{sM_C+BH9)2#>?WFy=}#COFat){Q#P8_Pmp
zFZvUSV0!Pq8?fw;io-re0y|mOyZd0xBR-_5Z9icZZe#Y`G5QY3&XL>iPm=Y3TC_J6
zilGbI$rj8)HK7)QefzfTT#f{u&wtiY7c>1r`2-tFC7)Il>1vM4a|D**Tq>lasbnLV
zVy#$lhEX?~t?$j?9BQ<fE`mtss!Qy!n+v2<EfVBP88haA1Q#{(dM9SsT(8?>3$Gxh
zyq&YoH~B`pb!76<f%EM8x1-8VaWo3|q<=Nh5^958o>5h?kPI;vlfjrpRP2pd;$`;s
zD!UJgEf-9rwWzR&h;bmL_}i*tC<p;ePWzOkAudRPmun1guE}H+bf2lqRdIHKWfKy@
zd!z>@Oqj)gn4k+De@@qKwMC4@$yzokDkZwZWCZ(&Zg(Yx9LzLhWG>%>?DL>u!VkBZ
zq^(d1Zw`P1w5`VHnkWXZIn4m&y~vX|rt^5TU#L*r4OSvn8jP_8Zs%Kr;3EX{6d&!6
z<7%Z}NTO6%u!sbl;p%0YWc5Pl)8>*~4^=Kl>b>S2X8$2Ddm|{)z|-~4&J1<t;IJ`A
zE|;KWWb02NvC%{9kYFPf59v+#1S6CLUoJ=`-lYZGwG%0<NHd)Ft=79zBx#5jGG8XK
z;Qh5Krz_3*M)s6<V8MqYpah>6a#X)>KX#KLVkho9OlpWB%3N6XbZd;IdWwbb=TtCu
zqRx--4Aphg2Mr3}J!heijh+bD_`Ubu4i-O(L`)(|Cb48=_NY~S*Xl4{nE&%IEfGl=
zL);TZvLeG#s=<+&tbzwYO{Uml498t(#-moQt+r(1tXyk5FT4FW0*-_x^oq2MQFfDj
zvd%)qBE@_xg+sN@T*;zTX==%k@dd1>g#YOe5kRu^XE(e=BCS(M{h8dK%FVxWHb<l*
z*gxfTEg$$9Ys=ygut#ULXBELTk3$TJx-X&+4PgK#e)G527h_p}@;1*LOHVLR9e``H
zauK`^i1aywOhw-iiqQW9)Z=}`Rw{O38)>*{4U+$R+im^u9cU)x+Zy$abjbt+1bXh6
z!C3m<^XgsOO30=~kiZ<S)@6`(NA>jZPy!nV3$WRBOe9zuiuCU8vi0l@MzXjGrBuzc
z7Ycitz{7(uBUP*t^j)fV!4RTBDmSM(LL1B!==`oz_1F1qZS@g8$Ne|BK4biFw~u?r
zB6JyZwhhTGusP4=;^y`KhZ1v;Av<4gQ)k8pyuPi?8!(eZ0xzshHUnI)KKNR0gnd}&
zcONh^gi8DUfqlq<o0!^a!&=2=yf^vf*Qkw96+Fx);Q1otnW)qEMVosjeAH;Gqm>B~
z4D*T7RSQt3Q@J5WuI%VHh}7HqrilFT8w4K<kuG5RIg<9(QvZgD_Bw`DXrEx>@XWkl
zE~io`KZK)lJWL^<rWxR$Zk6{l8+JRC<c}kw2$A}<k|NRHcJ55<gQDJO#AR4!-4Xth
zW%ZDF&KZqsD==g`pRRoSjdA=vkya)6{Y~pcj5%$!MG^M;+wq@iBZ&F?wB&-9$0Jf3
z&0l-^uVtc6g{?kYG(C`ue9|Dn@SQNCk}Yr{;DnQ~dsQtF{gT3BN|7%WyF}C0F>g1?
z?DGWttx~v~qF5$rkizSWXfgp}yjsWTxkJkx0FJME^=tq6bqG)i`E+wczQ@B_wtmn2
zQh!LY)cW=zo?b<JVbo4Kf?)3J682{p6apzX;`^!T{1I8u?_W^7=Nn#I!}0Xa*ZWGt
zU~4~!>t9@hP16b3CH7GZiFfYPLFd#Uwg?Wyv9(_7CiEjKURJq-dziU9^e*T*e3sT6
zDqY!zhvWzUy|-l#o^h?R$iFh1<YK>?MOG?g@Wz-<+ppjn;ei2=SbmsmDvs4vrgzY+
zgm)TIDt(?sZneRH{|0<wi>tUh(bX^M?wSI=CszjJUcCGAz|Z=&t4{fQ$u*$BTVD3O
zGMQgrMhiPM_Smr0_ayGY(RJC2Q@CR9$O8T*&xxnOd)_efj;(gMjz4j+Nc&M=l&>Vm
za76e^uae<QN9YYGD;=Zge`T6+cM8R9@@Z_Q8>e*x-~njmJ5Nt%ox42d1Cd0#Aj0XH
z2ic(b%fptqHfj6{QjH7&?aI4oqHZ##<HO$XRRM7JiXje_8^#Ps4wlM$Y6IkD74s~a
zk3Z5UR5jD};K1nS^<Ai9RAW(y{tf8Kl;r`<<7ALWyMPbsZ6_mRmdy!3>@=G_DLxYY
z@qZ89gDJ~yZrDx0_nBrlma94t_)C#L*bYsss&a!AhWS<HT=DcoPUBEX4`~@(9ITUU
z?MP;JXI)BrrHOB_d>}xBAxPWcx;bnn6Z2UNmijZQ9g-HqB;etaatw9+&2%D7zB<<@
zN7#aoR;`Ws`qqbq^75C^RG+{k&NIfD<~~ftl-LFvWWbC8CDFxUl>tBFv{DFJ@snXv
zuvf1$<5N+E4w3qjqMm;#U{V4e7QP_w!tjsT%EI-&4Li5f8%O&a`p+e|Ls{L{H^Q1?
zLu&@@KHW}VU3j4FLs}w;qOM3KjGGjyLa7l#x79tk#&oDmEFSGgNNJQsU(yA;=`%$N
zmn){pXmo~nIG%(;CV!!fS!B=@cdub4rwnrF8D6IE=_3ZEbX>G}M7ki@80L~cxelpN
z@0QG<&ZDq9?vr{x9^>8I-+Z-PK9oq|b|6gW0r<lwGDvhjuSC$P(@3Op+GT&c!WpXR
z55xL0OwVOI6i1$knXxw*(`!y_CT@7;)V6TG(BeXXHJTLXND(F#^{f}`=7<uCrrT(1
z7@;}`xqDZ+l3NMp!AN(R9?&rPSoDkyUP19P+i0PDEkzkqWnjGHFkc1fhSSZEj&v#m
z{{t^dL9N?!v1fhB^>P<kV2GY*G>12m_frhyI<!hlpTmuNaX7|{#$Q_t+y~isi4?Np
zD}3+{IseZNHFiqLIK3ypbdU|JS$n#9XLG+SAvetJSM}%rcS}27fQ;53_|#fO`^U#{
z36k#)t41A%9a`Na9GzG{KL3EDKlpdo$4MBk+Yj3rR|7HLJ~+?6Guj>{@&5G9K8?|K
zGMg;Xb+WAx6wrFam2y|27k-~LLAojiGQhnjRel;csRniDpe4qkp?Nk;=)&2db$Y#<
zzEZgS=w){y2!5EcNC=$WjjvLcZyvzf33U>pe>jaWv+MtAYaL32F}%v%hehb@_V_H4
z)_x$}epnjN%q>sEDcjhHatKbsSp&+bNB9ROV^C8Edd)T07E~<XrEVD1<46Ky?iydK
zO3>&<vF;Q_98Y{9J^``>DkOi@PxIH}4@)jDlo6@y@!Iv4yHy#IXB*vT7IAPV&+BhV
zVpj8Zd8^B>@jT^#(G-~x^)k#zCo`qU<Hh}+ewmqU88>ZK0J6z&lK&?pBB{4BT?OX8
z3RO6=q@cg9<AjSr#c?n**C7elrISsWYydJW)K*pUVB|b0py_0a3&DQ&!Ys2|4%b)`
zYs@n-UjfN-U9MHfJT0K3jooJ$icY5`j`$OcigDTTvfG*Q;u;^IKpg3Ch*fjOa@I=r
z%?kME{T&p`uE#m>yTj8f&|!J$Ks`@qa@ZnjxpaAEy$<x0Eji(H^R*l#EL18z+m$A%
zW!KXj;m?Q=H)YLN2kVT>c^|xHs=CR1VJ6kje1DqHXZURG8+^AUt(y)(Jf?<P!08P?
z0F^J-J#^aC9J_a$nOB_&xc_P>(<7>hsWkwa;`2DLmVa+igk5CYH_$)+J8_0Krq>qN
zQLGyazcXRR5H{2lNIR|k)KG3d-898I(cLRiBs{BKXSw*<a=OU>M|19gykfHS?EcQC
zk7|)VMDp!k+QUr!0JHPqpzly`3sUN|kii?VH}uQ142gD?J04$$S9&lai5%PBXl@tv
zT|SkBq9Q7#baLe5>-0}?cwV`w=*`@H-E*827K4Cd2IQ!av9Z&1o~hgN)MA~3xS2XZ
z$50Z9_2lqgp3MTJHjlU8>X}~Rz$o7Q;aSV+YXLKv<-T!TrmI;88VL$LqOWun$!Ncx
zVb5FkxQNAfS_Zsb6gs$VRUbL7o~H&xMtFYU)QJLJ4nrk7mI5H({@MtQ)v%)R;b?YP
z;bo?!k6=I8HgR49PG*0KU%w<{f0a^*Jt{CV=%{VBEND;rabYfeQdf7pjL2>($b2|e
z9>&{YQ1EL^DwWd~Gr$gLD$kmyAOV9+ln~&>4!QYyfl+8a(ebfnjHvZ4dO}Ta_|Lak
z%mw-MFR@Dvwn9b{$oV{OXG;Bt^>8RiuuDp(JY=MuE>Dq@9gLz^*i6||+f>S7XPvp@
zM9aV-pjna8d?szYhIGmI>b9~@`LV+|ajw-b6e=cDC<2`Wb=Q^gWMU20%R&-SnOOnC
zT|Q@PJ{HQRD|xi4t>P{gF!f@ow=?Btc653Tb=PDpu3MO7(d+ahnOG@1A3vew((`K+
z|7L-y^*Zw{)<U*<QWxN)Hjgm6f(Rt%<T&EU>6|ai4Zb-X?u1kEEksUlB_IO6z@SiO
zoB$5RyPEB@1DucVrIy-UTFX%07_~YAz|di<o)qQkcD5?9{yG`TYC1t)e=OjOnD8Q$
zQr#K-spC@>n%}l7i%+g%Lz$Mz`SdQmwCW%h{=-7qP%5>X<ciDY1|C!)9m~K^tKWuV
zBGa`VUNWt28-Vy>tb_C<43=D{%Ne{w9{nobVYPD0-RS4PkPCT>G$3nHwBsFY#<<L_
z6&~S`)}3t5{&vlzN-{Ry{=bs+{s-OT&9tw*+p2Z?{FgU3vsKXBYG3pswsX|1MK9D!
zU!C~E{dhii9c36b<0~BzU(ND894;sY&5d7CV<Jcdy6f6Qm7wXFiC)Ym#9Y~}V9zix
zW5hW5@3tmf4K04Ni8)Kb+!tw@e<{3Ww-{z$TKSuLSzuc3ZN@e;eQR$B$;(~mfi;UL
zF==6ue&~%_w1z29gn-RbK7`cGohCrc>#b^8Ey99KFb@+*_04?!^`~mRqnL#gk??JX
z;gx~lx?WI$o0az41iV4bZZfKL2>d+3+*`2scxH~N9R3twxj`j#XXv01`7O3#7H1+=
zY_Y9Yd8y8#$iOGv)hAV*amUo(TJ|QhxTZ=jsHO32g)wM=1~T=0O{-SYzgf9Ia*V+I
zxX55yoYxmam%9Oj8iH!N=Nnu8-}`0+ocJ<jb07UXPTH>EEMm?+bp|cL-Vnq*UmxFW
zKR+BOuM3?(Qew!5MSyHBoRP%vb@xZ#++F&&1W<T*1l;?L8&M4}tr;CQD~f53gETfH
zxenIY8|)B&jlw5S<JRA@z>=s`zjSlZ!X@^>i4(<6H=Y3Y3twz>ap#DJ_jSqj3thNE
z%x4eDw!{TZh)M&dfoR9wQ8KxcE<AKtkYiT1Ub7+ea5{bA{s@wE3j1%~+3YxI6v2rM
zfx<&Wei0`UYDMX|Qc*UuiCorv=|dYDn%x<Vv<3I`4gZy2N&`IPh%iO=YwDC^X87ui
zjQtTNj>#_KZYwQF`aDgR_=l>4KQDL2895K02|0g&N_(@ARI1q?@CzIw-+TI~Q0}V;
z{ym!zW))te8X>i8c@Soo9C=DGFz&BhxcX{0<7on;8d&GZ$lS4&4TwQ|0gOe9>VK2p
zBd6hl=CiaG`Fey!$<NN)W1*0Sid>za(vn`11(yzOR-_wf^q<OYH|3Ip&#)k8#@Xe}
zyR5{bi1|Ll+LKj%6|;Ya?mpTimni-<Ng862vhI2J2V_Yaz6LmFi6G%itALd-U@=q4
z_FjxGAx#SrS+)}HZBUP}WbJrploJkY!&Mlw3q<9S_)(%$S}@Ayx>FT>34#O-``+0Y
zrEuN2Ew%Vi&tl`hzQG~AWUO+Bs3y`VeGU-B4@81ij5t)=<q!&hQoetcPFhncGQKeE
zA=YH~NM#j+|BY97AXMDewsSLEI`8~={0(o7vhEHA&WR&y+3i%>-oA<}-d<;o;B{|v
zQz)X3DnxiJg&osW93*+-0((U^V>ee<{fE%;WUch;m|mT#E&m5#J_1Nt$>C6ljEh2>
z7tr8H07{{9p=)n)J6jv^x>l1blIWHs0^`wIi(Go^T^>>E?PSzgs??RFq}S7JI$`4<
z6pa~Z=?n%uBodB<NJ1X@lcl<x5;eLh_3m6TnM9dx-}y}LLGQn>A4!z+FaJQUSufV$
zjci48nX|<)>jqGL9h4q}&v%w2_->&I<x7(mU?}?ZnUJk82H8i2;9EIHL)`9{QH_!6
zY1YErhY}OLW+w5jRn!d^W<##SeV3?*!wheixgu~tAXvP(y6a93^y;1xj%n!FZ>=|h
z!$I#a-g`LQUJWI4Xwt27lyJ8g51}tzPbh2Oh^Qx`f;yk>lgIho<%CH1gTk^oJbKF+
z{+Rs2BInj6(M)S7*Q0D7u$uLO4Xe(@Ay%tY7qi?a;qulrZ^ntfT_k~Z`ZDPqGIo#u
znP_Q?R79wH3r)x3qxLeq+D}q9Aa{vM?I)Yn>ev+D^*<@>cG(oE?J4vKsi%vz9vH_H
zz3j-cL*CSZYw(nq!d@b**o8Y^WMxR}<|5E!5&WwTE*@_P9$+?Ij<5w#IpM=&3e9*e
zQpBZ11Pm02Aun$T-r|=|A1*;ke?%(4tR64?RuF-Y3bkQrz|ZR8=_w7t6L|dh6`fk1
zP#Fi3w$l>SOBgA<kky?8&bWNh1Ko*VNa4y@C9qfAUoCnM;~hUj5D1c1#?j^#lj>Ja
zIspK*Gca~@ut2%0izpx>ZtwHrR-XUlv0n8d31#aLJqS}ikt?ifNJ;OW!d!<Y8yNU7
zSLQai<AQCf-Eft@Ac&_(9;WGLIk=vn?!u;5FFf7gN!YKk?Ae=->tE#mq0$R&#7v?b
zoY;3erM2SyR|8s-1a=|wLFc=Ie`B0ZG$p{*K<S@qP)EN#*xYrgdvd=sn!*rs7L2wJ
z15cgr<M|>Bm{Cx;Fif2rbOj08K6K_UIWbL&bt-0O7T=Z+S(PBWJpb&vyFLmgP&FL3
zD0$ucmQuHu!IE#N-PAtt)lZwTQ#gn!E$<9luV8g+EjchfEjhrg^JPf|cB0;b$Mbq$
z`J7;l%Tl#m_%0|pwp0C81Zy}UK45`D3LW_C%XWV6GU5SHCZwqdroaO;01g-SvDS1d
zSzBo<{%EiogGRk9iQR4vvE|51E2YanR&zlQC8~%1DSguNO=IRJs>%$7jWJZ**ZpQG
zHshNLl~$WqqP(&o5*bI}$)tUV-Uk`UgComLO?O~IxhDbI;ibJekMAh#!FrZp^|We%
zA%}L8aq>!=Ais4J#jUj6<;310hl|VfjM7Sz!$I5d3@YJuz>Mo<ys$zBY+*h}XLef;
z;qc-%OmPLOdns*Z7hxW!Zrxj#(19eLspoS)p?nN)atmxeHpp5(lA8B2d=JlYn7pk`
zJ8vRWAj+`ALzGBH`>zP&Qk`jgEmb9@k5hkrwUfnW>Bb9N3fsVSF^0_Fg7IGOh<mdP
z^7|6dmk>M%aX}J@A9pKGU@z4|t#<=xk8J(NI)%LFZM_NQM)O9-dwxE4e$t3D*uPVJ
zW8!_V#r1y;UoTY3FN33rsv0?^i=(ypzBI{HDdtl{W#_)_Me>}yRNf|ScG_8hR?nuC
zie79nw;952d}Na=#{3H@#Ff5M$#FuHy*~S^iQK7B`P*f7ZzE&`$W4_ZjaGFc<5z)P
za+wxJg89A(JxPHR=`o{TML4^FO7?fH0~(yA2U7Exz0nxm1*J@Cn`QyPse(vwR1TNJ
z7OHFyZtjt={d%bafHp3esu5EdVa&n&!OOibUJRLp&s3FQ`MVZoU^2dWBC}5QaGv6d
zX6Etl(l<Gq0LAA<f06y`<l4Zo-`F-hlh~W|AMhVHBcv<{d?Y%j04fXqBw*_;_ej3b
z;n9B0Gg)kfBArTE$nO6U>6YxHFWZBexi}2|OGHrbqLvfRWTzoHM5Mjf9RUUOIZHZJ
zPq5I>g0cLB>mOuk0|(*5@pOX`A)SGy?Uyjr9qX_V0`d^ohd_mKj5JQoHtV2p30-g`
zqeX_?P0bvz?RslfxfeShhsEL~L{nbP{qgJ>`3c1Cg+a;dj!&Z=mB=j9Ix8C@Saa9)
zupv~Xg^s}I3|#A+PY!{%V*pJjS5{n&0{>Kk+{Z1#1yv>d=K1cv-z8O^4r!Q`Yw!YA
zoL*V<R>38@41pJC#f_=T`|CB%<}hpc<d0ya<k+)AxNEZKl$Sk{nQYDJs(LOi(01Vj
zyWz48-Q5%3qw5r<w$mx?OVc&HuE%q5M#rn9k=t6n7xNm>=FLr2!~DqZ>Y|VefmzA_
zf?MwX=`&K^xRqTZ3Y^gMLx`RE_jU$VJ_%*ZA6)tD`PEz37@k&r&l5dHzHt;kWK(9k
zQt8~O3zS?~Ri@pm&4+#u+`9H3UY>+e<u}*9)eWjf&c>(Ge2{y!jHj0p@?D&!G^q&P
z9T~EUu;gC8s_N}DekHUprIP6AiDUJ_rrp<Vu&=1|s94C-sGxavn$OJhNDiH~RLO}u
z2>hKq#_qbK=wbQc_A&cmy+xP$09T!};Bn%+mT@l#Qnn>ak1=o&FKCkguTf+)@rf0&
z3qa|UMZs{PLSiKLSgIJ<gyb2T&?&e79m_i>!1Y5MFfL){()~cEQBbE8(#3jl1Srat
z&JVr5yX<$~TcSK<l;3=Aw4q+@@JjS_Ju|#SA%F~KH#!tOAZ_-IJ);3kDPQj#sA#uZ
zdi;tC8cFmq2ZG*qk|K!&$ih&FIsa(aSybNbIh~|ji&Qy`Nw;{OaAPt22*wB8k#~1r
zR&{?gZnlX1E2F|AA|6gPns_VVk}Hih4E#m*6R{O5`%b69I(t-Vo2$=*t1@D)nlmwx
z+C%{@k<(_$SXYR=`M*IUt~sU@{(<})>wk12z25%f*<s?L_uXnP)i9lZx1V?~055t!
zd4}C?C=&K2G7Gt;_?05#F0Lg`16yBH?WWptql$7+0}FqVr277FOTYaN5P}fuz7~g_
zh+6TLVqS*^E=OhVhgTJ@yY{P1DWH@j)6)`R2HXkh*be<Dxxzo<q^d6#=zVU?=*(UC
zlB2X)m)$T~OBF`{c%FRfyDF%d*nDwJ6hBy0i&W8PhFyr@odHGLDBtQ?c%BtlojY{^
zEab2ND(c&mOz(#6shp>N2oTXyY#b*#P?<!q5YXdq17803Spc*udG<mrt{3d--&~{h
z(Lo0<sWX!1K0n)PO=BKL5XP*7^p9SLn`52LCy=>aj0ILZTw|6RZ5U&-4<cKL*ILBa
zXNvPN;hIzEON|3gypjT>VyWuUDWQkI`izIHBr+(;-Xh%4UO#Wz;IDL6wzYfTLdoIs
zC`<h@wXOI8+fR(Y+XZpEE>s2cc0Q-ZU1Nun&hIH0iDK$37;JP%ordtVdK+(X__l>$
z<8huZ(#m<bZ{KT=&69Oj>AXGstr-v17UCyj6#(FVCV$!PoG_lw#V|%{U?fVm;aB}P
z3@3^zf`GBN>FYDSPD>nva)GMUH<(D075axJhp`gWfhPOEZH5QJo{ndqPL>}sH*WBj
zs!cw2$Emm*AEY5w63a|~t<aTosrg$?Q9A44Zc~TPT9U2?dp%#?(2*qs2cnOm=ew%g
zDuqaWNI=tvQhjetWc<fgmOLJCvkzVk@a>_zf3zGi>~iYn+k+f&y8$4b(yT`m6K^H6
znHAY{@1-tzJU)Ib=e`Q`RF#xyrz=R*9sUfp7kjs+;ika6qO$W+U&Ca_9s{`vaA!Z7
z6v3~4iJY=t>){>mu!X*vwuSfiHH=CU2I4dB!<y0yz6z1Jpq0EC<hBTE`#iHmo@mrp
zn^am3=jd-bU+m8mtElElqJ>?i4hJHkiWJrtTx#Q^Jne^mSpLI`Ixbvf8?Km`h_`%I
zFx;bcRE<%XN1Lgub21O&nEUGFfa;FJY?}6UGXf4|yMfDD&%y(R!PS5&kJ=YWw49Jr
zBbzXqY|Uh)$znR16c|CQ0u?Je{K>DFH~1o{HmP+yONBa833!=J0MM=<vdXoQ#Fod&
z#6IJ-)iF?`1(lAMVURimnPsSZt5qlce7|tc4@_sL&8J`9Mf;P0UdZa@|70WTf3wl&
z;m><ELOr68@KaPaW!h+cJf{0N=4#*&GHfD6sI2f%_InKeTE;{k_s>_KEiUmOBr?2?
zPbJ17w*ZD-bB|0?+i;ePvp}R5{f{tBUpnCL45SC5*I)qI8w!t*qB~y`=BvucPusVA
z-)C^GrNM$L$xGv|ZVNr|AZBghqrkCSG!d8>L;S&>9JGvbn}PA3-}a{3>~{(W0MZ_w
zodBY)G!Bzn$)nRpK40NtbayLV#)~C$g1ln%z5<CJvfB1OZ%W7f_XDqY`(f`P-j$BF
z;WEKD&j%{?x}4RrLYpADdv};%KxI6FoPfSdLtt2+edB1#_0~|xdb>0s6{Gg?uj7Y`
zZ1#HK*HkAIkS~JHGS2YSMF85jX=Zx_t&D#jPWVi`IiV~~mbUze5Gx?mv!l?q!0j&}
z`=K9G*wcKC0R6IeyfY=5L>}>TKwv$oC*FO}X#p07n5)K)A=YRT_wREvN&D**$)tFx
zM5c0sKWlu1!#~=bgTOwW>Ez*vlKUPs)SilLOLnPo_A8dMK+%KvK6I;OjzZ<3EMl*I
zO1sbbfzMEa`mJle0ZTpJu;UNA*B;#jZ~1+R;TIS%qIG@vAL&K%?Lcnb{U1al`yWJe
zi#t1%8Z(&NQVCZSYJNQ@a!X{28Na>0;HiX-JQFlKTV@Xba{MK2E^^gnf<=3+SN=R9
z^5JIM(Fdzf&l6L$pOqroLgK6~uWV>rnSq>(?5#PZvftt-vvMig{I;8(($QA&^`Xn0
z9E0RFw89Cc+wA1mAemSYW8T<XF~&gh@aJux%JnStT{@}d)47zNU(~-qP5wI%oVwb3
zRh3S{O<^(3%di0$?V`>%dnsr?OlV);wPH}o=<8=WcZ8|HI+?x_Q~B(jwyQQ8sA%8a
z>8X~A*RveNaP%Z!aSHLG7_3Qn<$*N^O{F``-Qh&<DXCJF>1Gu8tCDk^j_EI~C>)gg
zs?)@Ooo7zp|A<Y?f5m21ucMTDGK!{_ZZT+%H*lvGMIdjdheF<(jDh5N-9~5bbFE2e
zfr4PCs;n8i!XI}}#X;U#CyH#*9?Xf!@<t287Y65nH@+92TiSa1Vy6tx?L{5@>J#L8
z`uLP#Qz=n7_<}4(jc1GNNS>J-O*(DwXEf-x7s>dVpcHi0rKXSM^&7>lmiwszKK*~7
z<G8BFiz|zHe}#Np|LN86{wT>YlUgpDodvl6J4f0rnaPtei<FXRE}(lfJMm26vm^6x
zcgG!@t{$0i-UcxSTxp4TGRXF`KUH)1Jl`uh3BU#0e0y3;zM7j>5uP*bvdXRh6$F|O
zpw^y2Tgsh8z2d+`uo8*f`d$CB@f_~g`gx;h9k<e}VFH)YJx7e_uXvX?M{uZ#1Nc8o
z%5@S)lzes{jEG2Pn(H@yW4jbR<pm~J!lTdW5%tHOdu|oEdN2isv>iBuHx6Q$bViXY
zWLcLpGO`OLQ@F`nK=l)Y>c>pk0dS7Vx6kPoB9~SuJ#&32l0VJsQ94!OW5Whf)o0HV
z8F^gg|5muKjpGeWPVX1k$-^G;E;<~&s?k(42<ddAOlMo`a_g?!*8VU%7AspYJD42j
z%aobmD{$^-N_3ms^d%Qqm6BYnuRXF{S7(<10tOT&v5(h?*&Z|<55a8S_;vSMfKOH&
zla03J>Ltv9zU%RzK^(IkaSG?Q0-bsttjp;_WJ;sCh{;G?0nluGI&#&EMr%56rgDP$
z+GacWdjG;mYvL?+K;;Kg;3Ec=LT)DavqD4I#rr&uPPJUj-fuZi>0_=)hszVEkiK57
zHr5j0+&+od3B;VUO|76ZPPo{cXJA|RtOrA<)kb5p+QBC4X&eRC!|&&W98?mF>1~`#
z<6r(imcWTgyz2$ce#S0*o^N?C5}S7_2ib^RzXM;SzlKZVqOWGId}Oz%<MO#W66J73
z`&=}Pd&Wn(XH)|k?x-?yOUj+kLA8|^r$=PXxdptjFt&M1NP$YySwVI6*Z%0jDN)g1
zn1I3P?(5r)&xcCgP1vpN5vg=25;X&79hkQ#d^A7GaC?6EI9rqyOa;g|*8?1AH@>RP
z;a%*n?+)0}xb!Ua6x$x#r+U*uz03#Q`q~=s52Cr&_nvgkBeM1V6|OT}2Nq2kF)$jk
zXGTS#%vS0}RiyD5u=CZ334iD?KX9bw1>7Dtv#8aIC&@erw>H@R8Fz5PqQN(GgxR#}
zn1`csu~e<33w&XtCEb9A=53MrYPVVt&_9yI-kker5#&FpIC;Ex_iNf@Ov*}~DfiRO
z;qdchhC=K&9;DOt#9+i@oC0KP*|<vW*wC=Ckn02B)FhhG{*6;-v^Ng3Zh<mZ$u#KZ
zO)>NJCyhoVVVxq^l$!Z*A1ymz({QQElX90{`44A>mY9p>tl}T?7Y@MGbe&;7u1<$H
zGKtKz=*5BJ-b&n{h+WjosjG)g9a2wY&zY3}(B8hwB+oM4y>{rsq4vABTh8Zg^o0AR
zl6M_tcvbw{Wl12<5IFq?_{W*)%t2=$oPK*B(>%{_$Y1L69biC3XNv4T{gpWL(>&e?
zU-ZyOSvt_6Y@aNl#BbZh-jzAQ+wAd1#CaL93o-@tG<_+y=T0B<f$1`Fh~L;toXBJx
zx1-Bt<4Bjg5h`_u9l~MK694#g6BFP#lggO-6^Vpn&KSePaW98&`Rj-Z{G@56dMm$f
zhof3j_iv+`(L4)au|lViI53+La4_o7&c06Yb;l<Zb4X;Xy-oT3`SJ$`wPs`vhh1G{
z<oo?$f;~e40;Usc)Jlqqs6N#4IL^;{1ax{0;Xm+g-y4y=;*l|_xEI?;&+36gm_y`g
z-|Pes0C;3Px%gXu!5E|;fV1%%%mDR>LT3N)q)G19_i(`(QLuj<xhGzn`ws`yC@p4&
z?M0d3UbbZQ{_fOcNM#Og{(Q-o2-pr@Veutd^xbAy=NOl8`}1yOvpK>6XhP(cW(Awe
zw!p?PW);M#sGfr@;Pv0d&pGtAz=9dbfzw~<*IL%_NkGoVTF*l4&ch>_On#r7l6d;y
zI-PQ9%38=5M-%CwB*NjiFS!hNE9;rl1yc{ou*H&CMY_4%9{IAxp>Rl4(n&1WAgle=
z);m^ALc+2{Y&KI_=@?$s!AgoA(Q#D=?TxMq8*!Vxw`Uo>WS5<)s#JI`gL9veH>C;N
z^kU~BbmzMa2Z<?ZotT@$QIAp`9-^O^V5_zrsy1<s8Gcj0e^>>8=2&CBn&n~w8P;!=
zHgj@Cp#3_8J--^QH}zBhcal@41U75f(+TyxM*${WVh!K5FJ==Q{^{HiYKWIMXKmA4
zI`LmLR@^52%v65TRjbbsB)PyJ1gOpG5C^6|)Z1K5gRL1&LX}U1E{4_6odv1~eCwSB
zdh%RaL&NxBh92wcxKI|dTXi-G?!9tOjG%c3y|Ju*McyX<1o@9ao9q3Av61iJoKRmG
z^5gV4uIyGjM6ibA<NK>F&bRszdf*^L0uD%EpS<n}Xkt%S0fKhx9r2uZ22X(Qt^7EC
zUG5U4#%$S6=fx1?)gG!hH#q_sve%3k7~|JF66aL7P9#F%*1OYOscmrrP8`~^^lx@V
z&9;`+oGC^dAsZ9UyAG%x)OVx`o1J~Sk_Z{*N+Gno$(04sUr5d*Gm`9*Rq*_a{QbxQ
zMs25yoFR4ZbuH<T?`js46|_w>gmJ~FlX*D!uBaI#G}+J!bw$0ZXaDc^R*Nwwm)}{y
zMUTDQJx7hNz{`?X456de`zu33FgR=deMO0e6y1O#3zLoUEoHa)hIr_cMci7}<-)O!
zXF-0}Gm$)XcaI2{UJw4P)DbeeBl-EB9xf?!A$HKyJMDKv60lPnU$Yz=`etf0=_~*Q
zyh&~|QR->F!U{hCdBTRdxjw}+s{B@lrBA^vbu7)~dqJNHyvm6EKEfN1W;vOApPwro
ze87<-6!4I~N-%x?7V-(1$e?bGaIyVh!Jp65VzI3j9)sWE;5jIa?BWT9_>X}RvRkLk
z^e=eO1j(PfodY)Kvsa&DhfGd6?L%Oxgd{#><LOFv-*5DxYi|#yqx??Kmr!}wf#z>D
zW^64USsF?dEe@bImzghO94ohF5r>|Ly>}niRZhO)n7gE|sM4u??lmr4)AK?X9Z}%R
zW=zaP08m!v^*WSA*vHu-S-2R-KUg9i7F?xxfs&c9IK}8Yg3FVQ{<)|aocYjjHWd=K
zFv^9UJ_ni;J)94-)0SW~w(M1KVeGVNks@H?&!v9}EzM<wx9Zmj!v~ULOpqW6VydOJ
z-G%-=Zj#aHf3gjxa41r*$eSr|yXiovPXT)fn|t`68MxN7sc|caYso=h{2dwj^pS#m
zS>aE{NEZDXR!gdvm){kmCqajF6TY-4zdyttT4J8W7&L`?Yv7?+pJ^m}MEVxTF^M*W
z)tz^^q%t?lGOzn6m&9n?mzO7y|Nima{p;VIR!>MD$tUvR9eMZkDA|LLoa;F*dznhL
z!nAp$uV=LiMYRjoe)h!pYwOJIpL=^PO>#8O1g5~q83bG<eXO-`P4ViB3btt#A(AO9
zK*z(F?o7;3N$sqVHqarXKe-%PP5rg>$4M!8cRXGNzJeNo;BGQB3WFN?=3dP#LwtlO
zkx45p81a70Jbvg8B@)Kj!GTl+!A!1}6OQ6;C>D)VLAdafZ}c&J;7itfzX;J##4x(*
zZM|SO%G2ffb%4PF2QWIA&Dvp|{KxYE28D#|tW1lBz1>-ooAC1Wn^{L0Fg<N67;zE~
zVs+^tFnFa2u1&H*$0J^u>X#f>&feW+`*ryi38w7FQnNbC9WDu@CNCQ`XgxMIJsTmp
z&e4z62F9n!g_+CG>FVAb>+prFkV*t#_fxm|cEQ7(e{Q>ZA#r^?Cb4M-lY@UGRSKqj
zOTJr>v^Dy=nT!H|gGKD88_RwTC!ZPCh`p#6^EJ=cgcQaWSux;1;g@ye?AF(_ynuC{
zY}kcX$4{mBla{Y6dgTGD#-@Y54SGmf$ii@JrCb-hYBaXD+5lbcOct#{$pG*#QFduk
z0z*^KU=Cu%!|}kVVGWEw4D%YWgMit5yMNF5Wl<*pvQc=`=cohwa_!|SWN{(tMxivW
z4y|}{tXAV@NWgtMKJmmf#=gw&P9)~EW$$N;pve+mGS~F#<&{S1PbK^1K$_0yeWWoo
z;lrO_Khh?fG0@Ql&b*YEMl%aG_Wvv^MRjhZ0f`+k^X3vjrs02!RSwJf!7j<wf8%kf
z&iyxiX)^T>g3Xv#2r+t{6b8JFoZoK^qn#5~Dmhq|!>-R4=ty#IfQ!$%6h4sn6oJAP
z7)PTQS&ZgAc6V^3QJPcz$TF<ez&{y=n&!%4tC0q{DGw)o0g(;?_Llm~_p!=XnBh+p
zP#PmTJ2~z>yr{5CMJyf=+CHbWo9@FK_{lpy4-CJ}OS^DLpyuGOyA=$cf6N(8037|@
zVk01uinXeZA?i;Sf;wKZu_z^@1>vW07<J`|9r!3@FGDV|SX^+GhT9$027DMlGY27H
zQ9NsjuocMX9euS^exX_)GPD2|WS_iM6yknyzA_%D;M-Cavb8Fyqp38Mmul8lr9n4_
zjGFOwG){D}D!gG_XmMr2A&K$o<5uBT5!QA+{L&wamhMZJ9?Tya%U+B8@H?VM-!e%+
zY!H9O@&T)wiaOg!xXXU$R;F1Vr8rs2TNT%B&BpeGectQ{4VKXX)NR#r@l146nGj<M
zA3-_UapYfj6WTco=~L$E_!3*c@Jezu`xt8wY`(o{cbaULw7${~4x0H-g<l|*h``}{
ztarx}a<2PxFQvV5Q~Q0U29}$(H6(WlAI`Pt^%jYKuXx5hEIk{G<pfq<z;UL)0#v}c
zukuGb={q%I2c<!&tWg*_xwa7<Y3(r*Q0a?zkiFe(av^)6mTO5YPqHyt8Jl1iN^L>4
z5qD-~!Z^7Le+I(s1oXy1%7kwj?p#ds*-H`4I^Wryw(XXP$a-|o(_FtP6D+HhRXH+l
zL(qEhW?23%eYk;*-r2|XPckQ0#u;CNNCnz4BO6~wBnW9?sSul9O01l<VIO(m3{-5$
z^8B;Rn^6V*?RnvtTRUx=go*>zO*2$8>NXb0J;EVS>d)W{^F(@h$L?3%One!V5X2q<
z!d<|UrbNu!s0CjMYpVUj*yN5^m`wp!8Q?En)TFZRzEYk$$aeG>$eTs?Sow7vo3L&4
zd{*i_JQ$M6r7MNbgt^W^Xn8-g<C4tSoun^w;twK3$!q8YI9ZxK%pW4vj*rw8;jfg|
zmB_yHFuV@Kb@S~9b+P3xU@Fl>y2JKinEvw4JfsRhKN~G&wKZ!vC~DYn<`i7WU$61y
zfaY&jfQUL(K1vW61Uc?aFSd2Q_hi$#<o}ZK8XONspBj%w`}0G$XW`G5NXGKPCer3t
zB9Z)A92OMbWVkzDOW*N2U`Xe2NIHG|QIPhnCS=C(X-RCU?)EI&%O%SzQU3a|Ig?wO
z*<1gPQ=98fp`B~@?+AT9p-MUaY#dEvQ_Evpz1c~hJj<L=nNEByWp)-yX3>$hUNb0u
zG|um`@V5dRGH=#L;Usu&*F7vG%+x%0C>O@W;I&`R>u@kT3O{_J4-Vf(%hN%RWfdZD
z^t>#tYL_(g-o<>hza?a=A)ce<SA3)ZO+d20N+<a9KN`MNZV*+Q`o+C!d2)m(xlf3?
z?O3@_@X9#sdcQyXqy_f81t;Na?vZKs8#gGC+P1eqJQ5dVrOEH}e4Whwwk;EZke5}X
z*SXpQ+QKF6x{?48e~4<@YUR&YR91yOYq#U~PN4FP=hEn8)w}uTEJH)?uZ)lktSWCT
zZoMbwxjRR4!2u<dqk)d*@xPtAGX?Idr~+<VbG>oCSH#mNdl%|F8db2!qMYg66=;%g
z^@WQAZ`TV+g7UEkJD4Ba;Z&l^(LC`Vm2=gIA?w<>&2u}S?zNH<rKI|UZuzW(BUU{$
zU_<$pl_!{$xsW1WAcc<zp^%hdb6z!AT_nl*=nHaxU;$<Po9_m<cj8I088PsWvr2`X
zUM|VgZ@EviX*#kzkR3$3e=fy83#0n!$7Xq;F;mF&*wTC`jpZp`V4Q`sD1{gDY+j$3
z#-8GIRft0HS1R2qHvSejMt2xyA<9G2!)T3<OU@M&X#Ydb9)xh>9mMoIA~55q0s+F;
z`M1)vdo?Tob!CqPAZ*U`hqrw4l%BY`o!6ohwWoVU^f1uDywJ{%w<L8?q%gmH#%f5-
zlDhS`X*=b1Jp~H;HJH)F9SjGu(L`XRkxq#_U#aR3vdhol<0ze|H1<E;b;CnBTNPqK
zi>foujjNSAA_x#(^ev?drhqIN=h)4gw$0}wOR0Cn3u?4}#R}A>n8rGF6mY?*SuWva
zD_sU2?o3(4X5DR;r{AM~sb8ZGFTLZ6p)M)Y;<=bcK*}fEGHL35mO+>0eU7VF64$&Z
zLB_Up;QsmW7<0V)9!3aSbhZBg(@tBspsENM*`DV4827^8?T8vdN93}L0$n*$QG$>+
zAXts2Htr5=N2YI&#|EwjULRt0XqbdWNKzgbPy<+;uS@_#_%utJBLAvh`<0plMyPU?
zIK+>NZ+EgwGZ|}KsT|f3iYniEagH+))W7hgjV`_ZRZ88>3kvT^-)Y?>!&z%&f_gmr
z*@-I8CkR5MbWAdSYgHqrqb+096uD2A_HQbQ&VUlPRMNI^ENnrI6Z!B~uGvRtEGHTt
z*gzA=TJ;k#Wr~8}g`Azy!0fz$%l3nuQqrnYtl=#Lkt{xQK@s`&AcE}43bq6RmUJpj
zBqnH^_(8@@o}%vYh9mauQ(K}ujD9V;b}{@W1%d3zM$yN*x9k&uur-V7c!nr<zQ1}Y
znc7f>Cbt=T_<&D9Q$f&iRVma!V=an69J10}{TP8`8?wnZf)fvrz$CU1id5B#Ql8`Z
zL|c(t)4c8+9V^pYJE0fq7riW`iKKOV1m!vKT=LUW<RQDwmb*cwFR5C&nnFq~tyUW@
zqb>_|NpH_~<3})^S?EIFR;0uAKYlder%7I*fTFdcp5sbF|EaH+t-R$=MgsZyv(KL<
zL%H7~!v|!P3hq|a)}6^ZXYro)^{yvvq^>0)cAauT!2Uw@9yBISD+j~}oa(T4^|H*l
zXw8S6Pt_bI;kzashZ)<J)@8e;%5yVp3Q~<bC0Kfl7GYN`kV@|Z#X6n4x+IZ*D&o^a
ziv*3<X1npr+vP7e2H?71!nrPehxT*=wuk%cf&f5&4F}}gTG=j-akr8?uoY)L=+zv*
z{_t;T_&jU7E#VE9_XYt(VNW07Ll#+b5uL{8xc<(DRyI6?JeU>)aICvmRq@I}0g;tO
zV~#Z%;{FLMN*x5m0R<f{SMc;uop3UoEX^P3+3nMG9S<!mg8o|Kg`hoDWC#kh$>Yo1
zp#~h>yp6~J;0PWU#{830V}x^G+$VabiG4h>u~6ukN2$6hI#@_Zg3)0&V^erLml8Av
zXboVWXQ3jdQ4BX#fnk-VE~qdnnD88Et((nl&X6n&O%KpbJb^xJ3-@n6Zy!^ZSm-$z
zlaz;;Dp8YYh1uGn*uI+Ao`jZOvI~E8SSblRcSO$S-G3HI=aWm*EMk^t+dv6uJm}#K
zJ_C{tQ*&auzN!kIu8D+o*0I2_+pPOc!Jx1)a&y8l6Pz0BwQElG)VU2{dp_MV_J?AH
zW!0a_KAvxWOz*UpN&*gvUu=i0iaMI=tu!-^YVC?GVhvcFEm@3y(s|}8*sIxGdQeId
z?UgA^>CxdCyZKf;DXkP+Z?)%IZ54qb1H|`G!LIf~g2zTdv<??;iW*xP6a=vZL?xVO
z{676Vf1cqn$FE1;OSr)gW_qOeiAL=9sD@96a8eX-rMGc41`efP;+s4pXBr^Jmo7E9
zWN-4}ZHgJo<;?wc|B5((a8NdDAC|30D^QQm{DAq}#e!|?@FU~5+sKEKD7(wbbpb`}
zemAof&8Z@(ZcODY6ccqTsK-Z+)>onl==r@EM!velN^{`qPAkYB@k|GY%HzBcm2g{q
zG(|HdLu2c>IVIs$Q<FXtc?IVO0Dyt4e*u&2d?PR+!}DPOh)+tas>8FJczdd}YrBqv
zJ!j2JKzQJbEu68hH@4mN>Kz&P*kjxO$>#zQ&LmS21L`d{P#;#_{1h(Yt$%BCQLN3o
z9yKKkwX0DRBy!akmXa+9$ew&9p^wzS&n?>$Y7<)+V5eA6547_eBf3P3oG*rpWaa$s
zE??-&h>l*)3rjkhLeBmu04t=UQH}P&lSv*=(CLP!`iMX&Fui@EXR)9)CM4vw&0+5T
zbfsN_a#$^Ab@??Pes`F=R4e!7copCm9EC4|!=#(D^36$E&OqX=@YaUbVQW%_RUAP)
zKyp-TR9A!2tTLS|Jr6Z9WJ_dbsUG+}vo&6!P?foIt|HyvkQ{r&(=h&kxU!PE9Wo$*
zJVJ1tP(0F|s)bvx*9&wYhu>iUh3hZ*`$MIdK_~QQN^&IbG4I(utM;}b+9<_OD*J8s
zb6zx<$f2f^q{lfSv!|+;VzcxUBrgviYkN^aC}kNj=m^UXStNmcyT4SIYE5jx^ze}K
zTKdyyuy+FmJ&=Va$iF5w+&TBnooRmp+RdQ6_5mLcyq@j^YmK%cY8te;_HlfINn+~z
zr(teT-5uC@)RO}hpI@~^ro`~f>AIdpY&ID-gf?JkqsKE?iG@C5LTxHpk|k+5JXYJg
z-b-^?U&oCbm0L&*JzcI+<F}k}*5h#QX&^iqPm8y+e9SJu+{!NrpR^tj!N7mGPo-i%
z|N6sQ94Sbgclr%I(wR)~Q42PkR5ZXN<MTnUrObBIAal>(&M(r2_Szpj^exTKkTz-5
zk|0|iDI{NMO={s&pX!5NUKc}N@_MDOcV}KS*mb!4`@}`&b{s6;kaE8Czz8{9IKoEE
zXTPl`y!6ZD)l<~RLIaX;A&g*GS^=4szQAA7nDo*rRN?uF%z8?d{4e{@4+wRJ2f{{T
zT;ouXE%iXroZsf)Wwja?$UW?F_5HNz>~$U4>bR<MhMk)VuJO!kzO2n_1jxXpdW?bv
z5Q&fI$H3&e%)Xd9jh^qQkyO0~p04x?GzPP=N0|nNWo>31hw^pB2DJ&5n|xO@E+IH`
zp_QBUy$9e@FaF<BAN5b?(YHhRqlVm7KhN;cQ{@f+^y~qs2VCmE6KEG)JXZr=^DwC8
zG~n4<xqdIo3}R!$0z01qNA*ocR@7^yOCNf@>oY6W0uM~N`qC*zY5LJ#yZGVyfdVLe
z(T?fge`FK*C9r7yQyvV6|M5B^vM0^+mQ4Hip-T+ZW+OD;b-~K+nBaXh&U+v2h>*<g
zjW1mNd>Grv-Xvb+C(y<j=X|aL`z=n#V-{n7+0dxCrpirEDG7K`eH#4y^8uIZZ#flq
zqk=FGws1-$Krc4J097DSfqrbHOWo1wCk}bOSxv5<opEM7l!F|Z^`YGrTy#L@LJx!u
z-{vHmsg2-awp>wS{8HFz^mwU{34Yo%a%Ft7V<^9$KG1rlqIFso_0*uT+30{B_7F{W
zV`D-&fRm4tRo@pDVEL#wq{?1qO0@|~-Hf>;VvjP+;A&m)^)`VsULV{1Eke+;Wxs~~
zYYdf=wT0C>u!|w4<%fzvar$u&{&2zL$>4p&n{D<(Si)pBxdg2YNS57jkhDapJk<K~
zIVg;M_fEike$6M2RScd^uYo;+m_KmqZ-k5!4nvL&kE>LKXQ0tQq+I9qzBq}1Z=RG&
zp^U-(-B`*bO$>=@r30Iiof^0OIqC@pAeqf1Pik9Sr5L^6q-7KEevLVFSmHFWa6&Z0
zkll9_b~u{9eblsKBKStrsncqL|MB_&!)CE6g_qBTnb2z94s+=-vAj6=>K|TpQCz5X
z-NGdxZUDg&#2+$f0(gGc<5K(C)0wNLj~dnjV{LzLXOWt)&95xMF@9D#YB$g;)8)Bf
zIoAEgSI+wqXgStFTKcp7HCe6(gAk-GvLHRT<^1W2<JvtdptIha@pDEwD8oa#J>h%(
zRw<>81F<-1$@@JIFGm@kr!*7lyCv^;n$5Nhc+VzOCY+CdCoaFYtUzJ}{zD(r@Vk72
z3<>ZC$fI9{Od;1*O=>A>Qq=xy$58<V=XejXRg-=gIHWWRV>Sst4dJ!oIl!jA+(Kn&
z#HstWIZOX{>E)<K#UZ;)kM`LA!^B&MMcIB)qk;$`BGS?*-O>nBBHbb&t#o(8FcQ)s
z(%s!1!^}t{-3>!GGxSiy9DTpvch33yiR-?ez4zK{J?p;r6B?7hG}SyVx{GET)b~^C
zT(HAwg|LZ#%HUge&F(GL>t_f_5=BLa$PzxCrUj9Ixq5N-kvRS}Z!iu)er|A7Pj#Kz
z&(akBOsXv0Fyv=ehoF~!r%Hr{ojdfj=DAiil94pS)?H<ls{uj3R<<s6*QqCVU0yoa
z=IFw0zX}UX%rV&YJ?j)8$Q=AYOh=g=dffSqQDu`}mG}5-C9Ql<C*ccfjjn1Bx%J#h
z=T0QSEo-bx0KJH2ZBj7;%Jcq{Knbr||APJx&fU#Tnajhu0uC$nd2RKN<N>D}Pl|@p
zqab#RsgY!CDjaBI)qm%6`n_tD13lqEy@d(OYR72#8rV~utb)GD5MG9|zxr6fdtQN3
z(D6-?T2^GheKkcD?1IM#>Y|lwchl{)h1}z<TB?m`b~0AiX4fucEaQ9n8m;{Ad=|-u
z%opI{70nsgJJc8)p!)8d9GAC^-)7PMV9MzwK&vEq`E4pc@HN=~QfULffw%ZJlT~w`
zTxw?Q-Sg#x7(PY`AYA<c10!)ntb!zfR>VD(eXdm52zVf+#j+@D)g?fSLHM_pib*+i
zKSBI&s8@YAu(I7-?XrI|G|md_MK|&9?rl!$`BZEecRB8gwaGm9gBP=Q{Og{oGd8O4
z)r{Bwgy3U!uo@`8El37#=soi-5j(SW@L`fX*Ou;j_)mH3k4M%dH}?9x3M`HM+dr#&
zeX9QFZs`?l+!WAEc7CYa$@d=nv7l^_#c|&LHIt6;289W^mB7pY(iLW4`7L>2*(>>U
z_0V6^_7r7W9^iI*PbC}{dp^~88`I!F=hK1rS4}7VRFn~lxqO|M<mzle(zgYgRh&*r
zk2d^9{uJ0Z@K&O2yuKxjKJ~M}@yN_$vAfj|^yZPyTUD9(ppD{#24-5Sr#nJ~zxtw>
zLQq%h=BmK_533GhUdr;VRR#}I7Ro94J20VQ44F{nZ~l>6f7tAjhd-G?t>Zrs#Rh-)
z4G!3Ze#)&`YcM;<9!%oLNn$hlMJeE*W+d8NFu@Qy$?^I-tno%tl%?_h^*B!F4P5D4
zJN9{#G+1;i3q;Rl*q)g#P{}iON=LmFODPsl8Py6+;YmLAu$xQ7L^qiRKk|H7@0k0N
zga71quun>1yIg9NdSQPuZL|DUzWJYDgV68a$+wbTS4_{!`>lZP3sd5psH48~&#C!~
zY&CpMc>7<-nxvHLq|0QD2(`oBMK=a8%V9%8k;74}5-y;BbJb`tJK#29MI19d{vJ$@
zA0d#F`-xqO#Otq~uj`4=wR3{kDw>!*FYLlQyxx<);Xw>vE~oj&JUcUDO(A#wy;zeb
z1zR>BXUE6%?v04n@o*K_#0hi?$)%>e(zcr^IMC+M8abG;UULGlYP0j;cGdqlw>~|C
zKBYXH^RFQmm>o*tBF21)wfAx-xEAA4xhA;0gIVF{_l@2t`I-<K-yiIWTpT}cTE`N3
z89I+fFAuvEgCZSQ+f@St124AzXx-kVNW5m_HSNi@oAbB%3YD<Dqy8j?Wjj}@X3TbZ
z?8m+zLB#wy=vy@|QBO$`-tHu?i~hIuU)z(0<GoY;U0a*s15uCEFndYm{nD24F#hJ}
zX@Y$B)3=!G48ISJuV2abQ4h}z3<%Ec^N-QZt^Y_C{|2#*u-vMS+WoM#lT)_(HxWtr
zlKaK9bI~7D(|p)7#}>;0?;0(chvdoBMmM{r?#0%B3wcKGO;(QBtY?xX%868-l8xFK
zC5_?|OTT@M)wPc)!_Qv5=%l+xjs?f@bSS#POL^IN3A(&bx_$9VrNy6HV>p$U)hh5!
zea1}1=i-B|YKh^MQntL8i&VerM3~giNd6?|b_86X#IHi<gfWQdhaa(B++I>KDR4p1
zv1dCY48pi7m+yRBQMYvqx%MI<pUs(_4=(g<Y;jnK(gkf8(Hpq<C|+XbDbkz@of(T_
z3bn~c7_buQEef$v!yNlExN}@rGYOs+w&mco9TW9zIfACmQ{uTl&+wD|BUJaqJL+pW
zPgBURw_Ljc{xL`AzwXAKiV*$(c6iaWk(D0q{J%TAw9%ei_~-xI(x(i>C_Z1{fO(Wl
z(Pgi!W9@u^nHI}$6(%~`)ezfjb-w1JCKWJo{V|D0g&uiURG8g$ey_UD>N-EW<bKsj
zb;}w3Zqu4th@?J%^r|~xqAzM}KYqTDHR~PW%wSp~<<^{yLwD>=M?pa22N5s-H}_9S
zh%A;;lvKAHn47KLStU`wA1}Ht<Y2C1YsFuj+BjdTBr07$Pvtle_viRY3%ykk#!31K
zdCFOy$WNF#v<h)ck$XYKiKcJlHms)}F=|bH|DzTLll-nmyDI<V8QD*3hn41T?&)gd
zOMEs0)%!aFS`GMe+lP&p1Haf3I(Rr#u5(T3Q`aekT%TJE|6H}txx0ijopYc^uH9bX
zQ;U9Rn*r?%z*YNS98E+869gtz%#DK&MR6n)R9<Id1g~+X{PB3X_a@VieCVf?Nc{*=
zURL{Stb>ClA~m;Z<QK6P<I|F+-JHqwC7!a6p&`=>A{aFc)T{a*c4i?A`4F2s_!Ry(
zKR8luwQHi>v0DFAO~p=xivP+cl~h!$Wd1q9H>4n~2WwV3yA|G{)ZnYx2U7;%IIv~^
zdL@NK-acQK=p1`Y|2o`79zxmuee<7%875_}k0i-n_giZ_ak%s=uvhxrsvD5~120~N
z@5@=X>Dal-%05ciSYv2gqA}O!#m)Pz-84#?!>)K6qgtzn-}0;B4qkg}BT_bI;LpV&
z(@=fvvciqna~8hS%lZ?$x0`E^7l8~vH}IIsb-KPWW$Ng=cD*aFLe8SvJ=z$i3^n@t
z#{!<<PR$K4!T|J%1I0&Rb)25t3|7bI<*&C|oK;?Eb<h5+^Z5ohAos`tUZ>xp{5Vwj
zD-3p<4ImX(!5E9HwsGKZ`tA+k+!j-leOGxe_~#n`F#NcolLQp|o^$^iOK6#-Lj69V
zz~u>-j8c6x-endbpr{ghZ}Z;96jf<D5vOJ}dp)@`G-e?x#1-lYmLUZbPcg>7;ForI
zXjig=WMlqJ8@?CzfIUuU$I^w7s<B~tr|rk0{cPW&-++l%<E;cn<;<XUDy1wpY~F)~
zGW6}vpQb=CDgAo?Ct$Kz8^kJz=f|#(>vo|)7YV^$--yBFJ$~af0nIxlr=V30=UVv3
zkZQ6av`Momsl$9@8n<D`#}&UzvM5>ipanWXAD^N^CDK7WElCdRnS&<xn@EDD<MFgU
z^=!Gf31-z}2fi=4jp{}4DNZsbpxjxauPjUC@DC@t!vX|6H%J0|#c%yaMAOgi=u8=2
zKi(d2ypgv{jV$;Y;&>b_`dD{X$)aaaJagf}X<NJPooH2_aHok<nD(52-W~mK7!3SV
zM4R*>J(OBM)7TVpkR3sOzp4RBnqzIV;!J11j*K~;1>OvdSXjJB1vMLZB*jK(xk!?o
z=EKS@a|=w}ich<gFXS?>!=FOZ4>|s?|MUU+M99!vPNv6_G11)^OApCy4L6c_^N(T(
zjqzImyl-zGd)(ep0!U)tN&+W(<*-bW9z<>j|NW7DoX4|;@sb(!nHe9Y&nCW(ItVM$
z-8=3N_o4f{zP}4f*4Y4G?KF4^=noh#CM(DiBbSK>)8e>pZQ2%noit1E`A!X;cddZe
zuu<%(*ORO1zYq}eWW7aPir%!lnjd^#%(hx@1V=>ge}+E&fv^1Z?3)`RWC!A*i7t|t
z4#EGi^8r%v5XD5Ya+HKPu-V&`4hHV*$)Aq@VWu8=LAZGE>{|+^F~CY_L~VX0lL2dS
z-bamD;z8L`MFG`=)3FX`7C>DY0j2SvV`ZOZ&;~Br)*_6>bdd-4=_W75C<;86iz?KE
z3LjzU!{-%U1f(DHN_Di>Sf`2g1Tu#FtwdOU)J7+03Hd~uWbrD2(tD^{%FS~-tz_$5
zM8P({S_1jm_JW`~Ua>fZyx67r_Im0A@6Lz)u5<@p>ak%Nf>MjR&(r!x&w`1b!7%X2
z4fC_+FiRb2hXjl?U)+<RUk@<8EY&WcoOxO7zSxLXeCWhdKX=M_nK}M}f#&Kg12wO2
znj~prOgWC!wF$t^c=WdF@p}8a555B0*_b!ylU`hl1FFG^Y+B<NY{YCiZwuT|eX4r+
zw3)f@_<<$NBL?G0uBAagb#?WBBSf}G5)9J8yfap5dY2~aAD-X>lRy30=-unz8L|Go
z_U(*dyQ9G}@85~j$}dMh7d}zPtG9Tja|e=N1!OpWBFO7B76hqPP_Trg^R#gaCPB%A
zMw?7melNOcR_47`){%5_Z7Z|-Qo&@$a<&%o(laWmP+cmxr`TNI1@q5Dquhn0<M5v8
z$e4oj3vGRCT9n5DufsX6GklQ+n|)4F3}@JX1|wYch_Q<x_-wd>UZ|oj5WLl6=JR2M
zdblPeqLtXoWP;?h!~T<oSNR($e-RXOBgu(rf|2!|fs%K+O{cE%(x(5vp_Jd}#A)VW
zIKX&1zgGVH&^mKt>nm;^$|=)-BSTs(|6=UASnW~sjwmS^%-yf2rIL$eA=~Od<Fy3^
zoA~2b^E5KZLAf}Y)c0wT6e~n(z(hQ3o9{){4#By0zi!7E_OptoejpM?_19h)MI@H8
z-Gov5&!W&q4q$@)R8ov=Jki?a^MyFlZUNaRH}Kcr*0K0j@sH0wX}+eyFe2h3A5gx<
z(`yP7NivHa9>G(FcsSu6^87^HUu%y#RCBr6TxUh5d;oLqFeUoxc?cVK9Suz;#1%D*
zn65s{n<jF}kx(>zMyCJlxL0Ps(ff-Bdse9Jt`d9P9uGMtjZ8aw159FAw#jo1PNusK
z50>)><R!t8XTc(iIdZmbGUsmnGmkvtf7Tj?^sGf2$?)mDQech!`ka`G?46cxNMb7!
zZ5O?D*ay3S4}h%X_gE-?0*OSMS{nl2*R)S55<;GhZ9K)$?fAkwfGit*Xc#u%khG7+
z0$g(ne@Ns^OXy9KwzvO?C4EAj{Z{PFt%XB|vjE6tg#F(^6i=tSl<ZZg;t<(Lf?4n*
zzp<+kzT?>$JolY>X-Wxmdhu4Dq)_|=^=@o}+zg3bnRHIQ{8xZyzJ#XPIJjeUtHqtD
zyU?v0o%tdp62vfay$oHK$yz+bx_X&jP`TkOth}SbH&(>ozdxQ$<BsklJ(w&5SN0p`
z$WYevK7w#(OByup%HVi96>#jT5#4HZd?BQ=bFVUW4epgyEB(sV8TCHo>e(?7RWJ(z
zT<mYNp)7W>GG|22-00rroO!U%A@qng++*@*mw}iSjjzJvo*gsFIvq1&l8j1^U1R$9
z_R;Zmf?5&$$uL9eS*dGg3g{hoC++e8jaT80{u8H_MeFKHj%vWpZB%1v*BsRcv$Bx+
zOkLvPnVT;sDna!2y1SU4S>oU_mu>Ur`tJ9<K1|_ENsp&JTvG|ZVj0o}me#{NHJ5-i
z=R?rtE~hBLo-zJ&m^b0gf3zX(F)$nf0bwAlonxYvJ|{xuk~})IZ-}qetVRX2CN(H9
zr~yIM2J{T!;fiG1*S^ZB=OH%*MqM4|HJ5X}Uel^Ri(hz4>|fjsw611Q#g#Rd@{#8+
zHqhP2{?#w{eXskuI92jE614}uz@F*BjTp|znBrUog?Tt{X2EQWVVpbUb&$p{o<I5!
z?aDl=1D1(GT;RWWZSbiyzLpsJ>B#&Ddh-F!D8u?B&(n+$eFj$hxzeJ9&3;14N!@@}
zpa1*-o6E9RUX-Gp{~5;};$k4Wl^|C1^6gBfM>ZMW$dv87i{q=VGUK{GC8|2$5@n2v
z#zzfB;VfAAc|3@rp3vT3#BrteJ0ZT8^bajSr$N<UL=?mdRC9YD9eLCP)pO^gz17Qz
z`|`3j)=yNP@)G%^!fvGXRD?A6#x!Ps-xQS0cZSh3zIi%NB^n+R9$-DgjJ0>oFBf_`
zUfB}%J|sb}ml@jTOYc9L#sK=miMB$#Un^j5m++_msiydS+z+9u5&1Oed+F22Nc=hu
z_A7WrzN!3^8x2hpv1jZZrSwmg;@>FLyiOed@RaGbl){H+uy+m1TX0LeEBFo^jy^GP
z%n<Y1d6&=zva%d?(Xk8@Zu7|#0#W@^uZ?KbvjJJY)Kg=su(LcpLp2aZXW%2<nrTbO
zIu^^s?+DR=I~U5~iZ}^y-Vaw&H`xqt(gHJ#n!yR3gH|nqKYnlekw5QBR>wcB@O{4<
ziehodZCfUYg1LeR;tq7t^CgnnjXW5RDet!)A_>|~iHzIaG5YU+(t*wh<4ILElwnD!
zbqF`@9gPV)S6?CMlv$;-z&ejKP~_FuITjMRMNa=4Oa_kSkH(wxf{W}-0N<nFM^95T
zq)IRRR;@7*{QVb$w|<er2>!C9;o9uq)gqSBjziFqE^1wYp6U=K=$h&s;b#1+L`x?L
zHB@?U8EU_y%Y-qCMIl}DgKQ1ozEa0qP8HF-4KWAxp-&0w5_Ai{u+Ooyt%$-=;D%d~
zb~1K{PEFLxJGxSb*Ca<EQ;ihd1Yh7yjJIogc<b_++?G@KjFXWK6(1pYI#gHzDCuJJ
z3^mxCdxP}Q*zNEMsGCW3i>I5b_TNLU>G?K!<4*buz4$ZvR?m|s@VHS)g?1t{VA=We
z$VSTV`iiTp`g8}l<2NQ}DW+_ibL=={>t7>%<tv#bRpG9T3i%9WY|6z-ukTvj$b-RH
z@@>hF#bT`<hZy&Wj~nO;<uOr{*u2201;n8aBu5`kXc4+`)!)#+SA2d-Q$`BVQ?%J=
z%uHHe_{_tjfGqL;Bd<4aut-I|1h>9D`<PhQ!^oc~&gS*Du*#S#JO53Tt~ZuUmIh8g
z7AK%&^Uzh+IJv%EG1=d!r$MUcM>3WvVOxiXV7UQP(CW?MqvLHolMInpE#W1bgm?U|
zK8-C>`1Fvs!g|}U(WicLRle+^E&(OLsB-G?p#bQo2wip`_DP(<^5Bei=2H4KA`z8<
zO8?%+`5P6ZX|EQMr@_&0$z*ECV*c>S=mnOfybWOKm0xlayLG=s)d3F5F=%mF(PtfC
zcMO87f$zznH)?~?;}xJW8PS(dWij`V9dn|qR<RVxmq-$)RjYbTzu`DV+sgtKo!XYj
zNtcnT8mz^;=cpiA;A&%zQatG%;lnCg2soGWeUPtB9=rA90q7e;ZQ}1UsOB+_Z_`Pl
z-c@~<_YRH2Oyp-5T((m@n2G=IB57DO3Jj@g1;sV5akKUB56_cv`X<qK*M$TDbRQrN
z#}#J=Iz3achFC?Scehnyl>kA{*CuhHjdiW+DStMI7<G>rTik|Uc*$P;kqoW+Wk7L}
zl<|aow9EYSrY-mQbD>4(U;&<KqVK`vExXpK_-vR}P~^ooay%lE**@;7>UXz1!H78(
z?6id^gE$S>X|o^iX|yA2+3VXG8EA;Y2NVc%=7j5jc4>f9rn{RtohKyDKYGTphAXSO
zu|-aiHPO0sjvB#KrsL$~lz0e4shrW)kcoQsf@FN~veFTTZa&5FHl9<QT=MjpieB;D
z<YSX=*OK(CqcYj?vw@#zpx?%eauO=#%Yzv%QaqPKvye<tkH`l-u$;wAZh$yxhc57F
zXsro{pm_X>M)6>QW*!N}I{7ir@^xSA*ko#SFb>+T-SsY8F7$Sim50oMt>k%hB!TKa
z10#F@K+SomN2}tl+Xy)iaokUaAgDW0W_DuSkL`6C+V0-V?-$ZK*vU?b$z+r>>YkQk
z;Y#q^P4T8+17z+b0@?{H_;YomitSFmo?&EchZ9*9CuIp)ED4zxExvg178V5yUhysp
z_2+h=l@IprT?wMur=_^qz6&JuZQ^2xm41X3iJJ@Ku+&@>?-*11#Xt047D3OwG82az
zs9!!X;#--)7S129Z2A-x^u+q5tu(cB1YSwVK)+5@3Lne6dZ*&wW*zpHU%la`?rC;e
zNm=qXEwrNwgg8*l1mw=0`)>bH<@5RE#>19Sxdkfk!FFS^JVF>HEv$MpKJokfUd*bm
z`Rhc5<+mo29F1Cp5-e*JYSb(7rrC_}hKmD;tUp+$a@lQx%K$hNM4EODm-4XxYz-M5
zS*{Wj0v1QFIc;dX82Z-@zpLvMP8tAdHQ=71O*Pf<U*BWTemKzb(2QHI*bE&yG?034
za2OM(ZRbil{+8H_%N#90j^bn|x%1D`X$HhL`ZZQb9f|oAcz!Vz+l60E596D)dv5Xu
zz&+k~8z0b1`)&WIh=|^Le>3D5OP~?cXSAx2l^joW8psm!(VSggC7H*Ez}TJ_JF0@3
zg8r4wCj0|a^Nl)XB39CJU;IzOo3ItKh{m(=+?o}{u};j<LNu07M#%yK>T_W<UD5x8
z{ilOovWNU6H>LlKv+Z>4>lj)Oej|w!zKTus`#TcVj0XnQhdtuOcCPqx-3qzxFk{6U
z$USBRh|Y^iQZh%>^H7loe8Oo2Ja+b2+JL4^TnoeU3bR0oH2~aPDh1PKPmb+)h?^+V
zY(qv>u6rA=V4#Fr!BFQsyvAkU%)EGD3c6TwT#1-A^>Vi=*ypESll^FkR#{6szdXX~
zO<Hqya+P#cH@7)4yrMgb7<g2C{JWs?ClM4~-4hMSso)#B)EKXLj_t6e?Z(&2d92l3
z@ov0I44%vQVsr#%UxHp)P#YN~w@e?Kf1ja;cE!iUr2iU#nKzE`DR|N`Mvd;{I7G0m
zc8uw}Nr8C5ha`eAmCKlmzfXhG>!Us_AtYoYK3KDm+&*zyRQEK-Kv}_!DHp3L#l7@o
zX80KC3l!WOWh|34#_a$PF^N4c&&@-ddiaUsVcfAgyoVm7rGLTe1sops!N57?2iQGu
znJj1pQlV<UJSe1-X3q9TXtsEULD6u>o1r@3khwT(9U_?q-$fAcX|TG6AJ09pf0mLO
zwi=lvZStUC)V3Ut6Pagl&jGwrXTSSSoPwVmWL@phqpj*7wBZlsj(OOoL(v@cBz~#p
z;a~2(X28JUwt|^EadLWAF^s($g|nGcbb&fJk8ei>_Xx6QM6(}rO7tjSImlWbORmoO
zPRryl8}G&$16L;=BEf0yk<h}+okwwe!5Tap4EoVG{K7O$F1i!ngE7$U+o`VnE-Bn<
z-=|B&gp<8#@qKTQ;K+$_&*VS?J=3~y)Khf};6gLLf?RZHE&9-DwstHa0P@_vgAbTv
zYp?HY2=K`~j*@UlF&9xr?4?sKbgKa6z3*KECg`A=!MB$-Ba4>@K8vUKPk@tEEil!T
zD{zR!op(JS1S$d-EZ1rg5*NCXPO(@5u02B4o$%bA?&>8DG@QD*a`IAd$$`E`um9g!
z0K98MEakReM~9JCpr(r~ZkH)%+WXAmQ)$bFp?F+f1IxN0qm4cl)4%~dX}Q?`Q&`7<
zO|0^pJ}t1wp;!@5a@zRDm5Q<986lkm)R~THteG^8VyhGpEk;*5bVX~`Po!}`euabF
zACwWum8FKVs!#BF@(LXIfAqpVLiGiq+{tkclO9xgLazXsbqDAR0@hxM3)?`Ud@X#^
z;B}7x9N1!2a->B<e3m?F{x~z7AMCl~E!uQ56nRY@5c7C&$PdWZqDbgsup2ReDZ_Aa
z@P|L14C_P8qg!#L0K5{`eK&mS)KUMCy)klENems^Du_^lTpz#$C>NNd0lO2Upa%R#
zB*pEMC>^kkq$X)MYXlTWQZj54?<|DErSpasC`>t1#T+3<tKL(3^%$&Q3hRPuZM7`8
zb%@@+Xx`=AB^A!k@$=KiKEbSlA`7`}Y$nJAh(=oiXa7$xZo>iryy*K$D!bqzz1xQ%
zUEw^ye%s(?v?bJ|Sa)|T+z>UJgwcKZ{u({a<F^;Kv^Y-q06*bIqWBcgpIuq5MTAQ@
z9D{V62)vTbtqDkg;~S6F4BuJvm1<|YlDi<DM`gqc*{k@b;-3($T_s7*myYx~;JS0X
z38dUsj(?TG=$(IGo1)33>Im0QFvr>kcxp4y5%S(k&Vvf2O)SS@ZjkO7IMNanYxFT(
z0cCGNcBexuDZe3%?yG4bH%k7Vte)*9&ZyFZfN<s!o2YuskMxDj3FvkLwy|yA5~*lZ
zpMt<49iwstxTAc&`uXk5eLv~snl$+tlo>r;yH<ZaXZaKSmp|ZB*$j8bP2i2am(?|t
zP?+K3;E2LDdS^J|4n4eD+5^#9j<-w&<X%z<YnHq?-VZ&jE?Sab`4rThs4h|P1o$A^
zV5!2<?s+_5V14o%!yQReF!N6F999F#E0B-VjKc#>EID<QKR=NbzP7f)pxVq3aQuk3
zGs^ATWfGM<&XV&m#})jr=s{(Duxp}WoNeKusL4Ek8veZq;SPvuWhG<)Vadh+Zc1OC
zY7klx+axi%eh`H-MBwlePk${N&L#)mB^yuuTGWQI!5!h*kl?=iMy-{XBN$C4Z5?$+
zNY)Kn6$!aN71UNAL+KdFeh-~rmDw|;UTaaa8hzB^P|P)<Xl!H&y#!>o9iWfIdxG7@
zE1t~K2SERT{E47Lk~=d#^^x|=Ykjhz_*TA2JhWpbv)^@_kMor01+&6?5`BD}WwrzQ
z&#;gWDa{rk#w&i5p8}<<iKL*r*{fNXe!m|7rJKObr0xHhLNk{Pq$a}JU?|=Xjq`tq
zX3c7IPXD|hg5w<dX=qNj;(71L4iH<&y=G$X(?3q2i#MGeQ!ze;I^<wjBG`RQO7IPo
z;EyE5ZG2e$#4+|DZ{QrKo7?mx%DAZCnXA(t7PFx^WwNpy>4*}F$5vS={Q7+D&w{#Z
z4(Ena4SRerdikd6Zo-3W9yIch59>08W)JXwZTj8uqMqAD|6+f7JC;O_zcr<GX(#0R
zdqSMCdGpPf{Yf&#E#HO)vLh-?jJtAVC}ps0!fFJlVL=;_?WXQ`C7Y@|ala`s%G{dG
zc?di#XTHF~NhRG(X)PW<o06O&zuR~`kwo{=CFZr9mrDaf!(P)2?kC!`B=m7rm3GgT
z%#odD5cc8r0tUWz?N^Gj78X7wGzSQmD0ac;hbO+v6}&;5<(5--6~x5&tHhQZd&|r~
z!i2WPZYk9BcvS`lkOU9pN_stD>bKbnf+l>zbm~lQm*5>K)zo~c#=vbRqV8(|fy4*Y
zozbVRmry5IfIhVIEUYHUXXQ3#y4i!nL*&~=*u``MI<KO-HPeyS2h8^q*=qjC0d~Uk
zZ<0Nz$hzX;WmIgXS4b1f<nD&?G)Olt-{9G^XV%DU^T?PcM>knqd(Ti40O_CI@x0IF
zu*#;9ncor@Rv=QC8{F3MHIk-zvdEZ8|5AZ`u}`_a&}?qm|9JON+az*iH<dcQ^+1g3
z40Z0k6Nj}qYCb2yR{)zzlU!_gzY>X%^7;1X*G$ZNqO9K>1Q$H)H(i`2WYkqce3gjJ
zpu>WBKYao&@r&|!dO=o(I+~t2lZ<rULDm{0k0M>YzIU}2?^_h3a(*+<<VRw+Hoae7
z7=)z~`g-i@DLlr^Q+k9lRva0FQ~)9ps&B>r54RXXu;nYHnQJT}(;%-0Y=aYBdSRWH
zbB4$NrR;|~xxa`^WK!$Ypk`YP@z1HzR`SA=Yogn0`ch-I(BIVhD-qFlwxsN|^fyN#
ztvYZ-9#^6{>2pMY1%b;Hgk<>2Xrj8uW$DaaxVgo=Y-Te(g`#cez{zmZ+O)=~^|p%R
z^vgCwa~5?oB9g0TRx2tQ+EBGOH4Cv5lreD>>S~De*PK%Jcp<Q%AV4=6Ta9NeZZE&D
z<Q8!GkT951vx7X5kHr{Iu<G{Qop9BC^Z3OEe_dHZ2odjs@wB;1+v}l2!w4>UMM+8C
zEb3mS+LRpr%{uVK)u<xvj~_={Y~Kq_dn4Uf1iR;<_v-FfM|#f)2ugCS9^w{j2_bGJ
zu>y8iH3HFq5wys$o#0oO<Jvrd;FZyPAuoxTHY9h89lK8)@Z6}~^Lj>L6I3$|;8Pec
zm(flCGxiE_IHKcY(+8|b*d@?!vX?fPGqALu2RR~^$2r1CvXURQ^Rc?Dtx2cz?W6L~
z^V&DZJncS0*-Q%UanL<QDGV;NUP{K<J@_sc&*gWUAfL|p%V_KXC7E+qmp+wZ-b&yB
z(e^K-UuzJ=Xt#>_Iiu_tyf=V{TL6Q*=||emLfxBFwqW(<jaXwdZ{{HI48293?$X)j
zD=ewp)PHR`0Izilfvepy7P*uhP*Vvl@jjQBnv;Rj4I}t3@Mcj#EypE4vRb_@CjO$5
z*`;mrgE5S!?XSlDx<W?ng4gH;o$h|Rve<kq)r8P_%Fz#BQ}$(X)~$(M{E_K(VW10)
z6JLsLxiaDihP?W{Wx}eU8~YZRiXulr#IxFK>GtucbWDZcv9Knmn9cKlr1-PN<~hP(
zVy?^pY*25>k{@H8ll#Ukx0J{W`wm?@doT94LN<%aGA5AAJq_EEH%EB4cBh}(L#*n@
z(vUdd-_qu?<~r@yg~0u!Q(L*abdSRspI(*YHY6;23!IpirYIjvT@*&wvc=^U`lBwV
zhcRl17@#JDjbE6Kdj*)(t>e!y0oF927NDbfb#dad`xF!x6Jd1Y;$E?y$?qAfZyHwg
zZ%5bq*?HIjCJQM>Mf-6^mszF*DQl%g`|7`@8Wy12yrdaD+CSNWJhCs`N5ayL9~+sl
zN<O{*^=Q=!jw1VgH=I&cd*QPQ(DC&gg@1&859rjpdwVB%fmbaU0zLiuKotg<OA&fD
z2Q(Vh^0tFD@$LE5Gj{vbb%4@`ht3F`&0Ppj8+Z!V<c3H*@$ypAyc(zI+M9T!IW6>(
z!>@ZsIYg{Hj;t-DZDMZDGUw*yn?R;cilHFn!PK84|NX7Us=k-43KYjy=vyiaE31kz
zRpDrjnQNa}`}T*OrBaWI<mnPGb>I8=x{F75%U-&%B_$lp6f9~aE3UWhIuj37!ZC%O
z$BSe%c4BAmZ(gw>RMS<|CC-Cfxh^_EPHZ{1ayxXTK7T*dAk2L~bT7IHlx|hKTMNo?
z_5?CRkTGPqq1=FNy@1e<p(zp?_vig?kBs*5EHpmfevk*`RNAvw8IE=iY;VR6Xm#vc
zHCaj3be8F=Lg#}XJxyf%9>61VC%z{ra;bmDGcRm49uU-k`gh0N<OQZRUgrYYSC|K)
zM!G>WaTap8s&!OyHVZN6`s9cWI`*=Q8@ECietLY>DY?J6GhL?^{j5^V%Rtn^d{=Ak
zw(er%x18f$7BLjOxpQ2E&!+czfHYAW^E94^+wG%-q3_Rt^?@BeGn0E3qfv}UVRboH
zguLd;dJq$@kl4&oEzM{ITr=ZVph}4YnOzt`UF#)$SDxJ?jsNkYf9?PiyCTZsB!@86
zdxj7zl>i{kwN3W;fgI-F->|2C#)(?}49#YOclZ3E)0|?CvDt)jOxxd1Pj|bcW#Z;%
zG<qXR_uP}2H0l}W|H=r(??&;=F_dKvh-~r+yG`TIcN?Kr%XUs0-Pp<WR0mr(Xl9^5
zlCk%VHmWQimCThsQ4S?1r2#dZ<KtYCnmuG}_D>|lNgR!Gy-w(Z@?VSuUgo8N_J$WQ
zZHFDl0@g*S&yYOQf0UQJMmI@S{)(w0(&ag$fq>oW{V!R;H#!BJ0eyIGzu1!QhF6!Z
zjQ<cWj?5u_YKZwe2#Uy>eN}hYy3jWOo`|K^r?egW{Txh=Yf^*F+!@q_UmE#?A%=id
zA)3KAF7H^R(|8<c$@zRedY<e`Kf;REOLxGMUkVsoLZ<03--tWktIz_SoT}B<0mbpB
zW3xInw<Uvx-!)-OlRzi?vQ@=fyV$E!3Ac}Z^XJ7Vi5}ok%_vNM$omQCmAUVUdGao|
z-I}v$1Xq%VlceQLMGFem+^r|zqxVc*VDR%QxuY5Ey5*cjj7J~<I^XQ(6&u{dW$i&Q
z^B2=#&)b`t5Ru|Kj5#bHkvaA8w7ackGD~_#HLj~OPAhEpNbxX-ZeK#GIu!wTdFo44
zpf#*1GL;=$qd@_r&)ev3Z)M?+{;K#4@H_N8-iXNho;2p&7kh5}JO8L>tZU^RK2LrK
zTr_YcdNXFwwzqPFf5&fmovTcSEu1i&^_G85uf3+f`xKeyRpU6A72xB$wwtvq6<cwb
zQ>{{zvU|MwNJ!fT)1-z{JJ2zwIe4~#$C`IunWB4vJR$tJPE@ZCO_8VNQ-+;R0KmVC
z5jK>!v5~jFRp;MN&Tt%8)5=<GV&&n<c~M+A-h3S4z6&5IwPRTz0AnT3DBE3ua^hWr
zjH!{7RR+*NI3Rc4*|@<H^Ac1@0JvO19#H(6qd`XJwZJkIt_=4sI!eXB7x(<0k_5|M
zaSwHl$7s=AugIv1!OOCdQ_KL)+j~R`Y#L?I;!si{A8p<uqr5!GlS*{s77B@ISM$yF
zGO`#ZH}akGa@^+W_IjudJ~;n<(6c_G;J>}XFRpkAAnGx^g>=5ZW-}dysFRLVpoh!n
zK*b1+Q!K}Q{e*0a@$V1o7%6tpc_qEJSwISS&`E719~7XFe?a~F`tfYVbNjgLWCf02
z{Myd98sin>+#)!b^j}|poedbvK&IW3rGjr>fqNe3@hR^1q}{*q_zTP-T~1=R2)0F;
zxNHe!<cPwjima9ANiyC!HW&U-K{F^lRDD~xt(Bt{s!QXWZvvQMc6_h|@h!LK2VF5F
zzzj9dU>*SInkq17H28hXDRNP~-*wNUnQF!58<(%w1V(mdS}3ATZJHw50Rs^*gr&=3
zi#%ik1B8Eo>r@Of<XZeeMpb8H+cRiimrDaLpEJesK3Ysi^SjzBLcc1pjY>pIlg&IW
zC;?hCHNWUpWw|#LH4)8c(S{XYRrDrK`G&oLZg(BMqJ)>gH$ALjWsyt~PI2Fhy3=$h
z1suOh`uqk@vcEyOrS_W)=J3__!WD6-p5?rkokm2!7hode+nC{xsEEhggHI=;f+@@V
z?gYv{oQu3YCVvmAYsgyrK05e`Y%Y4dOn`DJREx}mZG?ZR$pYMPn2`nYOu&N7Y)#Y=
zOyajh7c*Ch*4o9Mv3nn{1wP*Ukz%aL?R{|xK0A}1J`Um*c&PH)ufnAp-wx(3*cplJ
zeyyIx7a_Oo&9NeI@FEGIqVhcAkjkgI%6)%1V#_S?zTwtHaQ^MVi^I-G6xKg;q4IGw
zn^Ms8JxZgMEO7%9C#!wT^>oRlDn*dmSux^1?GB>w!tw6tdnhA@Ipf#X4)0wSODqN0
zE`UR@z*v2IYYKX&K5=(HN;y#$T5gVw_V(FnF<1IVu5g7}wXZlp?@0~Bm|XQqIZ2_T
z`2BDT6KMLNnll@HXE_}~@h(q=UPB+#5=C{s14zyQ#!48wD7=oSv=&S$YfQcO%^?)!
zeM4ZPtshYF{^FdQ)EzKUP3$O%jo@$5^vN2JKDKi5eNpw!SxI#kw3)wH9;^ADSDiU0
zJRbN}d8m+BK1u$Lw}05k@AkA9nu2IZJDT`*`Z(+JPo+PSd7ngdXIB%3y{ao8W(N=2
zss+|m4Ss`69{NTWPK1lHi=Ac*%P4M}9e!uQ+BX6%>~nkvQ=R;v6`Q{icU2y?7UN?h
zBim_&vjq&*->+|dDM9WcZp|7`0%x*kTJ;MTS6sHDm4gGe2i{Wl!#(twzAPUO#QeE`
zBK6U7b|@NMdTs*m9hbC7+r~rp+P9Lw>*NW+6S1ylCbVCRtbWxOXu}0LZ#uVJj#Nhn
zQ7apP*>c>mERUZrm(7LF3@I~$YVV#jvbYdCEOyAYt(JZMlfStI_Dh1HnmAI(R<w$g
zNN=t7P2B%d;>d{DyzA@Me7pLoLKg=822)(7<7HVEHS<pyvs`8M<QR@sd|ocKz+Isd
zw%e<=RL7mQ-y`WlF6q^3S&N^Y`%rY6YW)0vWP!~~T2}R{f<v;O*BoGMfHfA5Py5Hv
ztBTGM3aZ%JA-;U&biG@-gwlm?yrWxRHflvw*7~0CA81Ew0%P|TE;rwN#*b~d%i_^3
z(Hv1CE%S-lKxP&k9PdyllM6M**3fgO6WstPpPGgkKhDc$Wq6uMiot*7w+b0E^?5Un
zKzeE@)I8X`>E&@<y2b1vhf8(w2hgWZH7JUA7u%{;1H{1+t5%fKFsG6E06sU)3L>gr
z0|)``lZbuB`Y2!tfqj`L%|qCgjJX`xcL~PQz+vtJzF+~R7!E%UG2Lw|mcjR2Fti_Y
znP*jT?hJo`2+Ta~YyR#f8BcKxnv`go_YPCfE~+M;)fd#$g?cb0!=$s<p^=+6s9xd@
zribqDWZL`fYPdQ#BBD>XAh2w}++eOrn=7fPS>$+Z?{$$DSs9s@;b+{|SKSOJrUTAQ
zBX%Df;u+dm7h2vEU0g!L77%V06=~6~4{5L~Ryr+mp~{bZS|)|6s-CFsg8MQSw=&P|
zr}_gtVP^Ns?3&+XebPO`j&D8}l{cNX30Pe2s_afC)-)mqhVRvAxFk;+-%|?P79J<h
z+dd$b4cCZ&=~g6j9U1X$EziMFQh>XQ;(@0_Uz0tTK~@AMH#Pd;&?`nnk)shvMAZth
zkzBR3huG#=5R5;B^7|oEF5lU-<dXvxn%I*P@oERjF0VJ^xpUSXU_Q7thV9=06MqOC
zZo(CytdNTl?uGXvR3^O}8Dkms>6$g93sw*zJhYB+F>A%}N5<S(F%J8`6u->OXvkJD
zMwOVB5p5<P>S?`LrRgi^@I>?N%eNT+3%ttrrfJ!?<uLvN<0p~-oM`yl|D5RK_je9I
zka}eqN&Uj<btr*t1UKJ1r>mXK=^DMv#hlTKBUwhFP!8$4Z+A*B9|Gc&s5X9l`!<P7
zYGCW9pPYVub~BJH(AspaO;x=ZT~k9f14}8+?3^c4>h9!3zM3H$wuX81&i`d7C>Gag
z#_pRV!6K^y_QqJ<YIO6Z2zn(a&m~3JHA!3@{$d%{b#(;NnXMM}IMC3KoncFn^3Rew
zFG+OT)TqOZ>-lU(_ZvPa0bh2y_oWg?9nuO`p!wm`m#%C$r2dD4%<qW>v*rENCj5?c
zO+I1@C{VORTVlDh8;x%c5;P02wN=gArgT5H<Rojw+BP_OIQ8c6XMwrwYFA9kK5-f?
zSFmv1yAf6ywFKJ78O<Prs<ZYH&>byr5gx${jULsB3pyY(g+OWA+G==7Slo$?t4+&8
zBuXR0h+0v@fj^=JP;wDy_ad`6qn&xYf39UMw?Z(nBBjP#YWnC+duh^F7x_(9y^&H~
z0HE)xOaMzjw7*#$S)*QMC+AlAmmhdlKD2`YdI9Vh;TLZLw>I8!2QBong44m-taMKk
znf@2S!CxgUV?zH$(1P?Ll`dfqPTLg4{K=;a55cc)?UOPaeO&U%{E6?@FW(CeD%bW7
zP-y!J=a}UDZ-!5>7}&nG<VVZtznl|1d@}ZJX=GSzaq|97bbL_0T_cNq{o-Nc@hXb&
zdfl<$_HL3y@YT`>8<Es)+JeQPzg3(4wi#(49E9Uaw`U|@)p38{m)v+;$IY1igi&si
zQ22;9a2dJOfY8@Nh<U+P4qsUMbj|PD2o|6oqD_&wPwbadEIhf0Bdu1&IKr_cTF<6N
z82SCp-|ZIKr6>5~{`5Czn|*yHW>}(xwc;<WdK-eT%V(g87r6vc<K*gqGEk{fj?=$*
zbTh(LRcgeQc8OSQ3T)uJ*<8F$v_DvEs+fyA+f}iu!8_c2o?P`s5pRAT;;yIIA|Xdo
zWD&K~=-p{}Yc`6WPg3Vt1g_EF?E(`Zs0}R-&avcNXe=Wq=WvxC=n763aH)N;aCSd*
zi)={gncg+`goyDc$CQ4Yb0l}Iou863xA;Q4dX+Va)4UCVXXUBxmp9MsiC-<O2Hk=6
zGGZxo<V-9YV?A7?Tj0i4PIBNuTV5Zh2)_=;4*d`JI7-e26c@&|-DfhhdcAfhQr8&9
z#iisC{8N!QPlmbW_GgYRKi){?a^j}e!OpJ$I?3<BfM!ipHpS(w7;sGC&-)?i1@+03
zv6gXhY7$-Zd%C-6`X&xf@0_+b=F1-Z&2(ie-S*??fX{aAKY-oJCt40&#&}s3M)D-%
z0pWi!-s$d+vW`^IP|+67bXXs=JGtZ1c-Jw9BY)iad^hF*p{$ki)gG^C{cEycv@M=x
zBck2(=%B?B)vDkNuMeFLM~}*yn=XdOCP(=DEnkZf{9kB><;ma-ij&K2WUO+&>GQxH
zZasE(NqhJPRD4sl{C1B{EK7b4P^Vq=SFeTf*y$Yw$t9w<ctA-x)>RPGM<i&!jT{AY
zcASY{@n{^L>WUz^yZK}2^Oxb@hKAoj9`n25l_`^t8C9jW8C!BJm5RZO3>=|R*?@J6
ziu~=}U9F|j6`Q`XhRRWpXsQrO{$RoDk+x!~ysCj$Axm~Zc6~#`{>45g<=uRO`|*Im
z-uB#TZZD8Wt3SAgQVs7xe~aC~5Xl>uHy54z4pbFWC=pONbOb&Yv)IWzt~^es{>+{g
z1GP+P1G(*$uQPOT7gJoO4??h=9xTNi6xe9}=luI@5nNWJDGT>!u_+i1gbUR>MUn6h
z^{nl~IJ>`5rL*Yem!jABUsTCT+kd)J%^=b=YpU#e$6sYz%muIH8vo`-iO$xw2HWY2
z4^}A}c<0>ISY*_p%y&u0i5t8)o|-Pl<G#2XRexkZ>M)R$i+cI@`PD1rWrf_hre|E`
z$VG}L{X=u_xRWPuIh03LJh+Z4F*H~Y&RfovIsrM1tgC>v6HV&`;3=QjsrKN<DI5=0
zgFj)}!Vm)xJBx90vnw!CQc4T$AwHxYynE;%aHmDx<>(C2b4#_Nvt*U&;W$TXP94M-
zKItdTxo%NcZloIb>RnY4O)`)I`%r`q8)r7{TTifLiQT=fs+drbXmLm466M&@aW}Q5
z{fswMUH#IUqjOH*Rc*F#aWZUl|KVv#-f_8IifS9mT~%}ifIx+gXuR0UZH4HJYsE=^
z7$WD<h+6=zJpx?y>bm#V5+=4XM7X|xs`MV%nd5$V>{am_lmmF527d+M0HfE@6Y@C$
z)tIM_u-4E}Baa`xqg=wPQm%%Sx}uPsZ^`$l1<SDK6kQMVu1RBZ4b!kJh}+R{RtMJ1
zhVgLoGLeR@o+~q;{p{smbz6*>qtWaKq^FCqR$mXQhz??-Dyr+Ne)acTPTTFa*B$&b
z7n8kyUDE}?Tg<hEU1(S~RB9MRXEgZVn3&Y6gU^jO(|%Gc_)H`2!R(KTAFK+ERlkUn
z;9{xB{EkLtw;=eZeqYwo%U${XeIZLgj`&#Umt%&8Dp@&C#ayZzh9G2Gm&Oh{tq%23
z`*{IlcJ9f2rn%?{^s+bS%<q21`g2RVEtAAz+@Yn}fBb967T<<o%*DQ^Dy(68Yk2$K
zukWatis8<0E}bXs&0ie<qmSK3!LalX$EzFH4qB=!H{xdwc!pJVvUM2A4s<J~N|nK1
zYzG>yEH)PR5JWZ<{nAY=fwu<p@qiYQ$)lS^36ZOBw(m{>GZ@X=X~dR4=W47gpA!@;
zSo6GL;E{~BElAcKsAE;dyp0&b2xu{g;M67bxvTjVFlAZi)U@Q|vq~hi1G~Pw%D=;I
zu*+iMsj7*!xapf*NpoLGS^E625qvD;xj;<(?L0RhG!($X8{@-h@o8hZ|2aoA)G+0h
z&KlpK!u_&3>K^X}qrrRAJx5J4&iknJ|CxC0+cmPvdkN}j<+vC??H;w~GRs<xr$L>m
ztnX>k++Gs0e<C`A5@aGpSK>b7#l7*poaJEAsiMa9r*>L5>IE{3XKblNiM_KY#}sL1
znXF4rQ?zSBDNjg9XiPp8?VBWoy0^gh_E?~8-|EIMk0qMh-tNRt!lQk+u2Qoh3jcNm
zmfS~>+<MEF>thHzI-3!7N2c{EfYaNuS*e~T3jXWgT|F5{Jw1M|<MWs!<Z!vJIF+$T
z)yJhlj8!W#W(yQlZ)wa7bujfm`;FrxxnQLVnqt28Dfyy0;WLn!<xGVz3NY5E{n@|F
z%g1>)>|FX+nDbl18@Qb%Z~gyQ@f2|0_J0*`H`GR*g~t3%zI;=~L910<BiTjVz*YUd
zSM{XFF1^e??!55$i6VvVr`BV)Es>W%Jh28u^4mS#vQIx&1DxK7pBym=Ag^{93kof`
zHV_YUprr(?=A`hkW=KE|&A7&s`lNVY10Nq}1|fHWdT|Gjdlt|}{KCzKQoghyR)hhr
z>V^;2-jadAdDDAZ4AF}GBD=Za)}-w)GfTQ$VHhB1V<S?R!dpuxZQuX4vF$$CPm9iE
zl_Owx+ig#Sea2~au|WQ^t(E)|K2xxl9TbGQ2xdt$H5=ir)%jfeXyK1&!gfn;YtYuH
z7$TpiIN(D<GEH$E*e)&TR|VvJr6S|kuznp2-z{pwgKPTNOblBITi~7Sh;A16!cqje
za8TebnlHNIHU<tQ?T_}QS<S=7x2I6|^WF|oNU+t^2_7E3VBfdU*nt>(Og~2p37RQj
zV-0gc#w*eeqv`{t^oZY@L3c;)RL0z{jE*nEidT@bP%Am`0VzF4u2vr@kRaL4J7*s^
z=@>$>qz<#xF8j}<3b1G<pRRXl`F^$8K9c$$es2WWxFhus{Hm|V=axJTiBEygOfHZ`
zWwLLZpTx+pLhnYbVKWzR0L{mKCe>}XRU905+1s#&`e*X?WJCno#1zw@VwuY~PTY$=
zD50MLc%b@!QB;<dkD-*MnyV_>*<{f?19BHSUwuP35yGrYsw7lW&cn-J{rYH*`$If!
zm>=K5$H+$3e?JP^SFFbO(qyvvq0pkxoD{V2M+^})e>Jb@DbdC#O(Y-C5{8Wr3Uq{`
zJiD_ADJT?)zLIPhV9L}`f4catRiZZW&)UR!L0ia{u*xGEM}x#w$Vh5F<SEn0!yL&L
zi?f|6mpzT&{`{RAC~knxoVimk7onf?YUb2Uv5UP`O9P-E-dc&%X<{MKD?WV*##_C7
z!B${}wKS|NMZH--PU7)yjQoBAkl%VFw1}}mC7P(=KwBq5aop2yBi493%qhh;snU^v
z39PS8jYIzkQwM8!5ux}R?31IXeReO+dg*65TOqzO@UCNwI{p;x!{3ra_G05857p|(
z^7o1`N*uD=J_~k;{OW+$e3`hZ;fquTjM&1G>faGrK|Qv!&+NOKkOtl>om!2O3}KsR
z^XRNf@iAk%Z+26);<NC}xmlXKYNKEXiE|eYF`JG2o0!bzYzWQ5d=394)fwka`_ZE1
zRk9ho!Y|8lFDy8lXB@+`%NE8Gs|r5PT`4sp>)xyRob7nEvIV<}{e5}5d|_r4fr|;#
z-1oj0ul_`z*4F;2pnnc{#aFNZ9bNz4qOR#HhYh$`$8jIaDv<7#!K^ReI$0MeHv%<>
zdHJSvSdOG8CMU~X9?lefFcuX$2r?~b_j(8N^3uJc@w%SYy9M}(?&IIa!n3`caLQlq
z_`uRKfNGZjLejx9QZ0)Gq~b~+&%C-;a?$u`mKY%4ZExl*)@?f0-(z*o_foqpp64mn
zV~h*G&IN?@KSu<3`OaFhb0HviZf)Am=JrYBk2MFpKyw(E>e39HQ04&*x7RKIlHV+Z
zzhrhy@i5<R2jl7w<#9={*WH>-h?7);s-V-E396iYz4$#j2r*e&PCq5!0N*JDhs6~V
zh=iB(6dElxAE}u?^EjUTL%x7xpRKCzk&6R|X3zq!Uss4)b9+IqFBm92&SP}uGHE)5
zZTQ4*hVk5cw}bbBy<ndtE-&}xS6m3PwifQtcc*?HqUE-jfZK^MMrqgAy;wpHs~_fI
zu3}6~Or$v%o2&2@kj=?=<uaarFTZ9dhbfT2DFdYpT~asq{Hn2QDjA$7w}?YWGfULt
z`3Ovie{Uu^+$k1Sr3;(JYzvZdekldrr4zpe!Gev0wQ3&9NC_9?fC>dBCj?eN$zGT*
zmn!N+*qG8b8@7pPesov8E6N@v_P!1wdRg!W19ugGc<g&xrRct-j<Z(>d-3c4VedVl
zqUgG`Q4kbR5JizBf+$&X&N=6t8xYAkLqiipK(a`ZELmbphK2?VfaFY*K{7N!ph*o4
zx6ya*%zX3l&a8E3?*HHQ-&)mOSXF(_+57DM>}Q`-b(A&<`w$ZJ0@K&c>GmwAg_wxY
zg-ABpK|!4!0i`0>rQYMRvB%i2OCQcFpGlA%%s&>34LJ1$;#mZ^jE&dVo?cC+Ava4w
zXZPqE`5dPbFHBnF-QdR`vo7nCS{8o5i`?&I%mgcC-TUz}nW(SjIMLAqqUvo0?ON;b
zXQV=GMkEWX8R-HKoAVp6uqDGWkRtP&kTT@&iLcFRL*ljABdy-AB!258{t~rY2QQeR
z$%U&Y6l?@M$GeQBSMr^q<x4C;%a@r027m8Lpts%6k3e)Kc0B}C1PPAC<QQtKF`PL(
z(J?n|rg6BWPu*psax|UZ1Pe{qL#@77E}JtW-6jDE7SnLqC-8%+dT?e!IY71YK6CEN
zn~FWGauk|X49CW#v+kEzh&CyUVl?0&^TT%0sbUlmu`baSk<K6}IQ8osi$_Y$dVSt&
z4LVw8poJW$R!jP{f5<m48^{8Xd_dpqSS!sS$P^hy0M#t$#77ZLC}#ZV?=S3Vc3a$S
zy(~;-F-mUO*?nLSt0<_S_la5UOIM?$_^wsDNxbfYw2htqQpwnV89xo@stpkDhWE8p
zpKVtTiuD{pF46+#C%F(C#@qI=L!&OEljoFug;RT#iHTWi6h{RkZ9TUJx!9VeO`5!F
zSqe>KbM&GB*mZ?$a%WSXDBa;S4X`FFE^bf|=*If->*_ytb#xgSBNlb^AY&T`NT7;V
zI*zyzZa-SzPL+U_UCFb|5>$+jTbY`@v#Aih5tP(?-gEjRSVTYB)?u<LbiCZC$X}+Y
zuu#+3*f<_jv!X{%Ht-d5Fyoy&cgD7nplyip;BHT)Xy$zwEum2*D6AANo2+r^fND9y
zN9nd^r=}|4;o)tL6ey9{&9{K=dr$A5EuI_=<rQ1S_yAE!DgN;w@zss)kC!XfR3qhl
zF)vDt`fjNdpWQUsfTeOp_S*^(q}AsnCQ41257k-qVGAZYxKd(;1l)fVF2ok=7jsmQ
zKg>7p4`){~F#gfqox9SLGE(O_y`CL-p%Fqr<S#FclrLu+yjZQZXsnorRrL6JA7#$6
z!5m~RUrgI0*Cq=`Ox>+b16%9Bvqh9Wy3!}O^;sf*nx(qspY*_hhZJz(!5GI$kS9LW
zuwF{V`>WA3p)?d(0;M*+hkNVa==oyP{ol^)-uY{0xAvUJrq!snm0<h0)N5<&s*6o~
zq*}*zDa_{&Cq9C8aHj{KxvUR9eINV=EOz<kmhluH0%M}M@>xhrX8z_VTWeo|j8gLO
z*>3X&L&V4d`?^rORFB?;)HA2YGMot?0U};zgqWE3F#hQtOii;e8Xj<O?12tPP!M)T
z29NzC0*D)Fxk1xh(BqJ2Q7g<BIpCJeRb5Mfzh(1SG^pNbQF>8!HLo~Zi>OJIXtK&P
zU9i8WN2~KB&gNT(3imlOePb;<FpPtW;0Crh#y{Js8SvUZQYG3QniPDGyI5?G+E}~@
zvDv6vY4KAJTK3mCzQZ6Q5^{0e<Ls2=(-%oQTHsD!KfL*1WqElRK^0i|<E^~id>z5|
zOsx^+@y=w3T@^*4dP-;@OJRiwa<ImB<dgUATuv<3A<i{o<7(b;X?&0$dFWYtaekl%
z+3TR#C|JwRIooX<fyGH(P%pe|8y#qUZg9$O(Sg4?T@U`|(SUJ-`kFmh72_649A-wA
zEXK$fvC{%)6?CNu$@1OjfOt))55M+LHn04(evg{EBFySiFnliU7p9C{=isQCSOMC7
z!o~YEWUlKTbvlfY00#T{kC4lwX=lWR-)^NTs&=%<o6T-4SD=7WKr)7cC*N(QN6TCS
zQ-Rz1LQQ1P*gQuNT<5P!W3MFGv^HRxLJ>8*J|HjwD=D1KqlDKRzOh9KOnQ3&OoU8T
zr~QoQikOSGp|M$O>&<FeD`fz$ij5wtTIW1DkU_0SnyRA10N48i#Rd}_N2%C^gkw#~
z81lz*W_&ECeq<(twahV~6Z5OUz_PI?h8u+O&v$GAOj1OM@t2FTxxJ-M+sP`+8uvBT
zsT$h{Joe+!-PNvjK?KMu;-#{;oWogsbKR$s9KY;4jMK}#Om;@DT0fP7pqUa$eg?m!
zG#Obdo^?g6Yov(<ipD$^82-tG?RRW@oU)phx37@ZzyDxP?EDuN0V#Ws-FR6sv4kV#
z?An&WQN_l=(5}60ZFx(Xq2PAfwr-%wP6B+G_v_cMqh7~5v!g8_zp)D7;ie4FM8n^>
zj5^I#pV(4(vhW0pf{=rb^|P|xb7Z7@H(D-`_Q>_TIGdI3q=E`AC+zFWz`Vy!5`56Y
z(yPOeh|UY7%~pF;ub$+opATqCja;dDQ9lP{URhl&5wBEzut;;QT7$&f?zQ2gf;5P4
z8$QDB&k|8`oT|xD6Fq%{Jl-{ZB*SpXmBMPGm?s;PODW`~E-fuxeDEdaIy*-1tJFbc
z$kjnCMRUy?3G`}(X#2#;;4pQGO#g^xa3kL%9;hgiwBNX1u3v93tk(S`@7PL3iUWay
zwKhYqli5sFS&ZuNe?YDRDJx{1W~v1w^!mrG{976}*A~QZW%(R;wt|C*gXO3|iTei!
z8{Y`UY6PilaN<OcKI|7=23{Tq2J3nxl3ar`=v`b6;}(b#f9rLSEH&CU(VbghRfmRc
z6CuiX)6voRwO`Hl7TcMgh>Ppq2)%W^Skxzja?da@g|M^y_Z5K1jUvXum{ZN#h6aA%
z#p&8c1bx<LEv<~d(F^Xw8|>>ik%j82t@rh7Z0<YGHK%sOqP@`?iIlMDh|b9r@*aMD
z^Nv4?VVUsS?x9|IhA!p|_eB+f@{Q2Y(C4CS^8sg$w+RUedle@bJ=-IRH!{8FG&~ao
zC9l;92KJ3O418p85m(|wICcwilQ}XXBIo`y*^^(4mI;TS&A)vzthGa&u4SjFVE)p+
zRwN8eFfKJih2DvaJNtcp_f`BWSGF3+QkjK#-YG009<xOM;n7a4$8aoESy?HwR*7}J
zU}ga~KIAwDQ0BY|prK~9MVl{GO^5%N2{bPt7xrO?p01_`<VJR0>j)UQq%s&>h!~m(
z{~Kv-Ia1-DKxaoa0Vtr|;718@Si8<osp5(X{RP46A&LkMQCJE_45b6O(N9FuHpk1A
zmOA5(C%JkP_j=e`%Vu``k7vEDm>zZC_?r%TgI2BT!h)vFJzSaA{hWJj)ZB(_c6is^
zeoqm5K}syKCI&J^`)k4?ZipG&ESdB0Ij^9gVY#diVn}YA^NYgYbiu*>Ac=W@L5YZS
zd9$OfNsp}w4#+{j5b8wRU3jx@+9|KB>;>9fvwJnD6R(TM=)H{*vth<H5{D!tRU80#
zQfoUBR;ZjRnPTE)z^*rDKKqEpF#i@8(Srw!!ooA=g2vZ3=pKU~x+hJK1jX^X-?nv9
z1uC-`*5Ddgw3zL}z*L@F6Y-+wzrJw}-*|H8Z?wpY**oZ+@^JVjaWL9T?Ix>eSpd!Y
zeZzz$-M;#%`T5T`$4W^Y=UXyM^uRkmXXWqyO<8PLE)&I_Wx;fV2%ndaPp@&BL+4BP
zCw4-yb9Ul~4+odK5_&#i`@FawW`mf~Wi-*eflJ7l!75BVhGa;J$u@ZGMsXp?q(aG!
zjg4;9;QmU_Sa;~~a<JduxSqcfhA0BmL(-TY)P~Mr$`$vP6V!&a)I?<1S4k`);^K<d
zvixGRd=?`X6Pm4i`ZsCr{!MA;ifDLiUF`UFuCnm!=Qr&FMPlILm>wfNeRwh&_u()w
zy(<S+CEc%u5W*6jJAoIcMQ(BkBpl4_#mLAwj^~Pfz24O6uov8lls_-!T+pyPPH{uv
z6-}Rv8{HmKQmVMRx*|F#oCj-R063Tid(=JPbcF<XzSp^dju0NJy<~#kI~kE&?C7j1
zs08Qs^*#MX9__Rt8l|7K?gx9q$BR=>UuKAUZIMz6db|YyH)stT-Dp?V)<_EO9A-8*
z3nS6Yavcp!MWdtI?TQ~Pt*sN--~0_GB#NMYx(9&65KM^oHr~)G)&6}_l8jNkQ$YX<
zIdlj!1!vwp-7c@6Q$(Xh0Xj&%V;}A*(J4iQPh7_xuPpougj4SO2#<?l*E!ES9ZR5-
ziXwDUFa>=SN$-bP&zT{bEcJW&G8Z~>^<dVumT7%unPmJ=*2He4rlxxAFLx0UMn^~I
zqYY!^fts+!qT#2kchp~~0?#?{D2(FGzAjea{tfo1*I}CtvKd|YKnU?1rfL*!ejO8&
zys?C)1?f8tOLtMWTU^1@PE9cZXWR8ey}7x$^7?-lCAyQ*P-fXfrFQ%@xG$YA0o#ul
z@6Me(w0eforf$2?7A)#-XJ$Ry8W>3Ym`_v`z>zTZ10Q6LcP({7tIp_3bu59xv*Fm!
zKiGSD@rH(nYh?<0+AMX%j23IrQ;Bl$S2@kMKws*B0V3_Je0(YQNcs3Zq71LaZQd)~
zHVR~bKDc2&UY7IrEw0^kUAY_C$VUk*`eIiFHSy@vMXMsI@Vk*e(XertQu#LlR{a(>
zQ{oq{y&s&<-=R&VCz&O0d%B((#)L+(kNYbuOM+|tnR#w2-SG}9#^&bcLzNE=n&{B!
zwf8Y?X5QIKTKW3O6ktSbnm6;pXS&Yux$OB<sjxw`Tj+qEv&kLXAKSrdEhZyzQ@($f
z-W<xqsdbu7yW*Va2)Su!aJj||y-Ut5fAf*5qGE0e0H~m&wUjqzu=KpM4pzfGrJJoQ
zROEPgJXJVC1s@TZf)&AzYHOw{9b|gT^kN9=u9b@#yStJD{#YGbEV$MlHazXzEPJrp
zSENd?RBTy$+k+p)%~SeB#(ae(tWT0-xjJ5}?j<}iIfwGjQ&fwWNq<xI?p8pf6*JO$
z5-r>STDYl}=*M>IJdy;@*9YLohsXA49c=5@(R{V+jCUw8Vo4^7TN?=jZE*jp!bJ4m
zVbeh4reoAbibddtjy`X6l(y&ai4E~6e^=!;4@*qlSkYL&tq}@Zw}F7oOv2{D+xEw{
zomIpfc>o@|tLLPz0p={UjRMIf8Uh!;<B<!kiXtuM9s@sF-!Oqy3bxC>xoNI~GA$^Z
zxq-A@;y}8rXh0BEUl2d#vek`^GfZ#afFkd@8*Gkw*Gl$Z3sK_e?9Oa3H$*VMpg;p0
z1<_~)uB=bC)G}7lNYe3(uZ(SEG)n^gy37qfvt_Tc7)zs_<(NrKk539*flKB-I}hY3
z9Q|?gx@7nRUfWlS{3#Vx)j~Aw(zp~1<=&}+u&<t}4{)mQF9r#En65F{O&1ar6c!@4
zYzIpR29tkGWCL0DTr#c?!$R2i9>VS>ii^>5HJl-u-_22sCpxWX+*=>R`>;PhQ*bLZ
zNl4Cfa}>>#?rLmL)e^7>A?z4b>$whU0kf>QMAvEjjWC)h;8oGR%g$DVvTezR*k$GV
zvF*okm>MT+@w<hk!QSSsyR2QDiYQQgDoL%PiCHaZAM75pFMin`8K#W|<iXC<OFfYU
z1Bh|?LawXi@l_?9N$7ZGt^E}Kc|Q8|HJgA)dYJO8vXN%B@i5|fc~39DgrXIz2!|iv
z)thwWCquGnpVu*nA7#1wz;iRnJf^*;iw}hCZI<<>*V#mQ4LcBJ3YCqWHk+er_&wxT
zLtW(K;_vQXmsrVPC05j$?r1i|Q=*sN(oo1+TS~c9pJus6JUIT%`je+5ce^xycFwMX
zp&s#%KDK=GZ0DTn)3x;t%)<>IT~X+XA7{)fv=Q93HC{aG6JB87s%GIX1M!egRS2T4
zhHnhqKLn7Dg#mdhG-;rG^mRp+L*-&m@|Ur~8E9y}rkCMfH4-?+;KFrnBZ+=D@@%Hj
zz1Y?Njau}n@6Axe<~h`3FaOX3m14AUHRF-&T$eyv5tr4U$49bLghlUW>kPbW(p7!9
zMqfDcmaR^+vz~?PC*eNNeUPE^5Sv<mzDju2dU=kDjaiQmq*Cd1B-t+%<XIx9?xXaP
zX_;a{!}@J*bA2vIzJU&Tpo~RiT!p;pWx6g+c|E-$EIQ{A56uS5H1Lp@`SsPn7l04{
z)pf9R&kY^22x*coIS=R|xy9bqUnqb2IH_`ijt0y(d%}K^SN=r&Dl@6Ef#G%YC#V@X
z2%9z@KhNy#qE8KO6<y(;-9z2AOT8x(V$?ap$=iHs?Nrrmm_0sMjm}o(VnypV65`g>
zee?x3HTmk+HAn2?jl(3T0s}B|)^F@?i6%>t@RpT{wPR7oC;qs&1U;2XtC9#DR|8lR
zK#6YV0iaH#Kdy_9)XTWqxQ*mEuA#7trpOAomctBhNr{|VcEJWbZtz*kXkI6BTyK)s
z{M)Ht6rrI1_q>dw2qhNV9#<xgi>CK+DO9<7mlM`}2GyGyHOo1wy0D|q60LUHH<?TU
zS!VIF4j)Mr#1gCpa?VW8kG;$6w<F~@pH_JLKQ-BGAJ)60=a<QYP<$8-JcO82V_i2?
z@$MNqiN04X%(IO<Rx;ZAi@dAX>!?ABgs<2z(L1|PCzBlE$IY6BXWDA_#stex!)nbN
zS^sYGlGstLbV)Iu7Id)onGoflj?rMg7IU_!*s1F_2@41a_G~j&aAqsth874^=rCs7
zr4leqGb-N+21LBp0P8AtU|!3}8YIw+m5=(S;`^o4B8IBmJwT%B>MApK<^6VFy(rf?
zk6h0c20h{6ImYf)o?Hdd9mSXrk;3Zl)T}2~)I1v67uZ;Xm_Ek?tmdX8^GU=mf$@lP
zJIJLn3R-cL_q9h0xN57{Wef-hC^UNA=7mDEg^8|b^#qt}x(4ar`MLRm6^Aso0_RRR
z$4g)C^|S)#8Q+_MX(;Xjoj=vF6l0fz4hFF5ipDBW)U*AZaZ2sW2PhWnpNLA$l51<i
zE+)O-r|HJ!N1C^X5KgJQ4_L|>X`WihNZVtX9y?PTTHmRydEVh<Kr$3R*U269TS*<f
zyvw()bBf!)or0xF=)v3wQ~iR$!OQ2DuiBrM^s~Btoz%Zr-1QUb>##)L+0@JJJR7f0
zXJv)F(xRN;j4k=hoRzY+UEmFNk~|xBx9tea^MJ|J>O5(Rvm7ZsI36+)vD4b6)3K2*
zBIow@6NI*yQO!=3KLXVv*)3nKX|%4kSG?I0ory}2UH|@8cuNNpSS!~Kb#EDt@Kcjx
zx^_g2R()8JyyW@ok8>>_GTHQ*-X!J4RoF|!^-al?k(oWhz@V19S_|M<SuJ<sEjuJv
zMHr`B#BOD=Gc|r>93Bg;*@l0A??02o;8lGv)>pnh?)<>EEm38&4s$E~vMV6jtT|QS
zw1%27q{6{S#e`y%DeK9ORtg;|Q<4fSU-0AhitWro$7pfI*?>pAgQ4K>_XG`n5k~7e
zYTv3srW@a@HK}>71JO_S8N6d@N0}(g4r9PXN<E!5IZPTIu_N--ljz>73Or-DdIbDp
zvxixa2X^kwyzNub#Q7_QR*P}IuI5FFi9klkW?0jUZUtV%?&9Q`;{@m(+snH)H-A{o
zUr^3Z%NGqS+Y~9*hF)P|fgOGW-gStQk4I{MN?@C4X?iMogtr?tM3&lx>iCN49_t@L
z!4rBb0$rUrYr&LSs~!G1tSj(^FTEGt@x-5$M4|>jrgkbGPX+gMKk6Oh-bw?^mJ(|Q
zUps2Z5&S-0Dhh1Q)`pvCq104vPmx#hYx7-)u)a%d#iboZ%(<OV#2%t2-H<E^q`vwL
zaxc?|d9bt|I|e$>03|nK`a9cYmlAth;}Enx@swbXOY9)lD=||~oVX`NuX>S^j<r3s
z`ZTmgWE=mAZ*pwqiRpjG%_NNvf0i{^`i!%G;7pUgKRdegIyRQ0e3dZDpS8I4<;p#5
z7e%zU#|*i#bERBygLF+bicHVq1EWU11i=<)nA`qsrKIpllbv;D;@iMkcj0uWGpD9u
z+q3$M6|?SXTfR=sU)s^br9}fWz-v>Ms|ag{)#A}dB>lly*YdC>p=Py>5&81aTrfGQ
z$e5t9xl8p%o^iD}*lPjk!vy~LE-auf^ySxY>p4+y&4M57X${NNBEs$@C@(d?I5#du
z@Dq`T5fazMm7uG+5$V2#na1Ip?U~shROi7Wl9yHYn<R{}MP^$u4;^)N;wQcMqY5jO
z3oQLOiw%`!AfL=FfnMTRk!AW>Ek+5lQC^uJvtpP=>g`89z0~)Gv|3YkGr!1>ODP&I
zp7AH2%L!9tOYPcPZ|PgKA*Sfn+Yu?wIvie*^^gFaVcNC63PB#jW7P#I#+uyHF7X}Y
zJxgUjxd|{=D=xGLUQYW*j~8Xob!;U4*v@_6d=pq(Bje2PB$y~eX1%|0rs1f~Zr!D%
z3lQd`Pp!+h(h$1)Ypa5^K+iin+~eBx<m3II9?BT$w@zHu2e!RxLX-NY!33v=jFAL~
zc}29tT^at}Z2WTGYHGQj4(qLP1cy?Foe89qtrkmA`P)rhryxH;6+p6N3U6E<m-U_X
zFR4NszP#@Nqo0U8KRr-NxKzz)^<@`-HP)hq_&W0{Yflxi(C*0)H4?W>-6W@)sRKK_
z1T~+(>n>7Vae*|+Ild1LWfq%NE^09D?=BmVn%2AJO1Aa+U}$)~9&9@KCZM$*Yj(Q2
z?!rTSTgc*JB(S`fYrk`EL8J7|CmIpDqP0&b7pjJLOh8<(>xbOLuSPtnd#n0OxGM}E
z9pL+=7W+%BVQ=<V?}pccRh?O924s(fBjW_*2nGfwc7I5?%DnP|lJ4y`f0~#OLDlh&
zHp_C)n?_u{Ehp;@mKiV^f2X?kXz(@VC6mQ@yGqHYGg@4DCya-l5x$_iH)^p6Ol!Ho
zH5sZ)gRovI1kFS!5e^eOF4}Gqu#BBzvMjAHDp1(4iV0Oo`i5IpJhq+9w=`QA6FVYa
zDW8wWXjXc%nJ;F!-l9=q@DV1wuL~X)UCHaJZ-y+x0|~WuJCcfJB=>x`A=Tv?K+hw5
zW2|eF6F=;-^6Vqc1>?4R3^De}!43>ygJK$!0j?D`Tf<P~p7NO-Yq1<&(pMcQeVmP|
zm=nL^{AWz<*OA>X{9K#zB6|JD@e9UsnM+1cX-3BR-(}wvlU=@?faT*ft=-3}l}9zm
z+qJQJ)T9m@rlH6hn<YPrpMq1!P-&m#sEEHKG+48%Iqf$rr(0!oO77b;rd%{{MNxU0
zMqWv#r(`q=FR1~}c2@6apE?q<%w{UUyvkIU!nwz-;o~i6+2H2#*N>YC-e6yco4r|d
zdWeIEn$O&}AwL^TnM5wKuJL3IwIGLBAL_O|KKyC0M*~XJ(Fi!a)1FrRfMaS}(c#UK
z6)}|~h$U_oxx8v*4SYSjX+z`W&rLQ|d8@X9Or^j3@j$@E;Y%Sdqd=Aq$C(_rb)ZgG
zrl))$sLIU_)PA9^2N|K2#I*)T{SH^(NDr%cnW6M4_Vb3VFAo|K_&AWHf&M&en+3V_
z#O!k$-XrVJaVXzoPlnsik+A*;01@xw%B|cp%Q{3(f^~60Ke*BOA)gq{EQPkN<$I9p
zH!kx%iIPq(WM<N7j_BcY-+2E@rlh&^1M#8CfFj#k`Lv#v{Sb`|j@vTOCx1P~UAx3K
z@xj@dnu&B8bMrby-ZTEqfk4R}Frcqj;6e8}Xe&=RI%H0FNxa^A<HbN4N$Vn~A3z#^
zw|4U8z%hJyeI8ZBKj9dC`-2g=hnTaVEY_rV02c35Ym1rK9!(GtSAqTaFfFztN7jmf
zSF$bsbgU{lrge|=er_#YN_z!=<D5?3c#TY2-)k9{4Pa5p=(EGzZq_i$^DmbI(}HIF
zN?^o(TpJIY0yg_9A3xfXuh-n8RVj6A#)T(Q>f$-vMaL19I{xdY{8G>Yc}Fm~5G8iN
zkOf=2Eo@CAzOTle4kJqR0XI9KWipS6GlYuHSLTJDDM;9S3p)%y8!aG2Lzmvggi_MP
zY%D%5{{xN*c!K<sW7hyq8}-bXPfK8%-dks+idL8BG%F|BKZK;xhi6B;2R~$cl`O6Q
z+DCM+#1s?H&o|2^e&a0t4!N9yMbMrm%GOlb7^xkwcf$k`hQmCuML-kpUp1QZ(}79G
z<s9(Xy3tvz5VWZu)`&b6xqfuj!Xl>@%bguh|Ar28`(Px_wB{JYK&S1C9cQ;<_gTe<
zU9uqPkJ{S~so<U+C+Y7X1?f<?$J*o7Yc)Vn?d=9XqvNz1xc=P;SJU%Gq4$zr6Rm4L
zcP*>bP>&qG*MqXEIfc4Mve()#1H--f22(SP!B6f$fuRGx(qo28f8LY|yUqtC)Y`bB
z90r$gXK(XH$2~utcrbl1W-;rn_L8mjQ>3|!=;^iztUt-xB9M!BS%}Em(#u2bQ>J$Z
zI^nJz4?<oy_JJ*k2``=q22do<k76%QT0^(D0<}m;L*1h=5Sx$ao?k!x2gWlCYk}cP
z(T8Np4X7*okd76<bcHV2_UQ=^w~rWY6^iwJQGrnk7gQ)1*txa=P_EURvDSyveS;wh
z;hgT1ChywkTe0Cu7%jo9gUlN2n>>TlwB6SYW2(Q$z(>-iE<|J9Aan_uaw1@IrJkm{
zt9i19aN_+me{$$mG9Zd_Ue6FGn$Qb8x0wgl*jV2JH6qy^(NUb`LGDIYR;_?Z*Fs6Z
zV#YwQ-<jH^eB5!VoYb_Ao&uUIC_G4coIkrY9XQO>w>}lilVN%P9_)q!xSE%;;Icwc
zStvFm=VU$7!x!!YEC~b=@u^lwe+`uvUu&EmW_%`;4H=IsAi?3yHB9=(NP4&#iBoai
zKoItmtAQZlOo11!jczMHT;Cdm*WcU<&31NHdf5z<7MjaQlkrBqlOD5R8pipSuA!Ax
zDF!s8DVOgbeJ?m$nDL{))_!C?*}1-S@}%)JRA!?+sMW*4`Zl4vilOFP)n?-LvD~%x
zCe^_ZvE5voyHmnfqZ*Y;H&jeg6-_VGS061BCg%_02$OQX-=zqYv6xE9Qu-!3FDc*~
zYG>9l5A-Y@{7*vxHnO5uTy!m;=;eBg+oqM_heVCU&ECly5M*^<cMuD`wUe+S8p2})
zn^9~7z<F7*G9k@G-PUK-^Ul`K8UZ`!0XU%g!ifeZ+gqUK2#AXQ)!W-N9EoRAE&54{
zu3KQ!=#~`j(vs+U&h%-MZ@Yz~BM1W6l8<hdUOTbLA0Z`O@F)-z9WsN6A;;fu9Q{<<
z)EJOW+MwVgKx!oy4SsT}wedJMyjnsH%Khmn<pOW<rJswOp6te0ru9>TYaw7A%~ruh
z_m|JgiE?e6fw7!Z_Btr{H9SN21@2n#gVOWCxXjup8m7b|YaH(U5)I<@4b$TZp6^tz
zq}G$a85r`~HKh{#WZbd<EHABuub&EZX@}3I&dlcVIxf}d1wc|&SC&iRb(f5JwfOQA
zrxS&qp3P_acTFZfSLsxB?8PVka59{U?=jNLo8R3QpPw0?%3QY0qSHnI?^EoWJ|6w_
z)?4DIz*Ks?6!DBcOT5?Si%bC{d6JDOnI*e;CN|f3tm~ii5U0kVLI!`N4e8eBC=)j6
z;!3x66qU7}j8W#N*jaNBQDw1{);OEuK%5;daX07)gMd<ZR+v^JdC`ee@={HZHKM1?
z%3EeXs{ujXh0=1iDR!b5<h**FFnHBTO)p^g&^Kc563wqR)QGp0f~IK+J(&c(tmy?r
z(gnw3Dix>R#MB3FF#uA{JUQA3C1t=N4UTp+!SUHhPd=P#5hZ89q`uI$OL{K{p>#?6
zLvLTAL4LaCW*9TaW6_5s*Yi%Z8|~vPAzZRv7xW9U;6r;sZM~gW$<<=Mse3PWWS>~V
zXlXdPpiW7+z%?+I-ReQ%!ybHpHE74yz>k6iftw!DCgaWQe2uu+arE!7Yig4&SV}Zw
zCL-&K##{^JVP8}ooqPMl<cJ1!+00jZim6MG?k#i2MT98z!O}wTX~YKN2KeG#Tb;AX
z*c+$DkA+@cKXs!*-VIbE4>g0QWq4w17b4K<f?8HY&EL#spvgh5*FrZ8qFrkgqUcaV
zSmYZ)miUcpV}&a^=A6d8j<U$FkY2+*tCFokVF#`go`rIZgRJ%kzBD6t-@wZ>qTCw-
zcj4Y;n1)!@c<I+T0-mL18|?dQ+#vL}`y8z6p9iO1QWmYPI&c7y=N*TJu75WQ21MZn
zR4v|vat2bJwSg7E#&aLYtC(rMrMaeh%(UWUlZC!ps?iY&*DC)i+a^P&Jl!$e*QY%6
z7$GfLvbbA4FD`yk+6^KO<r2lto;Wn57@X2vt%)J)K#g)&mU~aOz36?uVY)39pW3&W
zVEh2jr<77@K{I}bfN6vai~Q5SurR$bH<e7`%WUk*tI;Z?mdu&Tq)xx_k@arGkB?tE
zq>2>q>of@`E9L^bpVdob2@l}gN=bml9QjM4Pp=k(>p)b%?|Zp6KI~#GtfOQ-O1T_p
zNnnc8eL{{90kRWahtWYn!RX%LyJux!y%O6)$+-0EXv}tho{>8pwWQVnvDK)w*ST$-
zz1)}x)V%Tb(I9^o9*dhGpD9^;6w!X{$*kBHuKoE7cd@3;h$6m6J8X~_EnR**70O~&
zn7YyM{nQ5TP}Clk)&Q%bPr%wAg7&ChtPZxAE8?7c_0Cia($w^GoYS^$REr_#!)*1k
zM(e7M&JhvzVa(054c5Ek(@hG*({kiq4vF@U(18D5%qzcCyl+_5+ELIjzX(kTxIavP
z<bukEhY#J0-T2j)_Dyk(z-Lgw3dvXgM8VKyCKumi3vH4labGO2pV^dXbGNsRc;WK(
z1@??<AB`n3jO*pA6!9FQ2HgZd;CP{*bw=);c~d#z%FK%03~^MNj!yMj`+N&2Cr><e
zedcBdlN9QbMMyfOs?yj>!hXMhu+ZZeLCL*Zl)0(RjYCCD!gw?@R|9^aw9ldV+-LI}
zXmj}albRDBnUb_r5hoW59e{6#>(Mq_a6o$-JT|eqq*2IQf0%S74V|L@GDh<|a%rH=
zDl0Rs{P|6e^`>DsT#vgDx*@=QwWwmN*NO|$T$`dc@*9&B2@V?f+bvwkuALiJycZ<V
ziv7M#L?f&G7cj4>I_<@)dcl8P87|pNws9^Z<}-HwY<7I!%zkfc>q%C3Jh4r`$TnlB
zH9}7Bg_`Toc(6=+{i(25Qgt{E0q0~<OYb`P`|*iV*5<r;JH|x2gwOP7bdEf09-Rn<
z-cvjpxSf?33-3u`G(2Z^bx-%N@vb{xc;dW0BZrd!1t=8I+i$8<7PZwZ*%YFrO678q
ztMS(oao0H_@Ww^;=gsU$Q1(4`>WdOi@%mCq3!Ro*uI*gA<*dJgR|Dsdnjc3!g^*F~
zgJ{Lj(OSfJMI8G~mSiZl$=w*^KAFcmiQ7g7-oGApQOk5{mfe<L{t~OR<tGw{+HZRH
zdmvSs<V&ue+|2A)DO4O{nWHCn*yek!CAmm{^qH#4d&p14jy>bqN5oF-ranjt5O6X7
zrdZ=~j5p=M5akd@Iiu@gh<|$7R81M0O@+$<S?Ex81M+r0;qSl<WR7nDy3)xuS_K1G
zfL<Mj6*toWCSy3TYe${hlEnphA3^xWCXQ4er^kx>C+QoC782-0FinS&l=|NCQw?|;
z8lFVI2i;{B=ovT(&!#0BU=3Up2&!n9xF1i7Nt99lf;JuNj<>(_vrN&1B^{2dmpDrG
zVs2bN%lACEs`fnWx!b(X)C(ytq8bD~MKwh=70PtNx{>bm=Gv#r#2AE6E0DX!#B;l*
zXFE5O46(BL&5k#q=Z($#1?bgi&tI!H9gZ4o_66*L9gJe=SOw&pCY>+y?bC*8DeW72
zr(t-}gAw<ANj21rpxycl;g5ZMcIrI3fAkks99nF&2gk_2By4#0Z`i8_7H5jP*>n*S
zR`mGDnd&ysGyMMW_0tGh7{)v6w%8{D0=r9Ka@mF8InSu*Q$~6>+o>sT$jbD#-7OI7
zK$*85Z;{oIiP&?E!w->&FTbotYk&`9{#wU}dal4iXk<Zmk4cDQp~uzdgReMg&Vy(B
zE-Ce;9zbZo=jWlK3z&)jvh32tpJkUge>8Pl?^QfNEDSVhi?85)OB;&G=p0QdMjW=<
zKF!n9i*}*B)`uOgk`k793{l=5WhbLI>6&UIeD=fgR_lIDre%tjSk5<z^1wf6`$~p0
z96OVK*3%Dk^ELf$)UTUGj6GU5+VzIq5I>eAkY%hR`+2L8t+{5HA<NILRHccJ^X#mz
z3*(%O5*;S9l#yWs$m63jEsy>(A*-<n1K%9u>%6f84J`Adv4uZYMpGR)M;x38x>w3e
z<Nfbu3oDt>myM}5)y{%mgkvojp8X`z6OGs^q`kMStXctU=gtF+O!N(<R)>1TtMbK^
z%9KDN?@+o1E-OE;dkp~J3|+Hvmc85lCS0JYj~f;}SZd1@^7`6`O7O1|4^o{*R-0O;
zq!~BwMM3(5e0kM@LOYJ}B~_IP7k^>q0pU;&_4f`(H`y>9p~PI9JgG|$->!Q;lXz?K
zwWp~dU1E#Qmi?@kL?F{hoUm_3A)IqyIFrQC>a6q*9cq99npOj3cyygnMd0ZhDXJK&
zKkVZPp5l41&OVT#`e_-Ut_c%J21U-CyVDvUzDZ&$di^OGw2_e_?xDf_<*cpQYP9qb
z&eLq`ad^QEcp-VGkc-|bePJm^1CjP|d1=}p3(2U*@i?cFbt0gr?1m5fbu$EQS2N2&
zIwoDOy|=xdmW%m7mP0W_wL)UP;qZT|0C4s%Yn7%AA+-u;OgkYBAC0`hJyx<%JJ~Z7
zSv%x@y+EqK%~8%TBD4NzzVv&NR*O09;_(Ovb@XYjavoyS^PBjs>ulhu9L6mawdU!d
zz3gDdZ2PTovrmThyfaa1UL232es~>TXpMHO<hdGferOU3uOJuw3a9y?0bV>}sFL)Z
z@@n8u{vvj!S^|OOrDyX#F3?-u&0Ax$7dM@qNh5bNej|QA4?)gwC?jwk81_nG+yWh}
ze}%utC*yqce#G5Y$u1e@d&(73xg~}|wc~7yx=9z2Cw^@iyCNP(g=>TMfm?aeTf&>L
z=8{>2l|Z8iOVs<=Dv8UI(q`88WQWMh?O8_~Bb~>cd!}}FOUmJi5IRRa2HzpgUzMEt
zLa){nbJ!3`Z2lZ?HtP6BNM18U=D9N|o3p#{-w(s3!Ag2+<GJWkkq0+C3L1Abp)7q!
z!C75WZsM5ktG4QRFqm)d3sGn#neuzG16|7)5Cf|^^&6Nw0?zb}VnlbSMt&aX<(Wc0
z;337!<&IVxHWrcFr=NEk`dG4Sg-1!Or49mTTi;Y4b+2K$k-_|r60DyEVcfu_{`X&r
zMHu$$DH_Ji%*=<#ev_%jA|;$`C$Fs^#%}G0+dl>yIw-7`y0@Yyg%3k<C~A&7HK2j#
z$LmHvO(@bmou!V_`=%V1m(Dxl-vMTa3Q9^!%BXfn8(I#1+!0F2)}jI=lzD+G<wy^$
z^xh{ptMVjz)ee0PwN4jD18lP3+RNFE@{IWxFjn_rffQHI;fD1)|DlD{LE^T!>^Y|7
zuH6kO>3m|PSX5iEISUJ5KqlML*=^UHfwW(g+lJ|*qfKhqqwX66`EE@M1>xy6<<$TZ
zuyS$vj+ZY}ChFUJ!ZXRVPbFi?EJoW;UxB<{8kdSVflg;kAPy+Emv(gTYrqnahNISr
ztXhQ}s&{b>kAV0!;r~%%)ap|=l-E<d&ff!fy_!7NOPx;rMaO)MWu|RvoR`_g%3R>Z
zXS%e99F+^1yMZHF^jHz?uBS=k%X_GpF;`11E_Xp6rhL=OhC-e3^x3u7m+Nn8W6ydo
z;oF@i#vgmFttF=d_e+p!R7t+oqI%&8OKEnfGv!6dhXJ=p_{*gtd`#?n;u!z%6|9YJ
zQ_*|IpZgGq;LQ+;awiA%?}l?$BUnmmC2MXC+cAI`6bjr@H!d~sO0v}tKI=A5)QYc0
zhf2+YR!E7NXe3P&gM`jCeVEcrN3>b0J(P0d_z2@{<|{P_2nf1%kJp-g7Y(yq`%ZB}
zacKUTA(aQgA)blb>&d5WVd3GmuHU2#=NX_?bGs)6c8Wa4u(A~^uS@pAUD;I+#0lOQ
z47TRU@SO+Td_c(NVZ25`7yM6D9VC7m)0#~-D|h6fHQha2Y>)j>G8HhOH1;@PJh<f}
zUEvyMQ)+4Q!w)qt#Hy5G+Z3bHDDobp!yNsAO3_>R`<$-L*Du?7N>$yt{z>6&RreMM
zP?q<tY|YHXo95>Vc6OXS)U;v@?PtmFs4=L+HA3MXZhWit<aCV_lkAh8d<m-Fam7l)
zFQ*mtQhZDtF04k5K$X@bI7@C%xoHDq^hmx|3xWj7<ftY;rkGCW933q>>?;~Oe{cWZ
z8LL-PV?yP}3lrq1C^UWiA6tX|dVZTdC!0ZY^|Y%?;ShnaadA<HLZNoHw(*lLV~^wI
zrHP4&hpKzh4!KJ?CVjQ2((B+VzH}QN$LWyo&+5v*Lne-hz)|Hr&nf*{GwXJ%4^|!S
zE?LY73JQ)+O_lWbtNKFczV8szsu?edg?nu&vu}AYcGNOZ(UHsiqd}p+T*@#WAY%y#
zD5d_%#XnzW5<xr&k{9w%|A{bvz83T`=@66@<bq+>-4WFGx-GD`Z<BPnwu=V?y(!43
zN>@mly~g<8W0N-bFqgG;uyki*HMMkdw{(2zVCn2->E>l?>CNimWc}Y!2Rl1EFFQN+
z)sc^v=l7AFll#x{YUASMrRLz~;pFDz;9%$Fqh{x1=iueSpl1K?sQf?MJl#Fa-Ox>O
zw6VLA-|4#h(6?Dx{Y^Q4M4|q_kN;oUZa%`nxPgJ8Vs1gLrbVq!{nvJN4g(MU;Y5G(
z(Vwe-kN?+-l+f1J{OixFUwlA+ME~Iy%zr&kY~k$4YVP9VV9Dz0Y3^q3<YDV%iQ&q|
z&d$#vgtnWPR{UICLhd%^ZnjR=uY?46CD}P;B&1pRr6svpxMijJSR^I*Iat^wxj3aI
zr8&4{WcmLa)bfAI_OJf0Ztf#*Y5vO6jqSf12mVw4=i~SX{{TNgz`y6<<KV-f_WAD;
z^6$3)!~LI&o!ZgE*3nXggI7R+pM!&6fQMCpgM*u$hlhv%HU}pswX%Ywgr=0dg02j!
zkGY438|y#wR}mFy32qf_cTN>)Nn1lZYoH3y186QM>uavZ<76o9&ZBCt;H_e>E2$~P
z&aEk<z^<&VDW@rI;h|wK!>yvl?yX^`;;Sj6D5IjSDywN{eOrK!S_@sNvcca~`u_kT
zZvU@V{QoA~zxx0GRP_H({GXGXot=aG-}^rYA0IdG{~!PV-`FD5Rpkh7lV7<5fr7j=
zI&Nd1V_;xO;$Fpj5Wfr)21d~D{}Rh*YHQY8-&p5s=N=YU8-~;_l>Uu45d8gl-3R@}
zv`-o^NS%rIHXLpU^Yru_^JKB#t!#f}p`yH-Bu*rW_h!y7Rdg5Y8$K6%+dGWZw&R0d
zDz5qrznxUSpBJk$G-L3?s|qoKe*TE3``xg6z$=b`L5=u*UVQ!L@5_UCeqYQX`F)UN
zyY?i;8^G@aF7)4)_D@g#(<=T8_5MR?e^>9nFYSLsw7*LGFNpS6Y5$8w`<K%G7mM~k
zX&8T3FFrS;%I{-c?q+jzE@;wH&Mu89(e5U}>i~Bbmw3oYH`asKopj-9c*jvW3J#6}
zP=Ffq)h|Nwn_Eu8rLrxPtOg&FiswnbXUq6Z^Mg6M@K4jMm4QF!*u0<x<0Uq?@%iZ^
z4tA|(5{WpH{^dH|UO*K}=PCbO2lXvTsnxvd&Ne0N_a!Vay67kq=S_uXQez~#46=2y
zZr^_B!cHIhYFnF;w@a;Jsl)E-O1u3%Q>Jpa*)L;TuF#J~SAvq7g^o2E6l~OBg~{sL
z0?iL2a9j1$UjJ4QekkHAG-Dz<CsCszWpMg|s`D-M*<1QvWP}V~-*q!B^}c>GH1hk-
z&m|q4Kp;m=y0JjaA0>yq)y(^NpFL6A!A)cd@f(Brn_R6V*w#0$F5eOiURR+0qioWa
z$`m?`E$kvbtgG9?Sb$eO8HCw!GdT$Lr<rZ+KWRJ<JY4fA&0@2zy2Pw)Xhi7KSFrKX
zK-YfV3oQ>sI|<=7Cw`{jK@u@Xfa9*=(ZxafTB!~FZ?g!x|G*bT*TIJG<@ul_@|%Fz
zLwmmr))!k}T>+w!pL!Je9+#EZ4ctl=^2AQX$Trim7b3&CO|NGi;PY1#1u9`;t93df
z?+JF7+V@i8))4IGC++Uw82eUSO^RHxhM(?d4yG}CF?HeNy6=4@e(OM1dA0WA*paSO
zs3C63(aa~UY|1VT=kl?x=0BYplKd6nWhw$?@kK?Ky$#vssIOPtjDY^R_>4E1a7Np@
z(8;>t{hu>$6#XjR1(3=pLTe@7nEx?%PTy9o3L}pb0Y=QBmYsl|e$AXQZa$LLhuf2Y
zWOGkDk^uuQsb>X}I)?ir@nMgiN_(7Gj?ieHn-oaeY}&;Ytpr$;I%O)kZLW{<XH%0z
z+?(q7`d%cxnvVw;x~vTo7CgyS&``M8dEzJ`KW(V%c_Mz{eA*|X+tv=+?`=p8yKSQ#
z<4JH+ZXZTvyWlGS+g;40trby|Mp7WP1O=kxzGhEd?L;M*!28a)7PH3v<uucs$?-wY
znO??vmOJryPZK3daJ-9mAit0{eAW0w?}ce?eGQJz2DS{Q%V~D*SY22cID9z$c=D_G
z4juU1b0od6$#F73#aVnJhHs#Sx=lg<5x;By8#q%`%h~QKA!Ka!^6P0F+g-w{$NZkM
z-TTQUhin(;Kw1i<!Ph@w>Xx9^=Js~MZkn&3ze3-V^O{R5F^LGYO>vROEyAz`QSeFu
zHmq%E*?BV9lMOkyiu!K<FvDdQ^}yOQmBaHQ5b_*gadc$V{Ob#z8*@_4TX=Fo`*YDF
zbvxMi>D2IVDI#t|vlq$LF93*__1~wz^ab`~^EI7CPHqDw)vjQLeJ~Z)!dx&keq{X3
zE=I<SZ*Pxx#l#b&`g88tiZ2WK+ON-(BQKX?EWWZn=A%Wmwd{o2g}&p5G;QE#Iz^r*
z;M~@!=o)$eie}_~zOfmBQ%A*KFxU{zcK3KQ1Il=5_J!l8C><@+RC_le{<vP9d&91y
zx3VfPsDx#*@~OSMYt-A=mi6=ehf={rV1s()*~8Jf?J(O{PXi(B^5YyK75?y+UCxro
zOp*6RaR%?UcC9<-Ie#Ny+mC3d6;EU;BJNUYSU<Hx12<DIC~9od4sQ7HYq>A%3!iA;
zls!50Y`Nav%_(4Y<_<nrsC4hK@aOtz->;H1Z*=LO$4VNtuEkJY`Zj}p+Itnv04J9s
zoWc}d)5)At@!RrSPN15XBQnVjzx&1Qo>VM->o4YGX=Iy}vXip+zH2P{Y|MkeW%WYN
zR4m!aCtI)DdugwS3W{K&k{VXke^l*l1UT~}Bs5)m+}B#xv>lE-s^<wGQ43gH^y@#T
zUNh-?1^A8Z;_J0atgHc;tSFQCNHrMXw0KN`X>P)}cER}aC_p1MJbaK(KfQ!yZ`=WR
z9-h+M@g%@|;K*9j+B<)4;i0RW)U?n%vzBml0s~<SWNErzfZJ~ld6cE2l%2guTg!vE
zH6z?tZ}HOdGzCm0<iMeGKth#20AZ$uh#Yhd#O)7oVLOjF-+*+_0HgMkcqjsj&F_e%
zyQuuOj$iSxyQU%aG-G|;{mK&~lC`JK=f5<eEBZ1+t2B<CKHlD@-^2rVakZYa2_C7)
zz}kuZ8;%S{jB~?@xd{-*>`8?a&1XwHNn)<cBwBY<gw8zhQ%KdCy`o#3_WH7pkeOMl
z$3l{We~?>ttJbtuCjCgamfrw=sR^H<E6l$okJP_UlJbLnfWeuM(E6!169>m=>8>Y?
zTi1w8N4Mzy&*Z3sL(ank-Xh|B1LmG?V-e8J2=n-kr&SS4&fa#LM8TtyqLbCr(Mc;C
zEV)a)_Vpj{Wb5+?9sGdQ%89XQRyjWSjb0P=`>*3*nd=+DLYO%bt2wauuV!dRz?qNl
zZ*M%`_CEKnIXn(#&kpJ*oVqwKG4uR5>4ug2i0yOb6aiJMx6k7C(qc?Dl&BMs#D7#W
zwZAR46X5AB#btE=4`?@ToeMxUBjNz2L*L}}4i49i??ZpxZae@-J!(3H?gODAlk_+}
z8qk5s7XGAu4~g=qQt5eEwT~e4q^H5%*;_|4d5&Wgt<c^Yzbr>&PoI_+sP~D-@&oR(
z63gl}O5f8-Ecmpj(C5kre_~83aA1mjqdd^9b=m3Qduc`g3uU%Xj1_vtqlO%|n}-pc
z)dGm|o{{MWerR!CW23}mwxX9oa!ZeRUk>GfgWbb^k#b1gi)u)lzqrQEow(JOccjPR
z?w$Qn$DAVM`fuV;OWTJA;MGSij4Zzeq(00V<KYI72>~nPT-Bw1lBhW$2Wett0m-B*
zmQr0nqerk@8=|l8*iD3IXJn>3Sq&W-#`)E)$%ivco6p|l&a)(?o&l|P;8`t7y8Vr;
zlcmpQ=v5)2)W?;Ib%<(A?>_-wRDJ5W9CXO5YD-SKY0=#*lkJF9AA*BI93PL3EZ$~P
zG0^0t13DZV7)$Ct|527`>uHeMUf6F0#@owNYs7@xkE3>*guKV?d#ky<o#}qZvhRSi
zUTIXia=DH-zB=EyN2VpRhMH~qb-7AZDTC1X%k#(Y8))i!UbK{6)Y()j%`-Mx$jev+
z7i$R$o`&gDq$_Q*R&zEBgDbjT$nOo({9$WN3O|eZa`qtdMI3r6Y1ZVMi++<R=1Vq!
z)yQiE?_K+g(+0*v#YJaRH?@X6t;G-F*+y=m)~&jEi-+4h<XS729$iaIGfm1A{H#0U
zO(~!g`OTP)7nFzLN_v;@v@!4K*4{1I;jit-X$N`DLZa)V!q9bxFFe1IF^FCnnyCj)
zAy1np=;ofx$#~Wp^})%9isPtd?JJvzvBv6l%7+E`JTxdv4*;&WbzQM1r2N2d|DMCH
zP}HR_It@veIy0~GvQ`<i+k>=Q8>mis!GUcU(V3slW;8vluWhVXn$!UFJ;4F`cqcMZ
zQBos9GdUEp_;-+U9VP|_se0oxOvlkNdxyM9fDZ>C%FI%0ygx+hzW7V{8{MY%%_zW0
zdDX!abymQh)`Qg#MBt4X>!*-8ib-J-!Rh9P$Qe0NF%3&z@m2U(s!uHs&QUV#kQ~O6
zqH{3CWmK2y-3#%3Loc2GJgg%e{o{L*2D5{E6$O><F|RN&Ei`lKQ|SSYeQaz6l&{ih
z{&1KifU`3=tOHnH>2=%Do=uBMxn5|zq&JK&;Bd*))Y{rxsHmP%=-?2i#Ru9!8mN<8
zP+Z%g71b28S;ZgI?eB_YnXd5L{T2Mk@wgr84m%DMdf6OUEj#q8r;2GmyJf93<!*9{
zZ8^w9QD&@i%NcO8x`B+{=$XIV+cmm<kThE~K29mLkHng9&HvmA{xDlXe-~(obA=7+
z-u|DzKPjapz)!?W%FR_BIrDIHPHnx%;u(3n)Z*EVBIyQS+^3@~G8@D=1gKU{+mA)c
z+}t`IfzRJ5KU!y$4#|}qnqDl~EE7)q(ZEkAmZMXwBeOw^qY~=VliM^f2CJI7gHXDA
zgzI9b*3`$6RKSD740O2X`_a_xZF*7DcNttuUq@XhP|K{7C)l_t)+Fu^=T=#VzLRZx
z7d}U2gI(GXPka>++7L5~xSuJEe<L;qY29BG`mqS~FLqiHpVE70g5VE9$G|s2{7DRA
zD*$bP!GHMG@T$f9!4I0YfWHF3hbqWdo~HR+%0DR&`;3i8T0Ng=-Y|pB?Y00HAlKDN
z_LkXWfVrn?khP7B($jLxiN8z>)1`TN5F8HU7gThThh~psY~hIfc83i+UEY?aQ-P-l
zB{~4*gAei{hy=NJdih1#48&u^eU=%NcYjxBW-2;$q-o-=V>~!<hpsH22l?#Cd+nTs
zhjZpdGQ-_n-}9urGaK?hWVh)v>SO@tpsp)p7G2x?P!0+yk){7CVZoj)z_QLcMV=<=
zPOQ-asHed7KGD;}7+YsB=+|4TAd+UYBZ?41(LWr9U1L&x|Fi-YdQYzpGBWV&2m8A<
z#MD=w{YPlb*5<~eZ;X<br^Q@)`O^6LwrjheM~!m!5cg*#)hEmQe|UY^%}mv^qYmrD
z&GBN_JB97pQXz<*#EFL^XL&+Y{v7Vv_nzS894z-)9<Vj*^OJShj6UM>?ae%Ci^foU
z$}Nxz_BVNFU%nt(WJ_!~Bz5|S-Lo6|EqpugUTWF@h)(ZpIbO%kwcLz5iuZ2T*)j$7
zy4-iR@Y)l6`nr>mThi(+lS-f6!4ZCjsF$pyA0Q_F#R)QhY;bywyRTLxneex2gS-SR
z$;1jkJL?NRQk$rX{+(BI0d5K2-&Kc<LHWHewI}^dGcU`9C$*+DHN11j_SdjOD-(2!
zbq)vnL8>E4f3@pguFlwbe_`OpA><AnfS5|S&Tl6-0y&&JL_Tk)Giy{<#|RytGR}oH
zI%E|V{iPx5@AT8A6}C)qcS%7*-%HD*pKSHQ&Gr~4c5@9XwZ1l=zao12Ms`R$$)w6p
zZw@I~ZQG1uO}qbi@wv8V?OA>DD}6RyUQct6_vzmo3-$)XUPDb}>D~YYSm*@QcP1Y_
zttqEX!Z}{9=y9@JZyD#@l_~zop(?3kj5rKN-g;_U^YXkg-c2vZk%3|A)Lp~(rMP+m
z=kuuvAy&`0%=-{|c}q!}G^YAj+`K=8SNrd(c_fh9my&SC6+gJ~T3dI`&S+y_iq<c*
zdDr$KvTi(~t0~zTx+?ix8o(z6lng4>##M3b8^SWkCJXZ}1Fw+8-}7Dhzqt77u%^E+
zeyk4$0tTR@sB|gaD2Q~IbmxH414bwyNSAa;4e8v-Q4-S4fYIF}#zt-YM8Ci1_dLJ9
z_dNHW-FxmiuXA3n^R5}TFs@J2EY|+H9nKo1#j0hsvO_{<uXO<a9pgH@W9$CsRDRHT
ztl4`4$65^14d|ND2?RsK`%h*sG40ufp`ZgAf89EEO)Dsyf-`!!m}}SA`qT*@mz6{a
zGU#kruMYRG@@QG^OT_!UH@V>5u4(ags@?j&viMWt${9cUj)|>W^Qq5Cxv_C!HMJ-H
zY$q`Ux~6E^rtd#QiQCMnw>lZiOR()C?DICrZ^*i;Tc6f6zb6SsqfofRJlzyARUhp1
z<CFB}0$rqqnv}-b9Kdfwd(D{ZOsWKofN==nCv%b7AYA>6)Ag3jM(*)KY<Y3@&A_Dr
zDuD{1-kR)_>mUOwc4WVD>t>VuM#clIPW24=+>9d=TA%jks4ofWK$3UFG0x&Hm(w0T
zNkV+54C%9u$}AQjA3clXOrf@6m!BwlL#!g}x6~s&`J|lm<(s3bf+PuYi|N|ZQC1@0
zA9SI|I1f`T9>tz(8emDesvigGnfqAxRlD)=!!n^q8~|F_QD-q~Gt=`cX?8Bq+h^t_
ze3~Mgq~WTb!IyC!BO-nLZjtB9SlV?k(ceNhfcV9VRiTPuDI$RJdU!|C`;DuOmci87
zGc`kZbjP6i1T=F#x2t>bE-wb;{jto)<}%wqZ%-_y87^d-zf`j@9CYl90q(g-1DyKX
z->qB(>;i?m&KE#fv@^ukF02;s<0fSq{w*Ai&iQ`G8T?^M46DLJAy(rGZ`LPB5bnv>
zi;R)|I?T&EDK~pL$8fa13}+Zzx?jrso2%V$RP8Ka`{NRG&<DLFtX2wX28EmgIvB&w
z-pd?&M>ngOW8!<3>Rr=cyoN25=4XZg&e%r4R~xnu&M$$$FTD;1yB5agB@dPA{kk$w
z8imCfu53rVI3bLz$a7q_yJc$Y^po-onsp#+^^X@y^hK_G275|>-|p@d38BtLBHf_}
z)|hXjY#u^|R=6vsGtkoN#=Kwv!&IV--+Y}K^7wX4;pgY2D6<?!<L@w@EyhG8AvRv;
z0+YFTfnBYiKS!}^pe{Q`x@z5#(YtShs<m9nQxaX=uz^>~3+A3#TJ>w)xh?)4ozUhz
zx?3{F;&$ETy4vx<A6TDG153|-l}*Cz2*I}}1=ee5?Z8(BS_=_?XvjdmvBY8D_B<8O
zD_nL2x_Q0Il9H;q1Qt<7rXD&xzITqQC9!;#x!UMHGpB>+DO|BzTC}ZfYji=14B@zB
zx$zytG1zq+l`whI$qLrCdcH-BB?lM3eEOxuN4t~Q&FWvmGl_2s>!&03U0z)D^xgJ#
z<UHchCA*@s_NE&9X?8Vc6^d9acnG4g;q`Np*LzIy^20A+cXedoDo+<Xtlz7Ti3%LT
zH_^!djCSS8Wm1YCje=UDubq4o4dnuO$Hb@`EZ3d`lm4WIX-f+3KqlVrqjZgkzI%T+
z{ll`^aVnA<QJ7#qP5U^z5Wm;P6&YGPkf{T*=kvWwytipZLL9;C_e6;bq<?$&7Pz^1
z%abkqzM?6~Kg5Mm$aeGvukZ+GcsTbSjnJ<Q*;{`6%9PeV8(KrsTcW6bKCzSwAz9kH
zw`bAC#RoBX=I&TAJ>GWjflnl=nj7IpgpT0ll&&;IQq3jBCSU@$<_BD?{fo844v-5l
z=8`BO=%5phKxvG^Vcx$54^E<TV^J>SdLvHMW9VpL9Ql^K_(2q?UM}Hd;qzhnrKkSw
zKd$u%pT&I}RuYQ3MrE%~Ej7xh;>IV^reX+IkPvi}TX8rW$1H+fc(BVo>7rICSGT%l
zxvMjn)(~4+-PB2GUX+$4*ehtf1Tn7<zS;fp!*jx3h+^|Xq1vbUA~_CGh&QHra=WHn
zc{S5dd4j~{CQSP{n3e69!uI$}bu*!)jwx2<%AWb<x=z=eoTgx>;(wq?Q2Ii}emVg@
zugls<e{!QRS2KNUym;~%m=v1O87Zcir^fnqv+lN5u7IYk#(He$mv|8u`y*)p%x@uc
zGEs#;qsySdp{4N1%KGbnUFj`T7%0EKZ(5n-<W3_jnS;SCs}E!A&ANa#8r2zyWJuCp
zXO&h5C)MrF-XQ^AZKJxPuyy+6hW5Lxi0|X8@!t2YvB_t#B8Rd|8>D{_X-k^3LIo(u
zf#;!Cvl`+lwx(n<UAXkX0hKszrxnbaKy#IEF>mq9HSuOH#eYCUeZ~)(W*hYb&$kw`
z^w#)h$VaT3MLWsIvW!G_RNQYOtx7JjgI)QBd!W1V=NZT24_BhB=*k}X$vLkmfxzUE
z`pxVB6P<rP<RMdzXsO;ZE-ivMtMP0w8rgtO;=e}5g>e_F_$M%&Jy+>S5EP*b!B^#O
zMkj^TE__ws7oITo(PLO=csArPX;mtrYsNRazi_#F{@>E9`6@@x&)(;S$lo=x0@W&y
zukt!v*_oL2ak%)>U4_3L!&YLP&*|~*=&#JOX2c*)K_chq1)^8ZOgI8d9P_|8V`7*s
zDTTXG-m+-43mupgGQ_9&)~mn1xYk>lHxp^#e}q_BXRuL1r5@dfQR}2{lMoh&4Lq-d
zAlCH;(!s_h9VMNw$tC_9|L*8@RlqdHw=4H=Tl`Z^%ksIaY*ONuEdOYOb(USm6Y|ON
z`%=Kt7^P=os6Ipe1{EG5OV}3i&{{rNQOl0wQd~g{q<N82+x#fpno;OE5SAyDjLG*J
z;JL37_6aCelleXQ=vNzBL-%4|Y01Rl$Gq81I7eh5G&=*F{0ASM;LaZ-neOu9VXDrj
zUkE{ap~T4)%Vdk}A2eK-LW{(pb{@Gr(K6)MDN37}FRYedJa@vDoY6S$srYo+c48)2
z*4Bqk7z0UBo`X2Sw|hT~@r3qzSfylvo2*^^w7T`NECNIP;(u;-6CoeYEl+h{Q6V(L
zX=n3G5xluvwXl{&t?{?7igoH(aig41ol5tUv|BQBo~E)sUTcXNLID1al4)c4WhlkJ
zB0}4z`S>L4Jm<|qvO>Z98-gVp6Rk4w@d~$wY?SU}=i`JcSN#6`fICgC{L#H;eCC5P
zo_zTE6M#T5$96T3?*{aqf+2f&$!}bKjRI65YeC(&&Fojtkg^=M<Pve!+yZLp%(anX
zif%p%ICNb*p`yJwv;2AzTyZ$NWc0G4b1y;ZhVixl*5yi|rNoYO^z^$)OHR%l6qgJ^
zQEp#sM)UjhiyKYwPR}_-qJQ=!HG9>YMrh|gqAXR;RbwdvRFlve;vDd%7?ywK1{ZP0
zY$fhgMfvdt6@}=F1i=%LBjWRcG==#EnCa2i2O?4Hmbqn(%Z&*GmTLoc_12%u6(@zV
z`B53h68m}?y+9r^h3>w7LcTr4i+nItV)}BWsZV&ha@@6-M#sU<k(s@$bVSwwE_Umd
z7+i#P@hzh)*OBy<dU+{_m73&W>=7i7S!$yQ`!&gzuQ9s^UP>7Tln}Z~`-m(M>KsSH
z2U{FZhaz@p!xBD4#mnYPvL@W6V5)Cy^tU+emnNgiJU{PI_bC0p80LXH#{`|Dui9Pb
z=^832-<-YbxYDmG;!uaIvxjj8vw^Q}y<&e{E{qz8F0Y;Zm0Z2#ncew0_?Cpe?>unr
zI2t&W!%MZe|NJehosatIPgRX&GD+s5ZSc<mg`r}T72rKX`BPKukkv`7x4WiP_Vxy^
z@8oc2m#~7^fWMe(5gBFa$J{;P1n^=zA-VC;p!>&vC8gWw1KdSg7EqfvA~LzbWVaJz
zFmd?`i!WT=mSAskYs&IRIL|bcm@hOp<K|{)*OgQ*3e7GDzH~*)koZrJ@wm&o($+(L
zwlW@+(b7Gh6%%D9)2cs9YeC|x&H^;`%%35O65y*I8r<U^O1ba^)r+GQswO&F!%fTi
z`l7O*HhbN?VhIb@14{0bKZ)FlalaTq!5-|=m3tc^3xlX8pP32w*#Nm(qttW@8-em~
zg(J&zKA1}SjUR{|+iL*Sx`;sY0X$k*lDx#KF5OjhvNYIM&#wmPM&%WxWG)JB_dP_M
zvi224)07XuO!>L!Qo!KPY=zLR6k(&08&4#gZX&ZLi|y@Lxe8|tXMRPLRjC5$x1TbU
z`f9#V{kE<rEqA40Qy$T%qp!5BlBaLHFb~-(ycADD8f}E>NrM+WT8$T)c73EE9D)bG
zbHM_jtfh8Ki7U>I;pYa}P9}=0VZl%OZH@j@{dj=y)!+O|5SWwtWMLnya@TpTsZmAf
zkhq5ro;c2slFyKW%Peh%K%0%nHni*qiz{qL6DQ*m$BN$>Gt}sGCEmwY|JfQ~v=P%P
zYyP7M>ETp;f`V5|HRe>jA6^qg*Fmn6rHx7y<ph!{Ys6;aI6e{bLAl`lut4Ni;fML~
zb;CJdAI0VPzV=AOWf2vtZ6(i_zZBh%fh6!ed#Tv-T^g^mnl#&oQ#t6>Y^%4;IBMrd
z$xeM>wIs{60!tkVs7*S~H)rLg#!Im^Na&acd~h!9Z#ryEi8ZRTJ|2NSmMQr2S{Glt
zL*ARjPlCh;1ve5`zePn4;&VT8P+K~*$C+RG((|PK_v#;6`c4Y4N2rhI2&%K#M!mXR
zR&IIcNd6~%FL=AlO~VEHMFmC;t=t3FXL4H{{qZ}JkGcLq0x9vSBS~|Bq1k4KG)vM5
z<z{))?gl;)mtMScs@9epI4d^G;KEd>%2v#_9e&iypYYR`{%>4yTDKQHa`?DJ3rg8Q
zX`tLmiFxlTD**s{dPbCcCLwj;Yf$;IEN&*=c+X4!uJb^;1i0~R<VEX&)L%yEBXguS
z4%6jd1TC7*9m*mV1H1@dk2Cs5RHy42qkIeM7`cYocEeI?iS2v5)wxUr1mVxGTWlkh
z%eSbwn_~s#fd&UZ!ZrCTzPdWR34WGF{r5j@j_28U0*_rKH<Of`8G9v6Tl+xHF8Q%<
zzfY}tw{2Zh#Lt|?XU`>l6_g~Mdc9W4o<$tp5|dKVUK<{K#{DHToA>75g>0(%jzenV
zMGQ065C!uai*J~jzB0-vis>dOX_)`c`=FTn$h-aDdhdI8!kec~f=QRRh5XS8gv6uj
zkS>lNg>Sz@I`T*Fie>-hbYFedF7WyTRaqm=KDOtuR2~xo-#foISAPj9pxA{ByuGdd
zvY`7;WH!zJShJ5|KrGx?HDM3+sLExMCJ<gyygs|d?w&pHb}BueQ)i$)-s5k}!X&Sc
zRqPUQPANm!O+`M-MRln^7ktq*TeZ1_vBn%N^L^r?@8AfhCB{aY5~A+>rJYSq0c!m4
z>MdQkQLV)&U*#7mx0jpf(>u$TO!!(nOTx!CozFKm*b{K3BUexdtAG0hUHpM++o{E7
zfpMJao@=U{An!{d)8@BX_TaMvxiVDeJ4y;ZT#=A}UNnN7*c+ay{kQ)Fd$p590=v{-
zU&S0)mY;)fXT=JyCeqEtSHI150_qyZs?YT%1>j`_2vgz#F*=`II>v1eA1nvN**vED
z%eQ}CDyhmwS7(&pqt_m6UQuM2$6N?8#j1bi382#blNU^;Mz?BE_TC>09Wx#$3BU5A
ze|YmB4ZgH;5lFY8WpUnnt2}jp$MAnWzlcBE<F<wih>4@F?{=NN;Qjx8QDh$HxL&CM
zks*N(3tD_idHY{m=vOxr@=Aa16D`>ryRAHc&BK4bN<;^C@k7QRr&ViutOO8J(=hPf
z{huHpbQ=;jA!77MvLx9{|A+On|7(+}WB~%+xHYY<b=<KzT`iL(6ZlWBVRCwUI9woP
z`FwsKU%ZY_nf#~UXg!G4iLKvX?3NU~^uxpVIca(C{<AjtUVg(#hSJcs+f3(5B2%#P
zkM{)6WNmK}U<z~UPfTo2LT7?|6Scp7EJlpIu1E+BjFDrw;aFBFp7})T^5&eOZM@Ob
z34Vt^-lJ!rp^li%8Qq+o9!7}`whib~ZtrD2T^F&|l&y=tRGe5H%P808{~mPFX(+me
z_2D8AzEV7FdH5M#s8i(E%Oa|&r=(%>PVw%ch=h21UuOzrGY2A9=?%ibO4Q+tCIq+K
zELf=5BktF`cn6dl@OOlZ$JCxgK2#BsGsL6fk^<qc&N#0XEmMU}GhNZd)&j+Rym$VK
zX^eN9s0(^V$9b^hR+5KVkf5`((HeRB%qg>M5R@fgs`zs#hNBWzzui;r_{X7*gf0fp
z%sH3Yu20USB@|m4n4(AHpY-P7bri9HlHzM@>SI(q0Bc(#oO(0x1FQ2Ta+F0X9(>}B
z3zrc0Y<lfruo{6J@{Q;syk4MSJE17a!*aMgylrLwH5~AfdqYS^fMx&6$2vDa+=z&i
zB;~kTryosqd7SRAd~vx0M|n1!;~mZbwA@^c4l)0qXh#3;Io%w!K^|xGQESj*WzUjE
zJI>7Ju_V{$>i&Y~^@M$OQ51myvx<e@)w)OXce%<{7UHzpyFSyu0cvj2QREYCzr0`M
zl6kg2q?dt--jERD_X+cbNqPAml7xu&u@0ETw-Qj2#rZ4+k-axE<(1C54#O9MFKiqc
z4$baTvlknYKB43dKy4UkB5N3mgXo&x^1>_<41|TO4)!-nMwu7pPgP2lt4($pJkmVJ
z9$@_{cQ~Mlu552wtK`RyS);BTyOdZSY1l^2%h7@(*ysq-y^!~NkG4~72#Id!6uzEE
z06NO>Gd1xt3V>;%H@s-cIw6tB2Ux1SBZ;}g?U7|%w)jXbyN~QLbLCLSMGr>D-MehI
z?Z@yIbnELyKvVlBYL7+?POvMf?C>g*fFDL5qh>JhRq5HXlJd8pHfEu{)xpvf30VeZ
z0Zz|hIkRZQb`pZZzj8;3_FnLN$dwr)?N9t(y=J_@I@7bg?{5U#n^W}^8*3u_`9GUh
z^i0%_`_!FC(|1^oIT9U5!iL<9|B8hys9H@lY!h~f%bMv=jIQNCtz}TV8s3{zQ^D`5
z<$sYV{xQL^wQu|n(9_QZs*f{$RWW*lF}pB0Y~P+O8Fk|)Ac`y*MYnH1dE{5gdncRX
zGj7>OfoQIaze*>o+r<3YrPme5N?NY;#Ts|{!cYt1z2MYFSOLZUT89-;endMw=-OzP
zy95}%^ph~aQS<yo6zTGe7YN*NKII@*-RJr4QLv`A?)e?FJA2ozIYhWj9d=8qs@z%c
z&Y+cD>2_)4)%YK<(#8IqLTE0tSa0G}R|`MU9;1%(tG{8Vdr`2lbQ-5^yVqla`qo#l
zV78&9W%ty__?bK_ztD9!d4XOoNmQ!?jg6pQNBuov-EGN0f<*uMvs)^gcUr#)5-gTX
zOzC0qiT)4hIWt`hgN$c&5Bwt)?)?o%f=>vSYi<n6)7hVa%mI6w!}r3wIVAKk{LOsV
zv67mSU>C3%x|_YTK5*#vRC1MW7JLEQj-4e>iz-v%nq666cE(*^>|&lv|C66dStx#L
z{R9fF>QgO+6;(3S!s;b`Jxab7ROc=!RkZBL;`tSJ?oaM|Dcr5J8#J8X_p?@j*0(;s
zj-{V@TwaLZ42tr`S4|mpA+!?8=h6@5LgCiNk<)39(z1Ak2ar@(g_F$Ik-|XrergZX
zYvmjDw_DG)F7zT=z^Esl7lek@qBdP^IR9B~FPM++D@^pN0C<6=Nc_>_^9Hl^(p`sN
zdZ!6}cp=^SM5VB(=F{bD!<xH)O&1$>Yu4he9WQiqK)qH6f>IPmg;-6OSG(ExMN{(`
zd8bRwI|&g5{(GCu_0Da#IF-vj%zq&xe<Lw7QNKg{Lu&&1G=XoG-zKD?ZB?13?R6_t
zG9A(1WKiI3AapZ|VL<0C05NzG<wiTv9V9L0(UWXqj?NJ0dGB{=voc*F|LOpS9Ow<k
z@=GuU`EBKCjoio`I^r0?pDwsym%qXzfIkPOYOv4&EiEbSS7YT`dTlD=f6ndMK4Brj
z6ym?#cQVT*6dBpr4~IPfiE+Q93k&~W)<dliW~2Nj+bOK0{{Sd;Uj=s`r=znKovP}G
zKI<pM+rBqdvQ{lOnz{Eki@E4#6?eW|r^w*?9<}Vk!)AI~PKWM?dP=5VYmojb+`|S9
zt~Z>#eC>#ZUp;cg05#U*K05`n_58xxwaJItg!WQi%C1$1RmL))oe;^W_>}<b5btwv
zowu;a{nE1RO8Qsj%g6DPY->w9V7Z!*_~G@9q)Y8wfN|SF=i!aEvM9;TqGAAyj)j+~
zc5q@P^`5C5e!uK7eNFEZralP?16{@-4ddd+R8O7ue|}~LB4E$ImO(2|Y}7i-y8*rD
zW!Wpcm1-NE-rxqo`}tKBbw@k5)|mPsB*gdY%oJ7gYQt;b_?CP!M~TSti>LN|mxro&
z4;^dljs^<k<RmD<Q|`OLUtOJ)z%lV~8{?-P+CMhY*SX93c{2}LUfOfIXfa{Ct{URk
zWw7u2@J|LX{pRoZ*>@jM>FgZMrF_=|_AL476PNi2XG^bclzy7d<j_sA_Lo=HyezX-
zt<TMKISNLDnriWT!8dff_c!hoE2O%$&yAd;l^sDG%DPWcuXcoWh9jKs2|y>Cp;lOH
zKErN9QQ#<FtV6eE*qgcFlNp_I>Z}(icH<BK4P*#fO~n2}0GYQk%gTj-UC>-Hjx>SG
zoO%vpU;jF3j-cWbE>1!A^{od_+6`$M@u_e4&r^?iL$Hv!>guX)N=I^3*)xiLyULE^
zb!I=j$Av4?gk&Agy|czOz`}4l%h%~@Pd{X{*?8@wGYfT|MY>!x?T?NptU?mQE^9)r
zKsm?nhi}|qbJVL7ZGGn3t@3Hj=SX@j2EVrSgWy(wEF*ylm&|gL=fi2h2b7&R(#T<t
z*K)MuieoF)Q^ZxLe-l(L7q(}knRb6XB=X$C3(*0HRL+ONiwunoHGqgDnXs0vQyD}a
ze|gK8ad{>vHlYKsy6rkvl=2|oP)9d^uZ~iUJmvIc;0ed$a<$VbXVl;cUs`2p3?OlE
zC(ctuc9yfzvx{hG-`D_YVBQ^h<#e{WABvC_cHCa&O8mGz&`fzuS%!m4FZ0D^fDX}Q
zUv8Ao3lF_Kq-0P`zZ`un02fR^<r}B{{8Io}|1AK#FWU>93op&7@{-PKU&fv$M;Db`
zqPL*M0HxeD^MX1e+=3-u?GWYg;u-+A-o5zJ)3p`wQ9`KTxTvKlAOJY3;6f=q`ZJl0
z(LL{Fp~~t?p=B|@HWFrnwdRcA>nto<4j96akr!ERC{wvA=cU;;G&LP*k9cc}sTK%q
zH=s|-ME!jNwsr?*;@mmb%`+~QPBd7e))RKwN`T|?%elYSqcJ>DhkAU!z_`hoeTC5)
zV@Y<z{SobEMv;Qmpix^F$!{%hS3^`3^{Om)gv|pxA~T#E6s?=adB^wr3m<3VHA*3l
z@GWqtKPO+FYhBTrlgDAUe13E}c0B07#60yRO3vHCZqETs*?>^B`kAfmELdGKKLD_b
zh(2^u^&9K$fq1a7GWXbhk5C{w6yVnxtr_XLAh94iB0rzQlc;E0o_XCi7P=tDAb2pK
zY%hiPeF@?;7y08_tRF}W;UjGnN@4OxAgoObCL3C^eemUS;V2CZSDj}Na(A;!SBX(T
zdQ5%#RAudof&7=9^eJs}%ylwvShi4=PZKq*p*!vyFuWKZ*R&>y{H}4}+ngYO9v!L2
z%g;NBoY1n-PfP7nCf}PjnnY^*umW!X$$vzv&Vrw_XAt|wPaTBi1)Cf$j4_~&0{9vf
zRUqx-;oA-W5zyVGH2n#VNPHaF-bM9F*Klx&Y$pu4z`u)YFvAAu`qOM!{<eJHNYeC4
z++=5Sb^p=s)+Xw8(6|u&6tFEX`L)WJ-IgCR-Q0^9+EAqXDrmcwCL~LqCmfp6IRxo0
zZxn5dk^Z$J6<9l&s3-Ds)aI+^OJ1n7zi~$Mgws{Y36oKsGe*!$rYK!t=Immn2!oB&
z8WXMtOdS4V^W)}~jdD8j@|dXU$&8xRmUp@Ga)yxEunsn#^ii8D-S6x|?<DrO@Hf9I
ze62{E&}41gFhQwJHhXdS=oj-)-CJXWrw(2prqEjjBJ=FPmQ7dRrJY(Og*U4n8=o!J
z;?VJdr_5=8Fp@E`|I$B(o|~z*+Q)b-YXDkWVl#4AV9djG4YPE^RA658B7mSzVQA$S
z(_o5|q2&nZYVPt2^ug0vivG*)s;M`9x)9FfrXz$Bkab+L&aHGP?^;`HyKr3dBwXCL
zt`6|!QBs}}#R6?o!lOYFJgm#d;ArYfpU*TWe2qxsqj80SVY9=91mY<LaR=qnZtHa!
z|0Av9^sLMEGg+sMQfRG|)RHSWGDUS6)MI0<9~SX|zf);KMV}+e5CIf1_2CTT<(%>*
zK)f>DUX3ox!Y&tiah0Vz%h*YRm-%L6<ZAuLU<D290uvf^mkkHd0rF7955bKP?v*E6
z7My-!>}4D(+Ur;NTjt9Sz!DjLGB+3?sP>TQb8;c1I!RxF?!2eu{Wuic(r5`Ng;L4f
z>Fmck^A5O!te5u=@C_M_iVF#@t-X%MN-_^4Lu*U@>bslS#*&%lYv*TYVm*<XSD@%Z
zriHgt`uv_$hT%n9bAO7&YdwR(bBgb7n5*`@+6^5<I6;GQO!^`~zA2Sl331QtF&&s0
z3J~8_j9%)at>1LKa(tkaah1gxia=Xi9JUiO{T}dJc{`XhN(lz?QRYL*)Sd|#mCol3
zSP%KEbvcG2PR&0Jr;oRAVW>azRW(vvL!s8LzV_uNvB;&5)s8DY`%iYK(~Paz$28V0
z#`3Kw&BHV%=<F(}izn^Qje9iUkO8%tL~+de5H_C=TRb%i85D+LYE&Ke79@nU&4h~R
zYXyg%(XY|cP{?$hrRjZIr4f!DgSJk}NK9oE?9DK0<O&b949;g|nCE8}fr#Da9a)9j
z`w}Vf-)--oH7qpG9?FGxPN8>4A!8RF_UIm15+IGu{AM?Mis!`dX~f`ce%AUzL{8t-
zbhTqRe}2$8{Nib1^~JnL_o~HMKa@eFc@~XUkPyY~jkt}Sb^cWoXE=x_`Op5M2m!l0
zhAN$1MB(dDc~FfwCD4tEDd&@%=+a_W&Y_<nT+w??!cw|f_0CsGEfC;fIgp>&Rpq{t
z-~iup*nIWf0)bl<NrATrFy?kOTT`0C;4bEkV9w~Fk$Xd=DE$7q`pWy7>LE?z%I-_r
z-_5l1fydYnx#}!6lCpU&*41?9hQ3y}1%X0^+i+?U9o?LAaLjCv^xb^P{?P9yI3|UB
zRv5RYTJGbwiP)(7hkPe(MILLP!~mnc26CnlN2iJAU8`cP-#u(V`G+we3;PrfzVS4_
zXwssbE=Oyw;;&W*0`(cwJJ^gVc2}3iKbSQt>M_3hq2*p)*<t2BH;Ju&#RF{C8~6Tz
z(?JfIYQVz9{XTbH@mCY*hnD;P=o53AjfblfyK|nGM>1US-U6KYyY{MNIPxfh{`gw|
zTzK?X|2%j)rvkU_@$I2ir_xgaBK7VV2-myn=1&7ft1k+oNNn4A4vLWE2DHnsw<!Ti
zcQQvQ|ANfH?O#sS!;h%xYTaIh4X_kzeEW=%vfb+3h(6njHy5HYw@_i1dqZosbHC{6
ztz@z&`|0zIS49})(kbpFM~>$vk&JlNF;|MqK|W2(Cker$Ca~%9>?b;)b+=!6RUT^c
zO+@60fJnH$`A%h%IzX{tX8kglc3mAnCh3;WT(rlpM91egj?}*TqBZdeGaaS#m6XkD
zZXfkXbQA;Q;S^pQWKwS+8hciDV`%EwKT8jde(=cYn0F;*sP~A=Si)n<%>I9Z>3kf$
zMbzPO?&YyEEX#WlY#A2fOu5#k1M&gDeu@Eo!9zv)Ln?=H7zy-tqd%Ic#W~>6+sR?z
z+}h*HMqFB|vwO$~#wYZexv^%sX7cWzqUISa<UaxteQZqr$c(1F?jYnm=or};lYe6b
zIR893QCE$@pquG-%07^N`+*S~<?oBvLUI{F%^Ya{ah<D^l>rjtMw9Gv`d;t3Cg#N?
z<3`%1;1AK>D{+EnTfDi=<)(Xdnvhf0JeR^!S9f(5Wwd5UB1EtB&8&-V?gWPB$$4We
z5t=b?Y-fJ@G(vKYTgYR+KS+Al4x=Vks7mXz*T}@(nBhac_{g+LRlCv)+dbDO<h+^`
z{9T_&r=%~#7oFq4_27VifFUf(S7XE-5<r3wr}9~oT8a~YS<Zp9`ewXQoe@oQHao+9
zXDk0sLH)<S^`EKhi!vnC4~uZm>As*SBuncfgORO02V^obIIeVm)K?xxucw-y+=540
z(P^pp#S2{ozy8&58m;u{5|A%#U2eVU-~Xh%h+*%8w6di}q<ZoAxBBDmw0-fpzP8!{
zv3!W5g_aWF%JpY)w<l9<DU81`P3QAlA=wNDZz{Vf2162O6;#75x~!jq>npTQC+QNA
zX!st{NKsmXFg>;5Y<hpr7X|yOa(^pZ#OI`E<hDL+X@Yo<0tlzGB*$vJSR*vBJgG3N
zR;cvFfF4KhHt$P=QSi?6_!!kTg^8ZFqn315vA^A4+(97N=Z=9DMP>a&ijhYDG-&W#
zW)9UoBTsVJE(&{KgkPVsI(e}X6&%0Q`;mi$pZdzvP<$Gfgv-$B8q6Zy1Ns;zfzUHR
z6+Ht0m4vs?<>$G^x@>8NpzJ{X?H&3@95Y7Ex1Lysvat3^ihiv)|7@XCFxdB2OHX9T
zxqeN?U$kbXt?lr9{9-PBZP&~C17=>L#UqDfwxMnaZilCN;`2~H1~L0Jn)YgQ5Y4HP
zmT^Ky3ibTr5PtXg-lqH5zRKVpAYkxfx$r@>08+JHL+`p>u^2%ob3?Gy*8yaN8=$$N
z!q3x*%a|J|XxUU<iL~n_wl}WKH#LX>)kOD8e+%wPM27`GJ~1h^NYe5H>hv5b2n&0B
zREl^won>T3O|}DUaL~jRFJzXWR4QQtRi?X@aaM09V-EA4d<1Xy?KCK^Kc|V`opCRl
zEz*88{V1Uz@?;WNxMPI(if>oUuFb3Hf>39pKsdr6$+@FJh_$=-oY#s@#C^Hv9v3sZ
z$5~3Oha)>|GtIFO4pd2Blgl)KM})Kdp|QB{(vsnh2*lOKm@CBZd&vlla^`(L&)ma-
zPOTi&&a|+%#*Qv~S@fCk`gXc>vG$30pKZyb>#VuBn`nsmLHa&;m)Gtnd8-k-2Xoh+
zo~Sl^JV!Ip3(3M}&d$vUiOzVoEHAzd6<DRkSJmx1ov|c&Y*QW`kLkS<QPxZhF{1E5
zZ!Jef6$~Oj6l&%hZXAm<8SsA=ded0ZAb0z@I<VtZZLD^;(q`$$PZF-so#v*2ZnmfK
zO%^n+3>?V%l}1hp^19%8?#XfE88lL@li>Bldy#fyI4)tktBUzbt&h=|hY0%;r>oZ(
zHbK>#={AibN~%A9ZID_8@NGB~&2Y6N-t)B5$lW{?nwzQWw_C}t_8u#w8F(m9WtUgW
z{Ry~Y05MgjZI;wTPv0+q{e8IvB4VS^!*@;08%e!&mN)haa6*KaE9ZwpWpWoiIXh!p
zm#aVfm}ERBHx@3HGHv@iH<Vi}?Ykp2HVC%{&zgn>KgJprrm_GflHE=k_dyTJb_BnT
z+XgA;PS|Ru_NJy9+A<mSZS899*Vp(_4}kR7*7>@(yH&VF<9nP3(W*E~jY>!e>z!G#
zAymTvjWr4QP6*!m?DUy+IQNNIZN+?I`H$w;a28t#eU|)XW}SDxFC7h!!B(;`!&-K|
z;S-N;14}0CI8#5*xZ$$<(8Do<=BX<$cNrU!2hpCG<<&u+nrc@GooP)eUw{6`I0DOh
zI1@i=aC(tFVXk8FM(9d~=*x+$Qm6yo$#r7XRV4EsO|xhJCfYnD!`wY%Y>fgXZi*zs
zafC6hh;?3%Ieg72M7}Qdt;w9Ui`Pr(>v7u)>X7I)%$nZtVlAk-KGN0$A>MBXM`5nc
zyWaEp%7_x}B!1ZUoxv&(Wtw<5x#UL8yV42+B8@K5+f(S_+>ObNOt58ld{CZIHpv|6
zdBOR40OeP`mnSDKB!_YC)KZ>j(&3_GzP6&X!XzB2>|D$^o4M?9BqVbC==LP6ZiPAF
z?$R!MQe)?bU@nfnMu5AOhjD&JloI&C4x`ueOd|H^i01afgJVp_Wo<;5zFGUju@1#l
zjDp)q=BU-xyQ@5c+V?glC;B1^R-D9qJ~4Mbv|J_~6j3Jh_bJUIrR*k#@QDIDqsdVo
z&L@q>Gp;~8J64BLZ_oOR##9=FE#M008#;&QE3+*+wYXLm>GYpU#ppb{LJWFZHjZNu
zQWrypyvPG_;VqXTbAJI`MS`xuiyWej9I(<iIyn7G9mfOapCwQ41&65hbbW}%H9&Ui
zUsc7B-K#RvAL@aEYfIS=s-_%WwfNI+!hG!ePqVm3y~~k>rH=+Vqj+x+pntPnW^aYE
zF-GW3tnslNl8kB4>ht32g+eLLg|ozVUQtmjFD-8}JG;3hfu%(b??iv_);IR{)y2?A
zkSAlke#4vM%mTV(&ZV!14sxK-D%`i}0~h`W_UfUz(qEL7tq?!&f0fT2+2);#D-ji3
zKMQX^*|WqL=mE<)CAMlQEYIv%rqr#@du03uKc$#<EvLU)NU#+KJu`0cGJmBMs$)J0
z*tUc98b%EoHgYN7W$JS?wTUT-kOEqyF1%FSMj=x?`4SlP!^q|~iiRXG8<7l@$29t!
z(f2S_vQ0%Z8w)e_BXwwTfQ|U1?m{VE9gf!Ir{CE#uX}Txp)>?5N@e;mkQ(oK1B`s`
zeWI-|y(N$WDIIbc_4=|8Q6r1I0~w3YjR$^%IyjSAb}J&iWtYRsT+hnRkvW~5lRM_y
zut?))6cg@c$DD3E{D`Iat(ieeqX~HM(ulv0#Tiu+rOx*F<R<%>-uulvt=aF5%iq-l
zP!k@YV?ANZWvN5jb;4-O#pq7I^73eL=QJ)lBQW6oJaWdu(u&;=yfC=@F%lZ^w3Ng4
zc?uSuq-L+Z+dQi`b{h(nufCa)r6=?@XXQYv+H|k0v;4AZQ9^>_bcn>&_;e}f%E61P
zbhPFv>gVF&{|@{}IKS{GdE^OxNd-fmqGVpwf-x=@Xi8F>+e3ex7^w+NYI3gVJ&-GQ
zfZEIDchv#hBer|pfL4*}*7NOj_`0onF$SfX{cr{SgOO~puJ1o=i_JC+Y(t0*kKWHT
z%LB0EldMq?%Mu^U4FL}F1y2;8{baJT^@y$zy<YBHzHbesVk)4l3zogH^eQ4Qffm3z
zF2&#9`hetpIk&(nd5T|DO1g{MMsgUU7V!*ATkzJuVfD~?-*EjQH3evAu20<O^-0$t
zHS+S<d5oVWw8ghaY^h5n)STAOv4<>EY<%UDY2udea*50kcCXGt%y~ReY*6cP@9h{P
z;AeIrE$0(XCEw#O_2IW>18&V-Dgz>&w4H}{&g2MMQ$RaIje8Z^@3Zek7o1YgvypL-
zu}XnL0vr1~$C?P`Z3K2j52OQeS#0cZ4@`%y%deq~0XoP2sWQrJBywj%iLUZuCY{1h
zF#qiJlK|0&0{Y~V$l<)-9Khe(TYi0IWlv^bb8wXEmRgx|+MprD%S1H6USJQSYB=n|
zN5If|AN7kUo_QX8QCmyt%~`6VJeX@Btx#MfJ9ohwM-DCeE#DJ&Z)rXkS^iF&pAX1x
z_a+^=wYEyV%Gyh%IY-C1yS7dJ`h!<-QOE|7Lc3wc*NO)AmE|T&$y^QARDpGY<oowP
z?t5NkFs<b0tB-8!)b(s0uu_4vv7XiC^8lMq&`f>B$NO9C<6?+B9VFf*d&}3)a<Uq5
zl{2=r->^t$-p)w8;=4CU!d11geWyzd_q!xY#L61gjLk3UFVm<ix$Gt@jwr0!U6~Nr
z<Hf12f|17xiOJn9iCsC%4SJ;=6MerVhJQ7#(B)yKc(P`7rs_MGXjAxlE<x4*hO0`E
zXO|odbaO3B8Y>0QNeqsig3hKAStH|H3N(PO?)r(UFipM(DdqYb=9N{Tg)?<+{8mXU
z%{hv1!2m@-y1!M{7Dw=nk$^xy{kJ$KZu<FGZju%-q*KQ2@=-#6DE4CcUH)8o(0bq=
zxFx<v^mRk86<O`}vux$VTF1TYsu?Lum=bc-`=>tL>Q7RYmA4_wh&kwFv1zZy%=+y4
zy!=j@d_ul(hHXvA{bCY#D!p>`3G5pMUj-HC+T7Nh#L4;8+KTfSsMMCxhMoS`jIR?S
z4uH6Mz*Jw6`*@2V!G1>RIP?7csY-@&BlB9-QikWy%{cLvQWdOLnJ80h#mavw;gR#U
zVc+=5#i~1ZzIeGRvWyJ85^a0cq<?Q>U8SMyx9{rlBDU-s-<+M&`p$%xSQ%`+^@jWI
z_zRTd$<)by)LeA`_GSL)z`%li*&O(3u63|SM)6B*Z1gRo7UAq_c}_<>c}C65<sQ?q
zh{T^Vbcy|146?f>I6F8Lm>?`a5K_ypedVjOq4kcFHuH;hPYN7*U;2G@k^Ai|WdA*7
z>rEz#u|m?S{P5D*IvYv(qh48aPHmQO+sdWQYMOGbzg)W!_et=hgJ>n-(DhK+Sh?bT
zhvEraev4i1v@SDEY5W-0=VY|)h`ZSE>i-RddH3C{wQt;q|02|D7##baaW-hUaJBdC
z5JVQ4(_^M9q#|0mmP84isLvEMy{!CbDufu!@b@(B?vj_T4d)qmpH86@J3OShczUbi
zqX$-!L=gV`v*BqVJahgs=n%cW`^t8R%zcYGVXo2BCt#^tN@+lYs;X41Y3MC|t93wa
zV|j6S$oTtBbY$3iNmxT9pRb!C!|hY0lVfzOT!veNc4R|cS*4_d-+2wj$ud6YYOK%V
zC<VX2AJxzR>2l!9xbu5fZ2hcW=z*cL`$?OXA^6J;p1`JC88kt|2!4KQ=0aW~x|o?v
z=PRnV0)Zhsmc-hs&GC`MLh7iORd&(Z@6aQ2G9@KSxtJ_#GdGh6oei-i46fm@46TiS
zLrjH4HLKoregQ;MQQ`MryL)ZMB}p~C3c=|u9JW0Mm*?=lQDxK&3lmmCjnx(MYPoCE
za$CHJKDeH2vyqtlG(D@XtLIuFId&1S_UMboOVP_9?R;&r{3L2#T;M^X^n;O=s+%~g
z4)E=oHSO^F)p_nKqxGXi6~zM*&|}F`hZ5W0L#0EgVnb!q$lXphMA-^S)XBpXMuN{b
zempUxgK@>;9PM{!NuwPF;0GB`$tu++PPa01sY=B}Ro7|It*e`=eG)_QFD2OvM*V8N
z7KEi+3bK;E*NOhXrSls)w3X{ZzKa9VHou%K>-QJc<^p+E%q^@6Zr(Xn4v*wy>G!&{
zIkBB2f2E_av|}5Kjlyo9Ztn8)zqPW@Nx1ibhd9PFB&%V4I~oRm9rQ_j4-y;WiJbQB
zj)OGiY@rKm*-PeiKh2fj5g^yLXD_LMWiTIkr<CS+TYpWs!#Bx72}WX4iFX}@hg-H!
zz}H{K_`Kv~V^SKoKOfs=^0Ne3(KMsxJy&vA4fd)jsP66+>VfX?5PJwk_ljIjZ<+GX
zQmYqCLJ<X7W5w<>Q4i{Qz4NCjQ-T|Bf53si#5Y1;O1@gA(<-Lza}+(9F|)Yv{)K5;
z4gzm<cKj}-^W&Y+^s8mTCHVNgc<$%Fbb@lk>x5UmWiZpT8y$_Up7c4nmckE0z86pG
zW+GoZw}9lov!H0Wc3jx>3Zu6)l_FV-o4DbSdg+~K&q?3B>hBXwy6CLjC#)U-gDpYI
zyZteId}}){eSL0&)n>WYgl_YDxj&jd(tl^FDR={aF2fdLPD!5^rMxfv>mgl@SH9LA
zN;&y&E#-&)(l+*J^e%d?;FJE;f|b}RF-@P5J3F#(z*jw4CTO-tYI~KPjP+|(UfG(g
zvqNH7xwYr{$wBH0Lx`w>Z&?ONp!CKyiSEHFD7b6~Dy@M75WnWk-9FEL$ELEW;A`{x
zMj9a6*<{=&m=z%uMM~$)(ROsWq;V##X|e#XV5>-D{UGn;FNvOK2~_f~w4(Nc^+kLa
z?`D=p%z@$j2E>H}y#x5BO0h;X&^ry!VB;pa&CIJrR=(RSVkTQ8qktFnH&B;}a8ILr
zJ!e?9<?rx*eNPiHda3Whv(R~WiM8<3V<G8kXg;Hr&w2gomcI)VX`IuYoA(61-TnEQ
z@TD1>cInI*Y#TE?2ZQ0DbMweOM~B9X1&3N_H)h@u1l@yqG(wXm_1WfgmT$a#fhBX4
zlzq|t`m9A(_Acq0cdZ0jSb$MeT#EI&*LqsYn}z;<o}icua?!QUC#Jw#1h3%)pKmz}
zI3z~52>x0yrJsWt)EmA}JuCO?s~pp(rb8j9HE(lNiylW0JSNouVA&z>vO`W3bp3z+
z>=9=a>V$HbzFilGKBJc`D}=|B&XD$rd=k?DO?X_9Z3e0L8VjnKg4P>WRxZy$@i!ad
z{f$*Le38TYfN&xL#g}fBWXqM^CgWTYjVwm|02iFG)Zl~9c@zY#IS@md-8G?1#AjdB
ziR%XWI$X`sf8DO)c6sI4yAqNH1m=6*qrel=P<1jJZ>}~Ie|Kmh_)LmpAt^Vu3qzu;
z3BzQLPfvZaC90Th-R-d?sfId!%%3nv-FoKWV|6EFUb5gVr&iY4F^GicIXiiv!x`Xt
zgE7xQ`4&6*n0hPs1SWMt41(4CkEXkl;M$@tsjiEHl5`fom)I7flKe{Xf{ZF3+p160
zp)qtX7iXHK%pn~TlPSkW@D=^X_3W_z_?=&-Rrf}=<iaG<(7pFcpG~gDu*e>~F|712
ziX~K;0>na=zN3(pVLkf^9Dx%8LZ|F*_rroJWb_*KbuR}oJ?B|j;XvAFJd96%eX#i$
zpRJ}VazDI4w`wi3r0&8Uq&{HbVLb?s-)Pp@i*veIbBin(`FNB@qkMAn#nv5y7Zxhc
z(9jqYu4tdHBJ+#Z{;ldv<>w7B6=Us78TA6j!^WX$#;%aB<78TB)<tI>gUxgTt{d^<
z2ailuIF9%e5<Sy^SWEUyTJK%YuTU8q%q<?vN&Rn;&^_TkEmtWH3E2s;&07&Dv6bI|
zmi1HW$<^5{&Ht#s{hHc(^X(xcXZ&QneOK!DZ$$9-GSA)4%iaV1N>R_Hwgy%>fG6v6
z-thNG!w7B_Zaw2<Yu;xpV)+vqGa~?rz`<Rxu=is8graaC-i=q>_u4@T^0acrIDmO$
zzjF>2eYGO@@$Pcgsf_|Kb5&nY^wEQq!$*nb>JWq*XJLi7eJ4@}b5ydJZCG1cA~Zs;
zZ{s)oNaMwszkuh4PA%;5;P?wV{i~k5T)(t-!TzH%9LwU&_&)bV<|iQ`Y~#3+#7ZAw
z7m?+%7RZ2Z{p|Ql6H3FJJJ7e?;M~%pYUm3KfYU)ad*D^1ci8AwsV$6eVV`q>uglTy
zYEAV&c$@6=h^Noz2T0j3TW?>x4I>=RF!5imtz*U1P~c8k(icOhR#9%`%-ztnCPp^J
z-;9*ULWE@ts68Gd`FP_6^|D?2F{+dQ;<!0xW>A$q`&^Nv(WDn_DqFL4q;jZjVVIT9
z)QKG5_z_NLfETR=n}djq&p2NazIhJadZlhop2a*<yEY85)^3(77%_xeP-a)AQ1i$0
zgvm9==IlPPk_Y-%&2WnCyH+*qbXJaLoT6{Upl`P3O=cHFX-zzl<bD;<Ef=qJ3*9MK
zD<Zn>NA2MAt)p95UDp^RshrW}7u6ag7<$4yFd10Y!wb?=40162cyR{=2ilfhT-r-#
zW^XtF9cNX0MLyBTuh2pil`Yz(95{r5|0Q?qCpRv+4AhSX_klTms_nbB58Zw!T61qH
z?pHVlzZ?Dg?im{cEuoaXq+A%{M@Mtm5n@OiPYY*z<^?Q5K4YZnr8*Zfg>b^(Wm}j~
ziPZs!^O~zA6!>m%>Z>M`{F&>}r24KJfH!}SrZl+OflAMdX$<)VQ8Q#9RYt?)eS?vW
zg#T*(jtabS=PQknF=KGQ`=0cmV_%Rp`HeqM=5voCke9aTCqoc><oEg$dWCWj@xIat
zF_G~;K;+I37Ds^@?Fuk;K5$#)iNr+o%Fp|-@nW!=neJ3o@Uql>t$LT3$%oLUYPp}T
zA-1ndY08#jPd)8V6m?tOxXypjklT8!*TR;L-F2MX<Q}YO)AlYj3++TMEL?&_9qtlb
z-@<KBNtJTpM9BTKx6SR@`FaS_47p8j;79D1(hbBbK1J9k)(RyZfyvCxEeeEpCD$0(
zFk&$wXJ5)WGwI6JHCHN`*j3R%Elqsdg^WhOj+M_#3smNi_pQf-SN(f-y`fiG&U!)D
zplJd~YX-ng@~WJs)wXJiL`TL@tkiDd#4O>)h>L#hAbuc!yLw`PLBs{`>e?%O`}SYs
zf-JdXee?N(VyJR02iJ8SUobkJw=DmcX>WhA>K3>+mr|6-cjHDPf<AUk{Aq0la1Tv>
z-#n=4{o{VzM)*2=CizG~JcV$E+}-%W#W%W{29Doa1goK#a;Dy$V54SWdWP4Y(|23J
z(mSB(8{MX9=D0KE@7#|kZ`XaxtmGH-iw%<c_V(&BUm%=u9Qcu%5!-x6rRD6n8qA3Y
zdT%zpxq7x6P4I?)^X?6Th)SNLt}f!p7DABs=W^VaIx1r2QJ6q89Hm}17g74q@?v-2
zZq4)FM$qjUZlm7)9vSKj%I8EE`!Ue(U$`j?{dBaz-h^TTQljZcaGfxo)?OFtuL08{
z07{R&m7VxIAk*4D)r^e$34!nP?fS~o2nxJ(!{aQ*Bs<3>ZvmBvlRdL6G#6_+@6}R8
z26O19q5qrfxcYi8aL#G2No5{xjUdGCdhaA6Aa6-YFTqqAf@2Q|%pHJoeDoH&-y@WB
zPsD|tL|nYg9Rh$czp@A;KXT-k12rs?CU@=k{rS5pPCB+=wv}vB(r`LRHWw?~8JPa;
zc=QuppQN{TI+UJoQZH#~bIbM0aku_UZ1=Ju1(}~hUG18_J~QUwD-MNV?dIlLyw!@4
z7Kz|Iq<VwVGWYBkr5&Krj<8nfF<ti&F`eJ$@(1Na;nI}GE|c__?g{~hg*5#H4q)ji
z9^HaUzZedhABzc?3RfW_*+@o_nuS4voRz#-J8u8=;)I1My%ES`r%$eV%g#>JjyCv}
zGV@=VUe@#Z8D8i&z;abbL{h18uFygeY_*$fe~`RNPQRd;$&g$&u9O5*%HDuR1sfS+
z+_r^5>l;tnX0@6VrT&W+Z26I$un~8p9Our-bn#e0RyIu;m{mx^J1>9H_T|aEJvsfk
zc80THDIbFPx0g&*r^S=8+sfV#J3Ja@Op;^dp2WB5WgYedB*YH;^0Tah^Q8$HO$ySf
zg&Sde`^2OyV#_yVy>FP@L^_BMq&<JspEJfZ?0otn!m`xC64&MN3{3-%e;4|D>^!nW
zOmD`Ka?x|DVSJE3vpc3&$K8XlUaA1)r0MbF5GGx-S@X-WwAADCTXMF0pTxDO_xIY&
zQ){IyWgMk(gO`vZKL%+(KajL{WoTth-uVXOtw(1#alY5`tuMdB>J0EB)85jkBil!#
z{VaZuZ-iKi2bX^r*d0ou(;)H;h03n_>X@kpNMv=t_AZ4L=I?#*8yHd&-qCE!d6D^X
z<&h=aKV^%dNla2IdEGeb=ise;`OXh#s*;yENkayWxLpeUs`*(@qzpAV(9F^2baF>}
zt|<F1;;39IfAW?1-zfaEmtX6Bd_bQC``uWyh+s1M3g_+Zw{D?)j09y@4x!)JuF5-u
z7Dj?b*@$ZB6M$|GK_OE!AAOnyAJeJDE5VILp!d0T$MtJFdOdwnDzb9lKD7i0`ycH}
zh3xEgFF<W3&xgxaUPg|q#_iy?I80%IdA6}_MawiOUAk{isd&~Vs0w7K;N&74+@8~y
zDMxXa-L+)czMs=73z;4=c{*ADSq(<|tNX2uaB)(Qx*+0s754ZIuD)68NPk*Tdw(@`
zHh<gu<yh5LF_&igPkUDU`l`gfoc~Sm%xl%!wZ?pjqm?`TQU6J^aN=(K!542#_xS1V
zTLsCBemXtcQkWmkVQfxR2=1(6Nn@awK1yWkjUv7^Hxy3Cg1L8CBSxBR@SAH0Kw21@
zb-cEQsDZlpw7!AG#l=bV6Va}oT6fI~HR6sF9oe9fqiWuL&X08X_r_#BP+Wz|3IB(z
zw~UIT3%W)JNC*k;?iPZ(y9f8c;O-jSLa^Wj2rj|h-QC^Y-GlpZXL#Q4e($<BzgSs)
z`c&5`*}JN`=f-iQX0dur`!lY2W6VH9>7D1w>tQc+^Y5tA(H7C8dzmQ4k9u@1+7o;C
z(i`YFX=C3KE>84(-LAiFp6&d*=CWX!S{wYU_GdZtGayD6kl2yT!^Z6_hro0C_>Y7D
zkLz#oYnAGye!%IJgl;(9!X^yAB`rg{PAR2v!EU*Y6;%JGQNyPgn)4;I7M`1jt|O-#
zJYv0!#Y&ct8!as*<LCXW<a7SH<7XOixY$usyzl^F>Yf7gW8+Ky2Lz-HlQIpK8#@<V
zJ0tN2*ZrcfI0^v<N4e0RsSNIE>XX^t+WIq_qcTb|kVakj{6yPvLRWUJWm(ku)XqOm
zG=*h6%!FBu3qD@NaFWXC@dH~I3;*~;z3N{)?M7MNO0e$1YpGSX%Z0u^K0Bw7pPB63
zGtzUS|GggC=!b?iNuxDRR5<FS{fpO1>09at|J6w!kkGD0^T(eD(Yw-6r~+Ps0x8l6
zabMMxRNYU6PA!6I2QW<-;)1GwmE}D?rE*Yinys<F1fd8-;l%x!rS_p)$RrIIR#1rT
z^I$4DtsIeeQi=!=3tmaU;^+oDSAeC>;5f}=xrx_!c#W<ey7VS3n#trE9})BSx?;=W
zlVeB4bg>VmNrXg<d+<%SW};>`+l?nN8RZ3WTyerv@PG*;1?k#JvAyA*HSMu?{Ak35
z+YkFYrKHE(KSkXl+f!#x>7On`ea+{P%Aad(=zoMoYByC0p{7il)rgoiC&gb(g_<u*
z6c#aPznWy;cx->Fx?2eXVwXl8j&T^7D#-D;u;jn`q58}7q^#Y~Uw}2wSxM%S_7Jp`
z5HGP5?*&qv6U}cz-!r<5_f17ytu$!L*fr;2rn0e%%|&k!vRg(hUV8Pu9Zrb{LhaZH
z>V3hKhCcgf@T{+U>8G-flzLQj+}fWo&N7}^^hNv0=B5Y4#JX+AtLM{_1DcvMI)B}3
z+OKqOo?GFLS|u`H)VI5OJ|Ar+yovV$?!5x3I1(lmTCX1to65XNKxOY0a)N{Jo;1=D
zG3p&_^Q@1@?}$%aeIL3S71oGGPM3x#o0hU@7)|n)Z{rf9<06X)m(`37magJidZLvZ
zcPl?F3%;VxCpw8eJpnny+?HNgKgp-YRy5W9n@)(dJ@x(jn{E3>kDB_fjKN)gWoY)X
zNmLWRz2x(uiU8AfNI#2usAj9_uaDBYiC(^<yRLR8PeOjBh|-Jz0BMd+zMzth0yzb%
z_EEz-N8ReGrJ;!0KYSdFOUta$!`FXVH4(DrkMU;GOiE83yQPNpmgj!*;+!`x)db>F
zWAorBuxo3!Jydmh1ztD{%ssjBS`oceRaChSMzssF>}KX(1?+KJU#w0qm@(d*BQ?FO
z`VzjX+ql~_lrXLP;h;ICHFyptAFUwfd#oE47r5QzoQyUK?4>TA-&&kx$M-G~u~;i+
z*qva$s=L25do#Y6E$82%gFJsrWS0)RJ&v#QJ2veoD!V<7RIr_7GN`VTQTwr6$@m<_
zUF(=dT@S;=Q-_r{{wPzKzIh?h^j%(`Zh_VP%;`0H3e*w=CWzkfAQ!Qk!ySGV=8l<V
z4E*Owjq{0pwsv=U2!%iQ@IeC4A)>Q*{Q+}*da^`V<;+)nzJb|OaP(~1PN^@!Dhrx$
zzW0-M;zPXS^5`-q|E|&2x{s~=mQX2r@XGDcDTvX|v$5<v!~hKO0kPwLQ+R2eT&xqE
z1H&+r4Mp-g`@N52VcF`Sr8=zW+sGlP2q_e4S+kEx#H4##l2~b98cy`RYco>IG7@n0
z-O-7T<QQgeXdJxyLl|eh%~Fr(YZ#K2+VX3B@gs>Cl}2t*ll0&l=|n~nxS2|Exw`l7
z+CynhxKB?;@>wr@k*=}dl_p<xX!Nf=T^=+mK~1-~k)2qS#l(8b64o!pAi|0>-UxW+
zjK9?QZH8GRBbT&X`nwmE*W0{K(uRJK-<yjsKukr^l(1B-xtS?AiO;!CJt2S5h$agd
z++sz<Fkt$a8F%_AC^&=K|9p{plj$|xD9yWvo#j3q*02p#NJYz~NqPS_&hHZ=ep>vv
z3Lo2BQ0(008fy8nS&No9XtOH}Lz0iu+9hDX(T`tH7w<K;%Y(gPy;@<>q=lYc@^*)~
z;E3cmYJHhh;@BbRQree~{VEXs>Ef6lk_={QU`5pfQ&Fn8M0gsJx}oeT1k^tVR`s-4
zUMMfq5T71G|LMMI)uz)hf&CU~-`c|EJ$$f}vBQ0!&B5LkAH}?%-(R^bbq-2F6vLrA
z!af?tMEo+PgRO`6bb2l_p26_EW8ABGqb5i*P4QqVo8e(3|1kM8*-4Mc^E`6|n|h+n
zrBtx}wtT$4{7fL^2seX)<z<HVc=T7Ap?4vd3(ay%D|>Abdyk6q$@Npp74jVh{lk&q
zdj+2g!(7Ty_)owc?<4zTA@}$G(rg+6t96^k^R^Ul*J$-Oz7;wO_7Ym;CM2@3nHbdT
zuNQ%QGQm!grh>J5wGC%>KRXst{itQ==2fNUM$QgdO6&glM@ZH=w{wfOzz};m4jz$U
zDSb|{9&PsY$)ndC61;f^HZI1`i$?^z?v{JLlzEOOxS(9bsWJE8VQ#*{LzFp!q%l)?
zqZ%H?G$gsv_TdA7iAbC0`@uZ5TGrYjl>usC)02lsY3CARm9P%aM4F37b`VM%yN2`S
zJRo+}yC){`ASKP(!RWA<n!;E5rn+r^G?9I~Taz`pXt;eT_RfUd7KVgPq}i@h#dqm;
z;^ESWy6i+DVoXd{Tg<)U-&d`#HM`UajVjo$-@P()(Cj^{Zi?WF9(MfB*4=kf+FKo}
zy8~xiz5g0qM8s=0Ktqwj+;wZ;J=EjNyXbY6lvrsXEe-shTljOqLF@&3dNr$WHuH6R
z9(W3yZmwSYYph)5k3#mzn{Mzi1!uOUnR=Vus`SZt+w4PZj_||mTr(1MhV0Gh*q<6_
zy=4~NW4lj40saKg%80#FI&m5<4b7L%*iS3igrtBUN+xVcNO7|$BJpC<sE7cZAl>#5
zB_0ukPdewqr=Oi?oUxE=#A>U0c|Mrf|10_X`xP#K{*KMUcdk;><qz9Uo5O+v-tE1e
zy0s3*kMRVQt<yu_`4&ky()HB)Z<)3dLfW}xc6MPmwOrS*PrRBi+2vQ2Z46%)_C}`g
zX6t|sgCsf_djwt=6@Bcdu0;}}i%+dAYYfSuFMsBACSuoLVq<5@&0L4u{{?27VD78l
z)$l8#7`C=)Ic*$JzsO+gELg}KJXi}_+nphpfzki>v?2@h!<n!5L>Gp{VlW+q$^#B-
zUP0Bx9q1I*vhO`4WBJ4lk1c>yDQD(>{u%~+ZN~?7>Rp6PyR#<=t_x9Io}{i1fh!}2
zl~(%!3Rj1R>>P%3E52-}tM-{jhCbg!nK><$<QFHyFsTIuFXy&pUVrkBvKQEx;rmoC
zI9AO#E%Bi)nvNo|8rno974v*=>@l!%Uy(}s+%+%SK_*sCS)G$rT>cl-<(f;Gx7la*
zjr~gsfozf5fizt9e7JIDyK_-dfw0f4gnVqEmY$<ZOomZy?Ai<GamYMf=gx}G?`vql
zb!B}{GD;QQR}Q!GAS&%O1VEV-B^Sc7nl)cYbTn14<QyHa?W;vvTgSp2GV3Ikt_AlJ
z^G@D*yJE$wh@ULlb2BEZnI>aOY9$lHZC4UB0SCI{-E;q|YDR{hHwUe+q{PZ6<5mWT
zf{iId2S?XEtaiiY--J4hn)TGn3@d+#`IqwxOEin8&oT(H7>t=efXewdhLG-Gy`ovj
z3x7M*;^&mq_aD<73zU?%telCRoRs~-3})2flNWGkSPK3K*s#PtPO)0zZD~A-Bk=_v
zk`7W&|0+|Q+GpFE8{ytm-qn7CB+P1)x}5sFszG7;$UUA(r!Xc)8!F$0Vih&u)d98V
z9p~~HKasu(;`UXy;~qM}k01N39H&Ym19c|~DNDx{)s4mlv@1ZX;gv3|_sNn$eYe`n
zsR_;9Z7ucJjjr=AHV)sfNgndEYn@Dc2@`PB@Ce7OmK(3}ux%US*Lx5Mw=K+4yzKL^
z0TPo$R~BpdWVy0DM0y_pMh3n<t6|k*@u}+ZaSccRGVPxD0_Rn36|+v0#diI|(&QN$
z|JKThQp8jCk4Pu_UiT=OnHsT*@rbq))qa1FTawiyB8`_U<?6WElJAOQq~Y^+nYmf)
z{ZMag<Kgju&xQwb4PRN?o{cJvN#?cPme)LxY_lbU&nlhgzz@CEn7NJHE%!E9oJu!a
zJ4(46W{l(1sJC$Dfk|RN_v1>i`(va>P8vY@kKFjPV6mcEHK;{zF~8g8>9E#au6JPu
z!tae$^UT|)vUZqpEvh<-6Z*<y0Y*J+of?^S*Q!A4po!;XQjgo;Y-0)3?bau4B}!Om
zpj$##YmM~JpUS-oS-FAYy}SH~%$#AP8iSQEpj)Y-V!gDrN!qmCl0;O8`4dG(@xvy&
z&=D2BHDBXFy7b%Ukd}I6HLY?3s;s{Q>?&?^CEW#e#bYkhBq{`tvJ-#gYB3Dlje{!S
zWiZrYjbK`7Z#KJ@u6mu_pB(E~qZ>YXE)BfIc-sD%nfk-b>3)O=jDj=ODz~vMUeGYt
zU@=%yUdG+13oT_ldC8f#6ESe@x#4TDs=)jXN?)rioSbQO(nKc?@_m=FQ&cwBlMW<S
zHTe?^&whXB2d2t@GtxA^d?XF)@^6$4LVtaZzJVs$bsqd;{@RpJaXL0|CQ6Y0x_wvf
zrI=e{!|4&i_u}rtKZ&EYc3(Ed7N-TODz+c=R-rdbPZT)}Lzc;~R#9bhKEjpNvuhI@
zI4)D&!4?i3yQQd1#Vs7n$fX+ErJw!UI<d?tF;Y!m%~qdNl^+!8RJ(-xJ3hV4&2^>T
z%ONFxz4qHn0`cj6v!SweRysks=TwH5?|9hqxKBC`3$G%^&or9=o?;h_!)c?D<*X=c
z2TKB(1^z%D*G=G)%aB(~DPu9HdOub^e|~JICzO|twTS(eMzLI3lUS`!;aBntnB{bu
z4HXxWtkTqfrr|KyZ`6MNfZJ}bzz7Hr^-wdj6@?Fz2jWl(vjJ9)g}hb}N7iWVYcETC
zv5A*8<Q7M}cj3t>vUG9F%XV+>s&5hPX;b@!;LB+iDq2TCUdVc%ld))$Q1=+CO<HuM
zB_xgR=v0B@y|Z*76#8=*h2DKK>=^vdVO?#ZQ9d+vKmnbZ)^A@WsX+X#92ZrZhn1X1
zh)tux9Jmo~aTfz7W`BUq->+odNo+jstVSQaBU*<J5$ej)m!9~!Erl(k2X^W4Tk)Rk
zT+L+0m)o$&OjkH__53OZ)?06jORXq$<Lezwa_jP+MO7*YNe|xx4vbu-|4fQai9o+6
z;K+gkkSom>)e#oCD9YywnJHxGg2q|%@n?|oaCsPaQ-QTK$IENzwQ;3pG~FbGta@Yo
z?Hk?b{l#%{*@emSHxFS6>5um<H#?<gi}aejGFoh~A4a({(psOVWlFvKI5BMQ3%^a~
zx5%rNrfU;Fo;nfzyZKaL{$<JCht|gc@4=qzPV9Amg(L3d5AroUI<4*MN^jkTc6%*j
zxi4i9??9Bo*fYa!#mK?z!g+PHujk-xEE?_E;jA5nQ)ysh^*r5Tlf$HPrP<#FQEU3w
zy-m|14R@c6FBF3+>5toK>+N+rN@rXzffiF8yQ6>&4;!u3!em&U2;-m+yx!J7>@0sU
z^P&)&X@xP4Evs_P`WNul{Nc&pw=8*XSD^iv!m3Axa8vdCXj)KU$GddEmHXnjoSwnY
zjV7{Vd8$@Mu0D`)FCq`e$}NR+*3GIHX02G~iZzUx@WiRG`h4k&+$x)>;vb*{UH4-U
zrv+EX-FdAS|MIdKi$NdV(A-M>)aoSBI)x^o7mT5%?OvXnb@toSiITruMG*LUCIe3`
zBLuO`)#cA;Y2uE00$aWD-1A86Y!0-CkE46A1wLsyx0S6e`&x#V>%B6fO?Rv_Q&G1C
z%Nl*lc4iNI^*kE64v#ydlZw9dy(5|`N|^~b-X5&S+;i{{XfCgEp>MFE>{s_WC9SP0
z%Bu}D{3ixo`HkdAKDQ+gHc^Tj)Cy@RNSo+MvzxGue!9f=yreoY-~xnGrrlQACbANh
znsXg#Yxb&5uj&?5*N&Q5rO|se);&0$ex~;_4E;^P5ZZntwC0+E+<MVPf6k0cck+Q;
z(d%!8<>PnFaX6E)bf57e8~WGdwjevKu;y5!$j_#cpYzPJXVCL*L^lw$(fdQ@>?PVI
zf>Xj;Hj?O}^h&}`Lty%+%jZvms!EibeG>84y>@^6%N2*WHsO0;h2vLp?#M1@v(U4G
z&tL}&z@AnNBLCwY5Z#N~`3n{RKv<@yIff3HTAI5g9`^|{4_7=3Tt#%TyJGY?E|~wk
ztFgZR(Ec;aXI`W@^m5Jj`RWeX#475Y0nJ~(cboFPmL*%wDF22C>g)XRhw73NPhH;c
ze}$*dVc)lAeLk<o7SQo+an;`G#V39|8m+$M33g>OXP0|;zpVS@VMJTUbSJ^O86yI)
zFVyK(TgQ&;e+WxYjNN$FcJ*~d$@FKw%2+PTs5usRVq7<X`N16j61*J5Z>K5RiNBn|
zf_ga;D)$@8VcWcP{*}D@1&{K2FF7MbX0}dQg9aCii!+XqX<bRzz3nWH5`Bl+UyLuy
zy!<qS>8e4POavo&N!UZ5s3v^a3d6&xH-N3ik{~1S<~A7}AEk>*Op*q;^c5ahuFdXY
z0Ai&OseCZ~So&2>Z~6OI`NsJVE$*UNZT19R5byaKH~GDyBp#)TLYX56_u>j40Jz3x
zh}7XRx*2oQvl(2v-m)RCOiI0&a;)bopP8FF+;MKOF}dW26rT^Qc&{VP4h~9Mk4pFb
z;~m^1G&C2nb_=(MR768{HAU4mVO)kBJ|=L&{Gd%MD6^VAGKL;6^dSR-BYxywW_=tz
z{-HH4Cz}cdx9qChWtG<+Bm{tz2h>I~(btrLwkbT3fvV5Ak=@L)E9CY$EjUc!he-W7
ztn-Xs@xo?=ZO=}^r|!@7?;*`t`_Y{l_Sr(77d|d&G(mz;BcVBXKpELh$hRafWYWBG
z<o4W2R}vBc6GcNlHbC1{Pr7#3R+lG&BRmcp&<E|pfs<h?rR|)#<pw^V`F=sIgo|sg
zh1m&Bx*y9ONxA_0h%Qs{!mPA$)4Z=RLGC1xP%!e>kxFx6P~G>-y9aEKP2pzWWXA?+
zpBU8ByV`XCkh+4AjjIDYsj-q9Q;KlE9~Z1~D`3MBO61U{+D8uVVQYZbvJ?#p3P8Pi
zX$r9SOLn=*?NjIQZ(~vw2}-$y?)cU*2D`o3Wx<(7E2vJ<e!n|;N34Af4Gj&^k$(6I
zLm62B9o!$cD$Q;n^r#%M1swXsgdKN=d2tjF(ZHIMTjR{}jmu5@g3dNmtGX8{3#-Y?
zoz~g=ePG`z>f!P7r%6jxun~2i`P>%ZFZ7vbdH8>RH0tp0O&%#8->+C==dp%micd+k
z!ag|0w3j>zVnk|xxw$~jm&$|Z_gZ1!7oUUQV0cTiGwCsw{)2413+oB}xVIt{UdUo-
zeSH+`3nJdF_A;D7l?32FwA*m=d+9<FM$abdUiSxhZJWy{;SksBf5;(8GXiAaS@Ju#
z6h*$(3SxF;(Vr4!-@ybi<K%^SQ(Xz9pzTF5G9&_0R`?GU-0WK%i>sorP8`QE_bc4)
zuc1;1APgA}DSp-*v9AX}T-f16F`PvFh51z=B$f<E!u4+wXWd}SID-sau7O53lRC_3
z%V0!bJl*v_&6L~Eq&r{>v{q+t*leT1dbcSlfndb_>%A}+gE;so%knvKkpB#SB$Xk{
zCD)Eq-QX5vLm>^E<Pc=D4-lLPcSQyLH2Fw6(8m-$T0uuj(WS@q72oNrn1r%hlf?SP
z!zSnIKiDO=MLKH$PM@Px^k|>2JJ@VDJS2;bmYg5$e{0l4+3&~zV^L0fO*hxEf+q1>
z7E0JZPA)hRw+bljbP5yb*M#tKMd@a66{(fTd5(nb>wNRv)L%nBt9z_oX*9D){H08W
zO$G<HjgH=Ms%g-F?vo^gb+Fy({lx~>UtZKD$bbzNHtkk*34s`y!9}<7$K500)i(qJ
zn2yhe%m?M`%nEckt0@Q%+MTOczA|tRn<Md$b$IomLdG34?-mKX9t?3(NQF75S@gfO
zxr1T%kFc~d8|$NqcW#$?AMTqhlT=<>UmKWu!@9fL84=0|TewOD8Dx`=&LjBe$!KO?
z{hL|E_&7wMvW0t$eC=H&ln=gLCv)(BKWj<Ezm9W}nWM@cJWzM|#`~(ofv%y}QDDT%
zZM82L%7yUr0!$0YKLMB!fD9OQ7B|IPU7`VA*bH;)u?bjYN|rZjHm7cVq~Qb+5>ul7
z>8YIjVmpE63$efW5~u=d3Ti14ZF&OI5VdRjC*vst+I;$33ILOdYIx+*BMcz%vI2dU
z#7V!8XXuDf<6tlvqw@^!!i3YC78xCE=f=(8MkT`hEl&1c@B={9wXpK$JpVF)V6HgK
zl=U#%aa5{S**Qnsr?9QDPG>bU|DVo#(XZe6zm1SEj$1&r9U7n&{)kFM$f{9lJT`nt
zx(vG9bNM3?1p|X4ku|(u)#cTy+|@IT)-6%KUnHzu#(j6E!^m_nf^F;0xj?;Ay5&@K
zzae=0DlTbJQTSuHGHC@d$u|+*nO_6~^|jAO(XruEmeu7D+YNeNIA8#P@D*?X<BNNZ
zo#EWN<7^-P%eq=*GfQuxN9cWS*h}LjyCD*i=`=UyD|=+xehGmDAG5OLZW74%8)W;|
zS<J4$PWKJ~DC4Hjeehl9pDHNP;V`-xr8!=y2&BAuSnO@zQA;aRQMI~ByC2X$D8!E#
z-g*z`zim=&Dyf5wixM0Sx##5N#o{pcWuA#r70{>~1rY$S&nR)6jx;k5;B%c`HUO!h
zpa3d_&PM@EY!QUXSI%;>Jf}?+Bx2wvSdlaqT0smc4yW4Q6SX^_A^e}1$90RLfh$L$
zhDWlsWk3x8Z0M>}t&(3LmD&}9KqKv_h;Lz6bgpbRqrmnrY84k5&LjW;aP2m(m^0&(
z$!A5{d-+B&E-nNBfTIn<D5wgrb&^YE3wz*Bhz$VfKn%PWr_Xqx6_L##T+0#mReBq;
zO3lwj(8P2U89?M3g$Dp^trxGGMV_-#mXBXuFNwi81TxuVW}-I0dp(1pXAn>ifNXll
z-;v8Y2ToOk`h>JAX&TOWP*4!*?44m_4}~-lNsTSy5Fi(FQlkd0#8joDT=xEU<*E4s
z);(Cb4z<*hGP~J>pDlb_&#kBs_4$Tv*+qN`w>vgp^LX|K>j`q1zHGn=Z7FF0U0BQ{
z3K9T-gKW;W<PD#0%G#JOgR~n*2!2?wqXG~PM0A{NJcE&=V$pV@%)t-qw#gX1!h9zz
zGN94FW_7L#&W>Om0|=IZS~h1jnnQm##TdcHL=EOHebTXN$#t&;FT7b$Z-a1tTUaIz
zVN<cava#<;@Bjd&vz1s7MbxQ#=2lNgmN!@@;86Ud%_Tt<Riq3Vd5Z}b0@uQM*V27l
zM?C{jE=LEWfU|ONaBtTUIjTw8dc&s(>)~JpKt$tW=d_#*pKwp(x!Dh17h!vQg6jnp
zZ5uKQIjn<TQ&PyDJ};J_ZLhL_j&kzqAS5vuePsi(cK{}oK#2sMWug4v{{s@kt#tV&
zmumV2xc|M@nfT&S%iUs2^)*1vj+(g7nw8DySSqob-{5(XF&VSr1rJ_qhGoO#_sXc`
zv~9wLRR3m8lKlk@;JTLcIRs55FLr|N#>2(I#-b0s;B^q8())o7CXc_`5Z$JPNLq1S
zz543eH*ClKmtccpI0dK=*Y!P|gXz{04rBW>OBMj|;(So(o57e9o_E8V2iIr~aWtwQ
zG%<(NVaNlRm;_jiEF`uFmLsJ+t(3>761fDOALcF9*;DCxkEr3q);!f9@P&X^Rkt`3
zVfwih90DQ!LIsPA1eqa16Y=4BO43&1hQ&SC=ogs5uL($G(zk#3%W2-ub8BEbYjLs!
z@uPujiS5oyEs=q#&op4R2tXfXv-s+9;yCno-Rl4`dy>fmbGCl=`FuxEt|wpzq1zlO
z?=|h&%o~X4Vgu<+v-KAr{n7F+h!*VCcssCn)s|x~ZMuGc<TPB;*Lgf@pS^60k@nq~
z!$E&p&3y&xF`wAMCyfIg8v?dgj*PuOZzp(R|E=W-i%$;DX8)lSyjCJRv}kXN%$<RS
z1;L0FW6-HMS*Fs89O7PSe>gb`*g-@qNq%jRQ<arsjtLkE07gi3@0G~K37=o??yH-m
zFk1KhFbEIgTUiW)cy9f?SF)>gPMm=HxFh7tj-H(PCGZ%7JEfR{=`SO@@a@XMA22!-
zzZ4CNqd?M`%BN(E-A{k8Q~rEG6nZ;kBj$x;mh{Z&gD0nU+QM<y)2@`IWBW7I?r%tP
zr#B0QzagvM?3{)CrzV1zAVeG3dx9p1;j>h+4&aPNpX;z0Gb+<PnupU@!AqN9+W|R(
zt|IlKD|}3CDspnwpAZ!qS9tAd71|kRSGg<~Dxf0bw9D!gGII5a!2vo<H5@QWjfJKD
zet_2l`#2K~I6d6pj)QK3`;qiO?xqVZg`Lu6gF><cLV%U`W7p6^^?OL8n(pNO@<0rt
zaLDa<P*2e7SRRPLp^ez5StZ|PQrV%CKB$uTEfQKPSU`Hh@4{N$3Em9q4?dbkU|4=#
z5&o;^7Wsd%Mpo!c8putJKO!2REa&YEnv4wbymk6m>8SR1{=8s+kq`IsJCMccUO-;>
z4OUEtJ{XG~A%L-11_>kb+kG<u2LJwpzvGjxYIu~;&`bt@%kcmtJpLDo;`F3DK~kxE
z?*Tsstu|{?#xw{ROIrWUR^Ab5UkCrD9~o6>mY!d^)?#U=#V#20-A}usroN+&GMBZ7
zOe^h3<!a|<XL-Gw6cBmaj2<s@wRyvWVVOMtM9O85kzCe2nB{)#SCf-PS5vNvErMaV
zT!JK3+l@3LXw5IS4W{5D8X^uHSbwrE60sJ(7T_RFBYy=A=*Zb0#$IQ3?ZysX2R+^g
zGslSi4nBvQ(lvVs6!f|~aT`AVtbDI)aDh`oW8!T<@(ve03zD3vs!o#o_T+6yG7?ZW
z!obG>$myly)^nk&|MC<@e$H37GoGDsb!*6^9hM5OxmFt$)PK?=sr=jmm-#h}<Gie_
zZ+}Cip#@jJ3KEnBMrpH2vA{Pmb<G#t?7V1Fb$&GOEZf|Z>puBDFW!2ObZWQ|vX);7
zl<ZytobN1X{*GCYl77Cayn19qZsi2~(T~kx8}mZ-B60286Mbfg>+6&LZ}e+hPV9P0
zy!DcS&Zp^kWxlZVm)#|563@9o6B}#*4I}AWmLdaNa5S-z=x^UeQzm1HA;&VmCIlk2
zbC)=oiyrt5u>Ew`nc1JDI3F4!%$l@l%xDM`4lXOvhQ9;rRo5i+HY=WkDE2oI+AoMg
zt}8iCCuuKRXfMM)T={&%@IC&fw#@0g3AdLE8^{zQuiN|-Q{|6>M%8LlN7otEPZ6M8
zgQffVVUT5wS^l|;2A;eXh6_Oi;Z&`5QBwnnCM3k6)+OM5nbfC<tE=myuQ&*4&T`mJ
zL&)*PgsR3inMGAc`o9~0WV|94W=>P23&_Gw78N_m_>S3@xLD|%5hI!b2QI>+uCAsn
zU`&Wsjc|wzSYh^2shW=wXXEq|@_W~Qfe$vw%oXgBJ7BoN5bO8$7@LErD`}u*<h!=(
z?j-t)DorrYQ)JPYhO1iBy~@ku^NJ0Sft!RqNEa-8I*mB)<~xbpq`aVUsGrw)II&w?
zek$*DF7$1sAzScKjr%_5JXm~3_?R!D0N3vj0qSgbC$UKq47*Nu*n7skV^@E@_L<us
zTARJLG>EtC9!2iIY;>76$`?0TTXKAE)v=qsw%Yo2<oIgNHQ6=8-CiAcF75Bla;6z)
z;GMtt(7pOJvp^{M4ou8tm2Zy5;BQ9h-bQsGUel4;6U}{P7=J)>3dUYIa($V`WugZo
z1?T*^t@%d?tNRaJ!o{GfBR^f-I5qui@A1*pd9%Ji-1FPPiza3-Y(@_s*!Fypr^YTL
zJs$&ub%Jtgn7FkXhh7yQm|CWdI*mELMc(aNO2@18#>Vq-@}l9F>`iTWIV%6*I!XvY
zy1k84g@8OPl8p)o?)l-jlGM~0%*bFzx8k(fE;`^I>_VPr5$#Ujght+!FX_5gN>k3|
zvoqAF?mey@)1Wt6lQPVB8O13RSXT(E_K@_fWA<gJHZu0hakv2TtDa+vjfmChg&lJ1
z<6T-gC^5q-<u_^S-9`ypKfUq8rzT=|nL(@Zv=KpF$ak%S1<&WDX&8$ClhSH8VTXQq
zSB{;qj#%w_boOR<R!+yi<<Kw?%k?@EgTW3qF+5IdweSd*Ct?MK6oL}~u$?CEE6p2)
zt}Y-S$+3iwq)s^ksPg%(qSC+1IT-y>{C_hY^i0IR43^D*8ZF$dQ~8Lzdi&M-uJtdJ
zbfynCX2jlny3Cf@t+7=uP?sB|a+iW$eWnMpZo!^5h^QaeF>l^2+&GaT)*GvBVX)$U
zk=ZOtu+=%X@i`{;R=GdSHi-ldu+?NNSM85C##?g+k0f19CLjDblV$V>^c_2tNSNIP
zO*VMI!!x$u)(`7Mw7f<#ZjB~<bDgnU|9(Jy;=vHW+6suBY?}gu<v5H3>f0Si@&@>a
zAtqNc+2(di*$7<2n&hO)8wbxHp8Zg1W+{jlEw}cqjjobwYT$lKEEN*c7rArGVE?E)
zSGnXwO!Qi5W}$kUFDT%h7vZQ_`1(@q-83y|^HNW&*S1iMd-Up^4zwL5&0-OGA*io7
zU9b`a(P&JwZWpTfo*w=RVR!ax4<3Fq@5QdN&VB}QqwG6V2rwbT8?tW@k%!YH2$^mH
z%P8Myb8FYrwUHcOMaZq!z_)wq<Z$&P7GXEz<k=H=UYLATVtLG;nVv4Ns&N%+Jg1u^
zb+b#1_o%M%(0e;yjOYzK-Zx2+iFH0221U@d*38dYpl6((Rs^Qh+a2kx{x$?6tILLb
zUQlyx^kqx6&SY)(BH1PUsJ({3bZ87F1T>;08K1SmL=^IMZvWApl3};Kh*O30)_?7X
zq=lPX->9e74Dj`Ji&_^ub9mm1=Y;ECJ;K?%Zp~tUVdT<^_ir@x7cH@CWX%`sQofPK
zybj2q!)ZxV^-Rm|o3E=yF<cLc8>IZ7&9fJOcauS4d{%fKj0t-{bnY8pVZuP&D&(Xn
zTXNb=8P!bDZMhM}aw_roS(7XH-OOo6-|jV?6E>gKs(JS*n4xXVnO=2n|BLibN^y)2
z5;3%a*%}E?l13d~Sng*9ox7ue*x|t=k3b1(Ot-!VPjZMMh*q&YtdX2o1l1j#%fCKZ
zhK!@6YtER>wY@iQrMlhL*VP%1=qDiRHg>cv)5X+5G&%U*6i9_GYRpSxIbk7L&2Seb
z(cVN5dCoh{#6fZ32eGzaA<?b5vwPDmERdL_KM*|jeeW@FyzO?|&bRS;JG^2BjG6sk
zrU{1x4?5RUVD*XyJ9%SIFEwkGkHymR1hxyV`d?iN(h%GjmK8N7v=6_-f1IrQ^09KK
z)y5{G^<2o=1`_P20q3%!i(+pWXv0g#DXnNap<nvrr6|O4gM|0+-M2?-m&biTnsY#W
zyVu0wtq751Lp<TzX|=zi*<=BBKEtSPZZT!`-rsq?m7H={is|S&6{)gb2jO2M0S|~{
z%z39}aM^TBBy_=A>^{UI8zr=XBy_Yzpt3leCK^rdv^fMyb^N1C@zE&ol?^Kb`BJ|!
zOMPKok$08zcO5()xr+RPk6%jOpwgXplMzVXNSnOuTJ{rf`S9yzcmYkfg&-%3^ub>t
z<!@|<(jL4zD*x?tdwh%gU(Pq6v9zVNu-UIpx-$0rZ!_e-hv8rVS%Gt66xi%kT&0t7
zi=^IB=#GEH;QDP{#_&A!fHiD<mC=-8?>eKnb;{+!?%Q)Yc~JGIuH9il+1=}Oisu_=
zsJhguXCdjvFo-KgXL`-?bET8<)xbft*$-(jA9w#;L^!#Jd@LR9^QT5`6GGA*LCG~G
zt7XXsw&`DAdGPYrRLqC3aFn9xe3x!Vf))yWS8AT+{kn5Gy}U|#m+V|%Z@?J<AT1`W
zx{q#kv|Zm4bFi2nz{Px-nQ!}OxZSmG4{J##+H^a{c!qH7lY=hlw(K-)dr={OQoy6W
z(bTjeh?hs|wM4b-osW=FgDJ3gT$z+v8G|cwVn-h2=>bI*nC#-LD@qF?pOD@l03czj
z^jx}68;L!?Yr_xfe3|E8XRKyM<b)z>!PkwosGu9-uzR_^;%42abT&&DfcJGBUy0|y
zOgP`~v~L(WGx2gYR9L#)jmL=G{Bzx5v&_3xY=uz@3#)LjOMQ#xJU*JSr68*Y!=W6P
z4(e;s4CUaT?hmu$)V;VmEUlYs$^yGXP%J$~KpZw(k)cVveMmgrD_O11?xI?tX}@c+
zOL1j)1f2XmK=zYoXt7*CPVl?SMY=aH*#T4krf@SZy0xSjQM2L5aC6P@&gl&zBwiF3
zH>~1J<897dIT!K<fv$9d@rf9xCf!Z_y7(xN%YA`2azByol)0vewPIXP{5Gmc&FB!6
z)<R}|8M?*{?{=>I>t#P(E$uNR=B|r~41iK!T%4KTud10}FDyd(?`pL!mI}l8xNKUc
zZcGqhx$(tE*RObwPYCYeXu)%6Y;LXa2qk@{*xHWu+x|2T&y(ydehKy>-ZhV8wN=Rr
zf(lRC%YC=x`LZK(e)mNFY!U7ISI}uBHlwB|T^`$)r$)RXJ@3cn_Kf0kRxT}*(!9gV
zKg4UG%~_5ojs<i17n9>x!ZMp9UX7>Sb6?P`nVN~WmOwb<{JU*QEuQw*kzTV-RASpI
zmA0?z-od<xnTUJ`P&K{ke&JSB=jHtNe4L~LsZs&vJzVdIFN?>ig7O5a?kSfCUZ1=k
zCgQs=!WzZ$Eff`zs2m#&p-8{VbpysnH^Uly0@!E>_`>t8ieyvCt}ZEI({we`fSzM@
zD5#v)8;v8Zk9{9J9#R<5ri*y-cIo%=2-q37n;6gILPl_xdgtpvZ9^+^m#V}`oGGGU
z7+Y#qq*_#5+FlKb>e>?`>-R#)P^3J6s-7KzjaT@SAMQZmFR0Rt9mX@6`RRhurI`r$
z*c8(5)8XL2o?=BaGg@xFJW;5!TEB5Plh`|#bJsVB?iYUCRz=xbZtdL$wcNvg+1bjA
z`Yw3lH)hAXI~n)VAeNr6*54P<06;kG@BH&4_b6+D>qlP&9}IlH?&WD;3PQ~YU#?tO
z+Yb;7Z|vcJp|43#YHX47R=~ocbjm8rK#k9vboPqBiT@ZS^Zu>-vG+|)r&*quYODW#
z3)c>OVfKN>74g1P;QQMBs!P{I$waDYeUpmRtv>rwLAV1%2ps(e=3ROa-ld5FAzpxR
z2F9{~XAkW!Zm$__!=QOKi_4s*zJ%TlExk&vu^rukF-Yug=+SgnF$$*Yh~*^jKz@_S
zhP7*NG(_ur(S65WiCnB@FLH+4Dh5lSxKzrPDkYi#T|lD0)|-T;kk?zll#?bRNFTf?
zJg|9>U;A^rqE?22kzCe}LN-qw<13DUw>_{SA6wsW$j-XdT;(HIiE80nT@eSL$Qsmo
zs6d7Uuzha^jwN~g;Ea`o+9}K9vkVp<hu2|h=XzmR@@pMA2EZ0tvKxS|Xp(Z|_&+Z(
zKq_YdBkS4JC)9V4CKgku3$_NT2#W|M5uNrZ*oT<rxaf#}0yYHfg%Fq~*__QCkCGF=
zFJ=AHhxq$OGE>q-81s+jjwIb5A*TP0b${(8W59<92lye41d%Dur4%IQ3?}WpsLGI4
z0N5@Q4~HbNObg<1+D%Ln1XQLF_34$rWz!p?@4-UEcKTiZHw>%M^;Q)HGlGXh;CaaP
zvB#`~<2EPbTvR*zuZHbMZ0sPZ(-_>*-V+Uh0YQIc?C9&rUrp6nwWlr0YJVr$aU<^O
zKL+_p<U2K33qlFb0lB(kv5%H=t!PBm8&gpnMJe8nSX<@HTZVCXZkBS`e>+~TmEklN
zNN$24Qgez8s(pEoiF@hk@U#PwjOHtlbw?nCx5!^_bT~o9=;Em6tLsv}>J~PID|W!W
zLg7UuBgr`j=QWjCwEqxPqiX!q@35#6h@>%Bc%xbqMKT5a@eQm1?deq4mG8#6l(q|p
z>RyI!T)_tN8e%G44S9TO<7~PE{<pa0>e*y;0>Q$G2(Q9>&=jpqiF0*MChU4R5j+9V
z+A@Vh?b^x&L|#$#_Im>frcD1cX(%y>0)$w_I8TTlO8hQ+?m0g_C0XY!&uh=sIYugJ
z4=s;5l%7t4!}ci5=zA#-gNpF_oFlzoFyUy~N*c;H<x-a^zxP`g*%9f}mPi<*+$xcw
z!T8NSA|ittt8usc&rB&^s6BmQ=-1NvD6<U=JD)q5(Z$jPdd`bFy~2m2=#Cz!eTcL_
z<V#yh_MA=_n3jet1@;|bBpuVL2cL7uL<VA{Xb_>^0{}u}ylp_>|9<Pl!mM1H==rQB
zz95BhmYWxw<S-CQb+CT3$P@=gBD?Hs8$H#@qyZyoeOqo_Yj?_UIw73|AJ|5trR0C)
z)fG7c3B%9+BFrVg>b&+LGl7@Cmz+u4{i-Ig1uzmV9MW)&_q^;1;(SVv`681)ihw5%
z`8b$tl~~ohps%#@eW;jb4QG@jLfUE3Qq_>r2TKLLT(vxbVN|hTI)zWa{%t;8EXo@G
zRV@l7d)nnn477hZgg3$xdwS=H8!$x!TpnVKKN+xyy(>TcLMf*cwtG1}ol{AKkd}3G
z@_UKTkW9N=ll2)!RZBWCpFQ>>JP|)mI#Iieq^Pgk{q$>@#(LvpZ}9(pZvu_<f4w*H
zBFJ0z*(sAPRs1{q?D~EU%}2qjvO~^*f9^ehOG*STws#5flBccSNqMZQSS@qEgX(jO
z#n+tzE)}&*WnCLH?+-(xN<jQEMW1@q?PY#&FjaRtX1}LYyR@7)uJ&R$6it$*QAco^
zN+j{ZmEMhvpD%T~KC&mL;ZHHU3DCFg^>3E%Y|Jv2ofg*AK9{lWQuypq5P!Li@9G(u
z@&4E3fv%c;JLEl=R*oE1TlbQm(Sa6>u2Gd+_qwksx$~~=N<I8_r#B#{#c~_{1N119
z5{zn(Swi%f7N!SK(9rqX!ku51Yc<xG_f!>Ish9?kZ8vx1yXm+#4WSxuU2#y&nf>Sl
z>QBE!VnySw9|j{q27>5T^$)gdfd*(uv3~s6VCaFi*Fq&yHhzD(*@jz=No#Mg2c#88
z%AY)O)E#VE2Bja8Y}alDl{b11%bJ$6Es6~c?PlB#KQ_6CvYK9aBDtp0SZJ1v{XI<o
zn_p56>wQ#~FLvNoc2<<Vy$*6RT@Rm^AaGnRu5Txpc*-fUu^<?<d<vhBF)l77EkV@=
za!vD+M3h0zlXWE8If|<1A2le>k7P}kBg0D|fK7CT8kJCH4Yj;H<Vd=;CiJ;G5Su2_
z?d!q`-0$3SBcB#F5Vd%Ecf1zx_}GRwwnzg*Ep?=nUJ0Lzv?_e2z}$+8u9=dCpZe*S
z&3pGw-v*iqo|Me9i9(xw{Njb5xL5fmRG}{?R*qH1>l4gPz;DV9W%0DLIEhegEDFQt
z&yRUSdoCi>0iCLG-WrkvH9k(xSDv?p+#z+Y|DWM`TTS-7&o37<BiQmM=hJ{t6soXs
zHpPHWiU8$0S{x#ubc8l)r-nB}3!W?xM(KFrY~-h$r4pGEfo|cV>p=m>YMZLH%8~CB
zR;JX;Ycp6r6&C|_XPjogTsQm`7&Hlc$Ud3GZmZvfTU)AG{mMUDIK1QBCqAQ5yqxv@
zl>V&W`&Gsf2cvG_X~qC(rX=E5ev@lCk8D-1NyTQmALrF|1+4tm#XV(<XDJ;_PdDUi
zVI)8*_V0dGkRmx&xAyr-@O1TPMAH85Sr)hy>wEN2?^sP=9)^c}KBRZ|sB@>4e)l=O
z>dW4J8(HuJ+W$LHZ>E6{-P&VwN|kB--_z|ke<B5N=1&~R&$3Ne7&`=DKr&ils#s<e
zrRI}MR<i4zaf@lGbB6WviGB>*=XNchTFOCAU(apRV3ngafr*NHiOPo(Dh(2nocv(p
zW%2%?zOJ<wDp!uv`@+Gq)+@>nMt6Td!S)w&|2I+Y@itMuW_$X72FkCXvsuY0#VL_s
z)czY^&&X7qtEFw8a+@zf4d+>zhG&XAWh~AN&od5bZ~uV(pL{t!P<8VU*lqTzIb8XV
z|KJ!6tyS;ngscSp3|i;HQ9&J-)Z+ehT@Nw-xR>sQLc+h;x#tMD9;IKQ#j#zM9&Puk
zUkDf|_BX<i$PSX(e<}sDwwA4A1NxMy#R&_ef$6G{W(z=eX<MprVf|5AZH9!iv7naE
zHb)j?^O#g0HETXbQ~7td=FVl9C7K!c^a+h-<00t!IZV)|xgobjFZ94mdVkx((Xh;-
zO7R%Z_UolniObYeJQM##G9`{z(j)x(pZIHH!5DUE$V`oM@X&5roIz2>XVn2o2oj*O
znJdmO(O5j9WvoYTboV~LG8xrMb9p_fuqU<Zkj<Q@E6sSm(!NAz%`QC`cL=%bqg-1i
z6Rdh!$m=dW?WW;&gFmHCO#Q~#Yz6Gyl)>gDPA#r8kg|F=kHla{<mK7_!yx>FnUchI
zXB8f#IfZ+x29BG9-V_!l6IMY^1Cy+4z{i#5V^AFzk4JQP{p-?hy8AzcVC1?3&p-cu
z=d9Ns^(X9=;OGwe731@bAyFejnH<<>A$9}FIeB7It?l!$^|g7Fbk_4`gW>q|A2&g$
zkRORM`uGf2h<|(XT+2uYJhJdCwi{w>S6tIx&6_<$FrI7qumyW2l9Tmsd>jkJEwER(
zJ6#@zdZX=+6>~o`U7D~?_k~DeP)R6l^zq7v(g2yOm80BZ0I?wtV#|`B1mk}++a^<0
z)%SQMbE4(Dg98%PlLN)Di`4kPsEt_dStei#lTBfivn;T84*S+;%0ln+Nn``u<OXdb
z@LJt=61e75eAax%8IpB|!EHMh)RASj+fC2y#+TNZpMYac{@>(p_GJ>1bA&$K=B4S6
zi+yb!mDZ-|b^jEKi?>?F3fnPZai*?`jMKwMrRub#-kpL%7Gk^Of}+=xhn(r1mEFFV
zpt(M~^@qKANz6Ik2*hq03%r~rKPuFIr(HF*Z;~M~hUlrtB^LkGsK~WkwO7M)Tra&f
z3_6(N&Wpf7C|*A0kPXj86_V%9tPkCnuMElmzf-IJjFM&*&c9~YR$RA!&isWsD=>7w
zAPE1HQ1ea2p>IM)_Y}ul3VmnjtBryeQSw%ocBeQ*g;0MJx=h8BvR1=Jld0d@a)6at
z9FRG~>ut<wjE?%sbn&{>o))!(S-EDF^LAFQ;kZy`@WST#NnZMI*=3QrQ5|dAksILC
z<)emS1Cm+D^J#L1fSdgjD~%P*X>!kvr+x0J*1_3XZ7L(T6nweI)Z(P}B_}ecO6ykl
zIcoxi`;dxn%hl6-Jl|03(Ns|N8P}|5#Af>1=5y#fd&E<kGO7S=Vxk_y4M|$av;lu4
z!8FX^)@wQsD?fSe)4lHhp4MEPE=P-E6)G;7zeuMU0OIq^voKIx1V)Ldr<E7q-Dcoz
zA9GI7?eT_wBz<nk`2b4(t{^H}s6D$$86NuYl$%~#^IkN}!kqF$p)H(4UCMEBi4MFi
zaImu1?Eg3Q8T#MUr$ZfLw;nl!@4ytQ7hT~2xe*?_{+qJ^(Cxe*xBhr<sAH{woIs!;
z80q~{R8zfh+cv<aS8MpSseCh~ck4{0i9l(DKq<TwCzo1QeO(jo(0m?sivNQKT_k##
zCw;2r+RM_RiYOhoKK!ZjXz?+-#&$|M0ZLnqxt~o=z7Q7q!8~a}Ug2m`E~5MDCSG?E
z*^dij1!+uD3VyMz!7fI~&&o}Z(;lbKBN3WUD0)*LV2ZnEwEjDM+;;J1zFTRtny~Ve
zplyEyjlE)GSUfR^sKwoZwlmt-xAfzeqI?(bWyF)_Cd!3j#X*60#ILutW26ZLIe|Jn
zHZn{Hk8~E%jsJIub4cmBCOPl<r%kzcYvpWXE}~Jy23}T@EZ6qgLQ=~O9CUbP-2JG(
ztj*aB=wib{WI(ShW4C*oL>nxmt<{SlOPSj2ny!-}jL1U3!Dka0poPIPW^k|D3JEId
zC%Mw>9$B2X>CJ<$&#Hg+kL5Lo{v`^N5h>O1ri*V0k`Jg1m^s2t&avwe73l8EAgzdG
za%#x_YEo0-?YXHcOM68KQMC5aDp$`fD!1XG%paJ@W(#*F)=A+sHkcQk3d_OnbDer}
z)*bA;Tv9D`dkLrcg$edKUf`Vr>`yS692S-GAneAB@$VRi+K`yxd6_KYNq&$Pht&PR
z|6+8({2-Hnu6!0ZJj5_VpHbn>4gej}2Lqq<2_|sE+vDcZMB>FZ*%;scA7w-Kyu+Mp
z8vbm{TrKUZq_~@>_!AtYF9Wy>_HhENd>FfiZs5%xtZA=FMyX;`4l;k+2#o&|6~|^c
zIGoL49$F6nYubbF7rk-0C%ln`VGr;5wdie}j;#m!ZG|*rhnQnwW$o8OEQdv>pD>VO
z%nt_`GX+5O^;Ywz4qVUBVmQ*IO)$HD;|}53+foG5&?JZAL-6+OyP>x~_c(}RD@oqk
z>40!@HWAS^c%zm<QaCA(-?vSSJfr^d)Nz?c2zz*XE-B}Gqi}!hlMt+_!544C1@}%L
z@S=Rc8J0s52D^&w5bwe2XFI3-o<~;f(6zaUJ?(87|E4Q19rQ4HAO>pq={qp!^+JXJ
z`6W3N_5UpMq<3&XOqZ>5h(`O4l{;uVUnkqOF>yE-5j~b2<;Jky(H`{pZW@l+{3=vt
zT`^+6`o>eHD1s`63LeFAVIc%8NMC~6Vo7KoEtnG~y<4<?2PJ*Z)bQD0Yc&JK;bk@q
z7mWd<r7UFyl@p0}&&5gHBTNK$ZyEPs#T+?ET0CN2@pEkx`s~ejyhc*xP#th?(_cx#
zbYC6cZxbQVAITA85JQQ+)@8MhTqtxD<sRo2PLkkD=sP1cAsDaZQdBSIOpbx}hT(JY
zgcOPI(Q%pBg*m9}FEK=*j7Uh0<&8U3d%zMU?&`lR?CQc6X-aZ)JxFCc6h>MjHjUv~
ztSuMje(tKVp8eu)(bA_Pid*$%Cm@<qiRC01qU4#Shp4veQFp>z@z46Il}V|!SH$83
zV1ZOz(t>G^b-&nJNk~b^IjIfuJDO6&mQ5a7_MIoKbz@#4@w4}8;4oZv1%CpkYk6t!
zF5;2r?WbnF9t%XAS=H^%5y7wCU8tWZTD!L&89(F6NlBrT9Ump$C5L5&7G4lQ51I2Q
zs*A#D{%nS%-@Zm}rTO2vzK{K&_JhxeSx^l9rjDG(piC=P7>d?y)HMG2e%1i_?`cP^
zr%gyAzdpQT=?U{C3{nJ4S%hj~@X~D)jtt9QqK;zP_ag8x<1W=MeF?l2YPRiQ9U&s(
z^nlO1T+n3``K2^H>0ViTs`z1cQ>>Wpz6<!$(xr(&#a!cBz{pW|R(&28HaLh9rp(8s
z%jT4l0(2=TC>E@U9BrP|TUaG(Eo&)t)+?^-TWnyCZFzXjXnl&bI&OOU;8XKniv-G0
z>cr+>m*+WVe^O+1Fssj~Wd?=Zl7L(+n@uv6TzW}glHAnEf~GuyUBBnnM;vPoDQIa~
zxtbeKW{P)k(L(xt`!7wtB!_EG;}2_|y)Q?po^~t|hLy=zDd8qiA!f@XRwCTEbP_1D
z7S)Tjeer13tJRr2X&~#rJn;%NGjdd60po+AC|g31ipPulc?^z0SAyi)P;u8yD~c5Q
zw&@zrxzhCo*&!jtJ|>+y&VxR5lsH_a6`h$Cy@j;v6)q8eSOO-hPiyApw8$=_P6B*b
z93MI6ujV#xer2!L*^2Ut%NJYMU{l9)=|3f(vYg#n#hAExF2}G&#}jrkmLCiL#28bp
z4rY2XaPjjz_9chDs^J`el#cHc#2iHZUR+&gJ8g0!d~V}Xqi9i$woQytsY!I67aXRp
zi5SeF)@E}_n$vZ#UOOMnMZJ$LV1;MR_mTY|@45k>tH#!j@!)Yv+xbDhmY#{Xzl?>u
z>CJ4=LYOuSonJNbzVN9vwF|`;TgRuA(!;^uM~E1({_OdgNOyRRI~W$*lLYrMW=$#G
zo>wlQhRT8sgc$ptAgJD*D~p$vFe311`^=SW3#2e);!YpknS?FkyUxlCkMuGea8R=L
z1%;x)15GAQ?=~nmK}!zvdf*-KV-|Wk^B*7V_<tzt%l*&)f|W}`W2mCcYMaa5O=ZXB
zGuC8wv5l#JJDJ!mdocVzY`tYvT+y;MN(2cK+%34fd+^}y?(Xgc2@b*C-GggGg1fuB
zYmmly-Q=8m?>EN#myEI5yOz|fS!-3zo|Re&1q`pILu1IHT>swbDXf3tNFGJb(fdYE
ziEDWyrhsJS>C71oLQwx2Qi+T(TF8y?9O94N=|k-IY<^ZtVw;TIDAaY0_v$j;9`UOr
za&Vo!tI3iGzMNy3m)t6B;fN8E*`3PII)Q(df1A#Iz$U=1mT;?RKPrqUWKfwwZ;mLg
z-7g3?1FF(!dR1nm75xylv`G~IK_x5gYSzvkNUyeNzqS*L`XkiN061-1C!4ON<R)!(
z*}P2TCFr=IEMO>(4O$Ed*~rrocbNyHLiBMx+Vz4Nm#|`4t%(XgBlwApy<6M0-7pmD
z;DBxLZ?i(^UrbCdtA*oSpxh(-$-*_b44alH=P7jNP7DKH$T#glPjX7PBoDt9+Lgx&
zr6y)PNSlkB6{`23S~%ZSF2@~qe4P10<D<N|CTc}LaoAC6Oa}{R%D%*{g@#%;@f;|w
zG<sDi19|F)J8O1?TRkq@*22etfTDw6!EJg{sY=&RpVYa=AEe&r<1V?<RV9#DQYQIY
z!j7vcpfV;a<yS$1T2vSU4LuwQb?(uzyRtilxvV8QBZBrUYgkh9Op<j#ZTgHIL8vvv
zmEw*v7bib|X<+ffs#jA<A~Fk|)>~Z|((7zIcC>Hx{+jfCB2{aJ#IVWKeqXr59r$O|
zAwGWe^bA3d!tUPhxonlq$l<Ipx2@JHVQb=b&#fR4wo;PUbsBjlPKlr8>vfttZY%on
z`Rbj^^x0U=A3@w+LHzp+o!3mB?Nj~74r!IyOAa!s0L(e7N^NYrW9n-=n#iod?6Rk+
z{8MbZ^lcStpiSX{wP^w{a^CqwXfT3+jtqgVn!V*%M(z_LEWruExf5@yAX8O$w|x21
zo<^uFpnKLHgMyMvzOCP5R9R>s@6H?}>PH~tzo>tXWq3Y$x|sS$RfpcMZJ93)??IS*
z9^TosghtTiMoMujZu{&oA2oY0p*fD)N#>M;&y11Xc`Yba5bm-@HA)8ktX)InB?QVV
zSx%-^NwvD+lCrM(<c7d4kXnVk#cp*3)>Pb0udHbbicqfDioQ*H-R4~}*+w@+8m&+&
zQ<g+NeOOH2ySHgXqQqKT3Hgm^$Ij|gJi#_uI%0d?bl`W!M>wa^J|kjN<7Eq&!=~+v
zP>oyTmrVYn!)MZ!QPg_+_s6-+?HSxvv!WXgHpUCY?9%Cbbs7U{lUsb)>IO%%u$~FS
zo47unN1Evqx;=6yR3at<IIib$XK0l1F+ybXOYfrTgGG<A1jY}5@(D_IrnV&1ggz5y
z`;&M7s1$x8?GF;1%#}_PO8#P}q-rVUc;?=-$;N<*3oW5;JEp^!L92DEzuqvF;;Di}
z!-h~lpOm0l^zIdkMt|B7X`D)!GEzjbCc|6e-r>nfNd;#XG`WP;8Ii`YCcQ2v-b!v2
zSZV9*vgePFBb}uO%BvJnw)W<*Y2DG1f5v#bO7Q=EvB!T^`&H&FD5C4u0okmR?UFq(
z_e{>CHgL7ob#aAjkm^+?Mtuix;Tz8sV)3)#=yCdGxFD!m>1&A2I#BVYV_j`kny^$G
zlS(NNLw3lUgX%A`#ZKMgVYTQl9Sy*}dF3S5&rH&wwS?bP4cvIA9QIKEzWi0rZe8PC
zK4I&+^JDQj4U^T{(0qF@@(M0~$Csc{ht6?0Kh8mA&7xQroe5P6FHw`j=YGWEA^o^}
zb(&(oQRjyx#kM7(Z(eIwH8ZFT$1V@(9NF8cqc<n^g`q)t!8&G2i7FktFe{%qB<x#P
zemXX5dZ<WRGk)kN|3{eDVzd7~WZNO+F!Kq_TD>S$+#_ZJm}MP{Lsu^^bPZ-QXVWlg
zbubtYZi=IBiC|)<U*$4Yt4)|&xCr<F<i!GP4xfp$0!qa1&|;ceBO@A9uf>PU*QTJI
z8IYM~7m;)CkBR{3;Vz{PoKSD8{qs7y#W;<5;%RxAw=qe_gv^YV>NVPuFnXCuGv&aU
z8HLz<{6vI^%H`%+hm7Z!=6yQSlo6z};`}zoc)-t(!1d3?gOA~qYjX#9%DfsQw`S<5
z`427cRo<oZMfIX7e5zSQqbqu!-rYobZ=q7ls<&F-nP+N}7bMyr81Xh@D7g9S+tfP9
z>wq<Oay_*pFR{a{Y@8Dt5OB!+pw~!^SEprm+ou=Us`|1n%#<>-wrH1{7gWz7)JEo6
zkxs@7{9+Sarm+Vg2dLO>TqZbb2X+SrH-$h-(G^YOoK}&a(q~R;_2K5=lZ#S*Lx+V|
zUH)FCi5Kwp?ZA+%HU8Qh2%YI;4-y*9p!g*r;M%kLeySu_+g!mpT;OsUwTz?}U!<+M
zvagm3F#hq9Tlto5Dq2=YC_~lDOQ8?ZDr%VQkBwG#5<k(+crwjP)#+3(&KRaPDn3()
z0l$0AL+9sp%h{WT=|e5|wh^%~iNU9syd$<2_RfaCBPxdy9Ib>KMzbfgo?1h7<e+Wq
z#aqrMCfuAucV0+AK@Lb-xJ(@tw3ZE7-DFJF9Kylzl|PL5(6(JE8;g-Wb3&&^D-&9z
zbXkY?pWQPljC50+ssQjdW#Cl1t1v+dZYmVVGuYAT8bA;eE}Wd|(kFurfIP!?Mp#}&
zoKm5@6)CK+Jy$8LEix|fGh?HHFCvEy9yJVxuK9>Rh(VLM=m)F^7P#gR!WnL5Mbui*
zx^k!YUb^d%l;2x$a;OHL0PJ7318h6x<qwpFC|HH^9kAe$V+2V0?0@Av$w+O?@euK{
zs{SsW!-`fe$LJN`EP2~@U#ECjGrHTjLmHac6f*&Z%~>ioHA}0@@%SSWh<na-Yt6to
z?#?Th;Sdj%fG`KAmHaU5iWDT(81dUB9GjP!GO3%^(XWS|lUC!sHq@x1Bevy8iS+u@
zZN92E*5kuIk5!*s<S$8s`mQPhR&Dd-4}&I3b*A;9%r-wtjB+hDy<5Sok__|Ncb?30
zyboyp1`6#)-~<N#Cx<fHfLt;q1N+geY5Avj?>ZGIX}P5aqP~S?{3J^yM<)A1`pr{Z
za?ENDZ=R>F`Pr0SwtT<0_;bkUbH}RHj{_X1lV4g({FT*|K9l4vWO0OY$Bp#GMGAlx
z<%a1=fMN9>My5<c1>$y?*XO*3WgON~@tuZjPt!D?`JppL&$%kEOngrsqR41RX#%>_
zyY89xUGFGf`prSgtP-i3y{V%kzf#SkFhsUX;PcuxHb%CyKru;N(Neg8>LEAbsz_mB
zkmw3NVb&S}2e}_?gg5T*SDpA?gV>cXg&n&ELVQ}peScs;X-!!vYxwDYCsh^QS4w^R
z`><(awpmt{c~%ud+ND^W#0Jy9P&Hqmz8}8szb~E+qjYy6K7XLmdPhx{AsafXJ3kz+
z&*hvJ`V%UBb>h^AoF`spf3j3sh?ZQbbW(A3y^Pb^mAvhAt|v&{dd@SRYJ3atTV+Xt
zJ$Dr6i3{Kw(dob$*}=7%B@;u63;_d}Uaw5A)m6asX)-->KaxFA4>UTT^Gjm^RuUHt
zFWGXX4tIsjWvA~KK@9Qkr*7_Z3FvD|W^lj7ve_;lZv~=<RS2^gqW2&L+8CjAz@@_Z
zY8|w0!(oV(u5d!{-u34&&5yS-XmR7`9^iU<dbt()A=ob*Jv_OxF0`8kLP<$gW$Tqg
zDM?k8l0h^$avsx0^B-H^D4L~LT4~zS%stw>m8@BAoj>Ys55mVadv>x7wQ5fRr@v>q
z(9J>EJ!3N=YcZ(J9z^aZ-Y5}w+jD}?3#KB0!};k|PZng63aV}QpnuzDJ|g{H>)2fT
z<+IzK(swF?cSfs<gbJwIhd=pAv2~A+_?OHzlWbzVDNlj2Jgy+<Cw?I_UL<%pv6A@<
zj$47Z{bKrEtj6(9prI+_Kt?z7=M-c9hO47rm#kf`{`whXm5*oVsA%uKO+?q8ajkmI
z`#&GE`ZEiU(X&%Ray~m1ZrPBjB#u5rxX7S(3D+oq>yMK4#(Yd_(vhLDEHT>dlb!Tz
zw6DE_-b+-!+BaOen=UJs>82oWwlYD;piyfcu$|A$cj`?k^P#EePm~O!lC7hwoKVJs
zRRJKdbc_*9dK>OTKfuRqEqQAQw<srs%6~b8ybyL8CE14uQuNYpvMlPF{c6;Xqghip
zDi5jp47Z)<nlKcO%eE`i+9OZE4obOyo5wN?yKE`yVu?Z4YLf!Kc|QNJqcWfLme0BO
zmz3K5NKu2O&7CM*h!g@Cv@T`BaU2!yjQDj}PxfhYIlSDDs?^h245JYS{3eCGo)6*r
z>gY6AYVio6AL4Y40gOnWwQ6LwD%yms^iT)^ZPL0qwsBSnDE!dr;Iby=l6k7hX}IpT
z;ZmfJq<{fW1LZAdY>=6^2{et#ssfB`AHe(#3w?wgH2|V}FhJ+C1Ov!cKoVKsNp%I6
zmR2=;y@-7P5)v5b4r1HC_y|tHF<J)KYB-T@3C|AU7uL3s7ZQH~uj`i&f1JP_Q_(-p
zse>TI!qFw-NSJpef*;cA?~vG0!5{ivN(ylo8jf^*DH{A~no$lAw((IW=5rHQ29J~p
zC;J<mdV>O4$F#10tf5KVbwS=)sB{ZEThg%c7w|_&9(hsB;vh&MJv|>sJeUk{h;#r!
z-|jL7V~HbHYl^kM=cE4jL`=PoT0#hhOI^jPcJZwkIu@aG*9b@AnPMCF2L}*j!84Rt
zQjff^nEE#bpR1RlwAn#La8r>!X>mKadu?3>!dEMkmafchO)XbF38_2Bpa9IQWXaxv
z-rrH84=;fT7lyd0%6^QfMw3h`LISpsGx4n;#b`){s>W^~<BUF;Au<N!1K=8?-=QNO
zv0yoGd<4+TNhU;UDAO>t^H~Uz%l9Nmkl!X5;%%8ua*%4txo*3il3kMr_a%uZ=ixvj
zV37a?n+OJKETtmRkZ1SQ%9*B&(#muU{AmCGo*ZKFl%NeA86TQ?kns!~gS*lo#J*@x
z!}%AuIeV+Ui&U>8Ll)4fDx1UlI~%o#Ec7<upiGG^WfV|CK{?=YsRBuIH&LjlMlQjQ
znRHH@_|F|YK(h|+UxPC9wccSMn<a}{+JcfP`{ob+@|RO^C;44bKuw@%F`ZQ8qGyYx
zmxW6iCGLM09c-BrM!)zubv))Ao0`>i&3pRy;5<436MRNl4%-*459OnvJu}994|+SI
z?63&<c*LIm6T~-9kV!^EwLK|5=;3f>P=#w)g9u_@A^tJGhm<SDpTMq*Ja~O<{s=Ax
zc1J&x(_6*GzCjBZV(^S}#`ai{g1Kmnn9=eH9mt>8C%+jXTK_>0ixXBE(iUB1v$)Lr
zf<gQV6i85Nb;03$szyV>Q<!TDqPt}n!CFv((aTi2hpEL<t@LE>#8?Cc^koeO(3noM
z<rziID24I#Io(>`F|;vDIBHjMQ!7=<MD5n4z<_Bm$eEZkV7>>5lbAXw0R(M?ctH-L
zi=5Y@!nRKQnK7en#XmV}3sNo&mPbpgvrXH<NbG$8$Q1v2%pYH$Tq0h6$y4XLbQn={
zK6iPE?jz%n0`gI^)5&26EzJ1?<$p6G1@Bz61hyi~9vKfVE<FgVG+4zI{(;9^0vAYW
zH`{weI~&%2MH3-o4^b$tPMpu;((qb~oPbXZ)vETiY5cXUfx)AY(frl^yjA2=CQ=>$
z`S$6!ZY_~2h#ABrG(V=XLVkws;MO-+D6Ihf7fRz>hJP|Tfu{U?eYKE-izQwj2<7Y{
z5e`?=SroIIEsmZpvN|2)nqB#r2C0EQUQe0KJ38O{l2*1;FVkc^yEbaggT&y5qhrh(
zTwb8y6t~sKh(8j7wqDn#=wWG(_Km+0&4;}3_^`RgMZ9C5K7UPX<rW2ht+UotG%iRq
zpQsT$GT9y(nl<e&d$+M^65G|DwESjQKp;Nz7wzw?WJJTyZj*z>>-79`FdS5tpZjyt
zI{#XBnZNd!p4K9{-4^tl23>xtN!0T_2f0aCXn`9kO2l3~tD^efGW=%%<41F4RnsBV
z$~KKJ1u&2jxlgs^r@3WIyJ09nm=F*f;SQfglf&*B8_wNub**Lw(MO|#1~K7Cy&E6%
z2M%ivLtrC*;lIavnvtKkLnx}#qR+N^lu}czi$HQ{TomIDsODJWeH<2HVKyzQ-|Y^o
zhZGp5MiM6lkZ_#ue(X(B_MMmrI`$&^DU!=YbqRC|{J@2FUxs#=UQB-m;>gi4xk^+p
z%%H8EIOq|QGjdp`cRS+VP_{});GI9Rz4iwgnL}G(yU5zW%GR{bsK8Jk(m@0N4+lF&
zS*v%!gZDApax*rsrc`E8?eb`bjO*0eC+kM7U!tIGX}+@2OqJ?4f51?F9K*_*ZNmgg
z$H^l-p!^hIg2jxWYw|^ilt{YvK0&$dQ&-m#2PyK47qQ{RQ(^Xq677Lkh4j>9-Q7nr
z<x$0GQQVJf>6wef@}}wB_K}_BYLSP6Y}OLj9{SLqc?qvd;4~?##%2za-AgO{x&;k{
zi4GrESl@Bb1+PTN|NhoUPIe1m5p5gVxiTavMRUFdm%GI#aQr?Ty{0<8JHt$S_7i9=
zY(huF;0N{#RcBws6_%5WlEQEi)5i;*r~(94iFsUeZYzMaTEE<VrSSsQt<VHA=MP|r
z2B(%ydu8Df4Ur6OY^k}`9&{cbUr^xGMkNQr_97Nog+gbI%_%EKM=Mv(Q(1q&sL8}c
z8Y7h#9j;6+b?N%Tq_k49lcbC>9{v41(Aq@IW%*ox_h}Q)*Q(3NuC!9XuhrYERB+x;
ziYvk{rI|JfRpeyJvNy~J)<x1i=noz?<_eQQR6HY{C30>ZOcd5pGzN~d^;m`mQ|X`U
zv87rrrMhmC3f*MlpoN+H2D3~5bpv4s?n|!^zZV`Dj!DPyI4ha!S9r*5Or<}95tQ+X
z;KAAD>u~zR4jluj^9ps?&Y8xmL*-gb;bAJrh=;STH}CLqanaCd8v$?5;mhq`69H|^
zQSzV^lUI+El-<Gdb2*r^H3L<xr|izmyE$NZkO=R^L0)<Of-cXRw5Z29z{S_Be@X4d
zJF|J4dXcb0tFwPure&`3#Y4q?nRl+e)At4%_|vH7dNX9N62Z(}qd@ky+v;)M?hH`)
zN-OhtwdFx@Q@2ZHgG#apV3(+SysWGqbxZtuBFf-<**uazqMOp_<F<PHb=|^xI+Mr#
zz(Z6N2BHrOHn>7I2}Q^?Y3L#y1GL}Kx8Qe3Q!q|tH>EwitXy7b={uf})RJ$LE>VOQ
zE;)3r{*ZbCAuDgAv%aZb|M0VM|7F%E1rsI-05Whzcvpn6-URzqv{ofYHZ>4nqxu2I
zd)=DGD{K{8*Y?LGQ%?)OJ`82pt0~N3#2-HwL?&nHobYXv1qDe}d%dIyh+of54It>S
z=n8$S3%<UG$lLhVA^xeylFbdx3U`jLzzAISvE3&v5yTi;IDySTiTid5NAvS9_XjS~
zAhH%E)k@~3rGC5;9|vN17a*r*y3~;CtNnH6IO6lZyhd1modoGI`Iv)ix$81?C0H4t
z#*K6ZJyHBKdbY*2JFVPIP+JjjSyw2dOFS!st-*=O%17a{kx<sVvHHL`v!79~!?`<N
zLn$*Tqtkt66%i+~U2!7c1=&HWONV7ZvXQ;1lfnDRZJMH0%Bq6v?`ZXSXY1)bA?)9k
zs6x0sK3*S(<i1v|p^d52)4?*C@^X&)#x&4_ddun>3KSd*JTweSszXQACQ*N%o(NU4
z4#_Kxj)&1Py>!Z|Y8%glf~kJ<eCj*DeK&voK2nes*2E-kr#zrGv#=IK&E80cvQ%`W
z>uaTa)@`-a_~L{yYpnir4NE$IoA=1`?PNHeULx<`hq8v#LucicxSuTjX#f*ASDxI-
zP#q`7=|(fSWu>ldCBL>8LTr4saxvKm%<MAx8x82Q(JQqURK!5X<#o5`Wk8jkT<k8K
zKqr64HoMNJMYo>4?8-U2mYikAkojBHiy$uVTEH>H$$e>Ql|Xuo4)-!aeEL4ELv<k~
z&~Kggq<NyLeske$xCvn2jo;OnyjeO=k-emM#Kw>}8CJTpT^W;QDbw3xXxl8NXE0><
z{FsZdZaTem+<8OTX?qB$DL5TEGRtPTHlxClJOFW?^gT%u4{0%vx~h^6?6<;a*Y9|Z
zL;6#Ej&)xIFx7l-x7iSVJ(kUqV_aXq-OTyAanPiXADdKITlFnBzaA21Q`NT!a9_QI
zk(^=o#40_M1|FPgyjMe&HhXC}uH36)IpMnN;R50pVPTh}^aLGzPDp58+ZvPBT5++y
zVpHnkd8j^(qCiYpT8nfyPCAa2HQnW~!^k?U4Sl(aVQ40Zmy*($YK5)N4@k_m^Vw1M
z`YuvDmD~m;8{#%5WOU0_wQ%*^k=%9h)*}WZjpO2B2v1}zF2M$FRA6kJI`l8)aE_*9
z2eQWl45!CmPi{$CB^u@SaGHlVZ&@m6xjX~B++9Uq_Sd+B69sf>9XmrUM)Sfd%;oZr
zGeetAvN-l5ziE;>KkH~e3HPTzsn2UlKgxN}IN=vU^3<QPrcqsM&elO^^H}02FJuZg
z*69CcY{%9g(ZnKb$X9kP6;zt6kb>DH0#onqd?BAfu7t%geOR+PRI;a!28=Nhp}j(m
zhC;k>x4Km-YUjK8OQtljQ{H8YLIrKzN_z4xlgAa*_ZA-pH=1h?=HA-~WR)*B*8{{k
z+`Fv5Uy0Ad>_e7PRqgITGbA@*reLEU+wRT`#;{zkD&o^@%5DSk#jMumh8*|xIRvZK
z=DA}(fZK+C!{$w6Qi#|w#M&^F`no@Sd~*OSb?d#D&+WVkR_53f8h&FB8o0&%<QLd4
z*{{=3T1MroE;L47nR!k7*9)VjLY(-1ZtZR*cUw~z8PUA#PB$_A%LIeMovxDmjC&;e
zfLRt|-(AI%h>d{FQ3#{OsUR%!sI+2B<pcP)J|)Hl@2ypC8yWp!T4iSpEPR&H^<Lq;
zEccs6%w+OHc+zbIS}P+PJHpf<)++6J+jA_-O6t|^ECFKI(&p2cg_*IMQB1fd#0p~e
z^zlx7PH4Ul+7{pqJ?t*s_p~tJgU+@rWek*)5eF0TAIjfpfdhs3&N$sitdPIdqf{+O
zQco$)G6uL5_c?r5Y69CfOVMe#GA+z}6kK#w^0kc`)lZC<Mjn}05Lq2>UuJL_2)Ww4
zkG4NJEietT7Ec|MSTAA|tEtgN{c%QbK5yQNJvM5#VJR5H<Zfd9*&<Q($i6J)TUm9}
zXt=yc!thi^GNP-rmKK4q%imNFsN`>I8aDe7jslh!>8$vcfLBt+r=;ap$;8i_5(Um!
z;E@3mkB7PQFhuh*>&^GqMG#3VicYz)w1>^{s3cA<LOWjQl7z{bKRbq28OjoUo<}-f
zoe#Gn+n7lB3&efro(t!-M}Z%s1zjlci8k?sSl3Jq9acHoYPIOjnh<q(-F!x#ZtLAG
zm!|0Y@6KvQH5ao`#vc}VI7ADdzJ-ko@NoKAB{+R0`!b6E3A~>p;SdZ*V?*&}YwMKX
z%0S>V8KT>BZ7ifCvg%2Xm%b(go<oJseVJVL;%j`y@Vg}79ST$9<bjoCJ9@6e?jaHF
zgFfeGw*?6vy8Efj2GKl+Dzkf@q)*FI75IK|WW#0!qfV=lL$36%(BmkuyJs)wKdDMd
z&TYv*;}O5<T2y?Clc(Bg)}FPleQT}|-Za0ga!M2rOvz}&e901`e8=E7zxW|DFklh&
z)&7(PumeD~)p?nfI$nR2vFtDG!~n_^R4n~CV;N6J$DTT-{xC1@QM1+l=<!Vf8#s2r
zwUE%TSv2!)9&R^`dAz2-7*S>3?Fl=>>HH)N-!Gm1<ami7v2h8I6m$X|=Tp7Dp>-tu
zi8dN3<P%u^g4$44URJqzbaO?A{H>nyeU)JGlH(Si&;Xqi;iugoH)Fx?nMdmjXOoLk
zh^eppJRPkZUc$^2wKqyQ464nh&rdRTYtIi|DZl2h)hnVQSLa!nna?vCXLP=3HE$j+
zXUQ12jjr%gOg)P8k4<RY+tD9rJ57-S*P1P|{P$m2Nw=cS?AoiH;9YkcN#^JyvRmET
z_gR8w(Y=&c=kh$a?ej-h7y51Q_R(aU(nMq>UQP+CSN}GVA=W}qATj?`XIe>@<~-Ll
z&?>hGL(MA=<_yuN_^!K=rt9xu;8M%lRCH$Va;ie5_&NSNcz=Ej2mieH1`T<HgJo9$
zwnT(ZnB0caPn)*vq?Nk|Wnezg!hpr`EIrmK8>qQBa_YjVrD*STQgO>ssNv_W<k$__
zz|2ZZB8s(FNaa&sd;aBmdTogz`kC8hH!gZ5Fw?Q$WqsXPxeT72o<VxxcA)(fo$&m%
z)k_0mnLfQLYhibc^Z^yp%d5)qDi;50UT2kGI&EzDe)NG4Dzm+)bW#!yb~8l$LOZa-
z;}E$=gDofNVM2EtWe9fmA#$M+*CF=j>@oZJ>|vl*xxs_Q@RE%u9$)uIJ`hC;2pUhN
z#7H}h2B5=&3#h*MyYAZ1E9}7qh&GLSIftsLT%0w%p6ArQDod2#-FLS~4;!YJ>n&-`
zd&p6!ejAFMxj&1XH$Bx3U-bH}EAY1*SW&n~_{!_schxvru`#e<FOeKMuX&k03e>x<
zbzXdltID5s-oN!k5bnlhz~ekwq#N0kP&A}4?JE#z1M#`g7Qt;fsb@>OCKT7qwNnK`
z?#_H1a++v7P{c#BFOM$_`uQQTfjk`_9KK;-M2&jK&G;#T9Ot_zIB9kFp~r~QhB_8}
zPuGFII_jVZ(<0#N6SZ|wEU^0WsAEUTgE?vQa-VcgpL~r)RKY;+{difuW_zMOXL;6I
zH5T#R%X@}DRol(9R*SYnxuf*}F~;xJKvx)$%(;{v0Vm*dRQD*?=&pBb_#)fTe-aG2
zp6aO)QTtRKW9xfbG4<#aJ5uz}HN5_MyxhtDI6Y;1N_Km8dRiW|C*vn@QQo*bnA{+{
z3W^ckApymr?;;7S3C7!2a$5NLd{wmG!-Fr54C=J>RvxrK1KavEZ=Ld-vo8+xtG7Ow
zACEwlL8*b1#;I9gLr9yV2GnNX_p%sqRi?Jgere0$UiGEe|KeHTFT?6-GLFjwT|)vH
zWG+3Vdv-qpXA**{IBz_4ujb%sTt+E1V(}XFJI|7q+Lj(z2*^EY+(%eWkYU%`MkRx>
zYpWPGWGu17k)`p{_|=)@**L#otoWcOq-DbSUTDPA#^bx(>CG}UE^w<vWD1UP4SklQ
z)0H@T+_1gb87eURI{|nUwYxoDkFdT-aHMj2O<{s&t}>p<#QcmS{0q&9f)}tHwePXS
znY)CxKCVzly?kqMbns?<OP}<68$3XNSD<Kti}{+X1<Dxu(%wgYIXHTUrB@{xMiz8J
zAJ4-$1U^m8(WLz$_qjtT)af)~dAmyy3F)o%Oi?-<P1GR}S9647$MUV4RKJ7w;am*B
zO+;)oyMypg(g{nDGN?i#7M#xvPJmDI7$MAPKJO=;noTX@V)988x(j5w8g{Xspk70P
zIW*67es_~@mGY6+81=DJUf0_Z?7Br{R5qbUJW{Ri+$-EZ&^HkNsUDuuFsB$({|WyV
zZD^>VacoTy`>l&KSq||zd$M9UX)GnX184G0;EQAiuT=fFp|Zs_<dN8$W`+18gDrqk
zVAKTeO#K4upGcJ$Jyc>VvBcEeBF!7vdgo3J_zgU#AUQs^uU`fVJ!WCgCmVLL0Z5>$
zK%Ag>uV2iDAr_s(QWA%aO6VPUUMHYKx?d*0C)2S{XUrtVZ|`4sLMS5nBYyy609v}N
zRveNH%%Va7jTgw&hnmwBOcj@2;;g%(#m-BcPuDz#LV66KS!<q3BJKHQJ`>PNo~j!2
zCKpm9<YTdaUrmK}#0Qsc%o6aQRp7SAx(>da4ElL`Z~_SdaS+^V`f|6?i6`419(Ej0
zQs=D=>l2+aG0x-o2D&yHLL)D*y<8@Mp$X!>jbbZMvUbLv?XFg-@J%F4ZGUC6-|(9t
zgc$raXf3RWtoQIN3cADItZk?Kxl7PeA1G^Nt$mG_8Gc6mW&muZG&q79N43`Uf^~x+
znh#}`<PZU3FfFzZeFqcn0?B>)TGCh>yowt@YX3`m+7vz#@~t3kK)JN5lq&Pvter)W
z*qLkLOy^P14Ytm$n`D0b`p*UkjYQHyBuC|<snp81dQhUB|BcTWyn;k%|HRKC4JX?-
z{(Bl84K{=IKv_$+T^y#xWUO1r_T&5aZz~K?8B`P+_*1rok&XZYHvh|!LuX&d3-tJ2
zPBw!-kvn7`=%G-_wWoQ6K+^?bo`}>n439`vVfE6Gw^x>+4komHl{y|;k_q<#`ja+S
z1o$F*kze0OqJc=F|L5{}EvHo<bl8Yhu*A#ny8`oyyZaHdpr0U9#|$q2vt+IXVIRYC
zc>%qyC1n_3BOC@yn|z%&Oxn6zDV=nRNTL7fz8Q_Cfp#_Du4DR#{915j<~<S6NYp4;
z{Y#5MpHC>%@09OWp6_y$8r8A7;ZzSI`*+LkYhu9%?vaH5ygTR3yT1#XeXeDtpg)4d
zVw#wSgUF-|s$cK5R7)NC{yGP8NZWdK_-s7yV5I#es;2(n!hf=E`XgyL=oo(t$`OjD
z*>2us(NOj9z1kG2WkUzl**re{^O?Tu<oCf-FC%N2CU3$EeSvl8A##Z?Lj`pl^qz5&
ze-4>`s|SMZY|~)c%4Sd-Z1};e(k)U0^X?slKa;eP_0!WpdeG0Cx)(i80dD=>En=d#
zS#s>eEJ;9@h_j<lI;VU)jrXrP?F`XnOLJMA!oHeTgZ`yXQL77bt!)m7-x2pU3fkm=
z3JJH;w}jKb#`3kdGWaixcH>4gWDI1ZF>LOacE*ff!{9f)GFr@-d}0D$&NbI!f{Z3l
z6nhaq>ks<5y#E|ch=#M!r4ji>1nUQ?YYV@=I}3z+6D7Z=T55YqM@y{+CSjK_i2m!-
zI&hv4;=3;zvW!aTFmd_H{vM9m5LcT`CLUw7CVuZO`j8Tsr3-OEpCYJ-w*GM1f}aTh
zP4}KP4^UJ^rzVu5!{RnjSiv{R*>iccg3$n{h)Rq5LqAyd>|(hVoY;5@F3Oj)5QBx!
zOqEL*PI+J1f`tHPoAD>P*l65*7ED|4-4>ct#2?>*T;MLM-*6;ejC*nsgtA7V{|5&+
z$i+<&CN<-FuotIN5Bz&rd6gQSP2)H}h;x7uIuu<31t-z5wWw;%TnH>-2EPDem<?h?
z$Yi>iPWwNr4(2>!7!PSb;{CvhA8U14BLQE+L129`K(bPm4q822(fBc@3(E0tV2xWs
z7;K8Cu|UW!TKxHxPKT^32s(*RW&$D1f|5dL;KOiM5)7)nGdNrPt_5;!#Q}#@Nu$O$
zML^I00S?zCzRBFn?q7YWgAs%pH{M!)47DuM-?LMk=Sa%4vQs1B1|GkQW4!w+(6f?L
z8f#E&b?lZrIB!M{zL8^;^l{ihu#qy-ye=qQoJUxeG$UsSLnin3*r{wUi(mMN^!0y=
zO&<cboVk%9Ar7<}#9o2j`wPxX5mILMSa&rbiXfJ4)Y~S&ckL%AJ*HRXCq58`^#Yqv
z3M-d*#_)5uj1@M%A$Wrn?@4x*(pv1NI#N7YCO?UXLmVe6(Iv91?z<enn)kM@BIX^f
z7J9*kI+l1~sLPiMlx?{0X$$(Pwa4o%%DmS2$Yn+>^{$!NUIw&=#x9PuM8<xiQH!SF
z5TuXhz?}*5&0b?BSs8Uk5|2;+aO`)4F_C}@C8N3xbh#z?mJdtRb>D{SzjPAqU(}l@
zpQQq!U6y0;=;oU?ow7Wu>XY*xi_m}$9KrrC_y4FV>dL?%MxolXnGP$Y!h)SAcKrVr
zkS1bg!RT4V-0}gK@e#rw`8;GL20aSh@cYgWY#<TNn;PT;2N^SM9Imn*Yo>C^n`wjL
zR#&&^PhShC*y&$B1dfUP)0e!C5js3M5mf_AMyk_cbW=d<5%qY?CM<9LmtsJ4q+z2H
z!5aR%m4AaE>Hk8|H$A?4(|;;+{sQYbAPPWAMPw{q3f33!M?SFb-Bc><WIjCc!T@oM
z4QI@^2ebG3!tZItOzJv*RegGUQ1pNChahyL9|V2i>3!<UQB0<ggdtjfb0(4zKSY2f
z>v3%X%U*FcSn>m`S2EGnT%r#JXcknOD|nN44C0i(Q_6jUPo`^j`eXD@<<tMW(X@rs
zR{kfVWvQ~a+4!W4a}O-$Ck|TSKX-e|k^j%nlIGA#gqlv4+#iv*%=v?ZN8+yif&lkL
zZLeuse5;}36W)Iw)OSdYgmh4_QM=8n13_2SFPr_2k}?{M%N~wVUnJ0=<eY;bAfT*0
z@{IjVMi-Yn8*<44P=ov-BE?KBM+Y5MvMK~CnHtq?U|{Ydgea|{=!B4X6{yovDP_Zv
z;QjEaXOdmBUeknf*Xk}L*o}f-1mgbRdK$0KcYePV)SZ)l;>MzI5lJ!25KDgj<{gzW
zeJFwkq5qqIoQ6xeGJ^(vos9~85&}zSC+cB5TBqYnxj5w)e{zZLvy^ypWXBMjD2(*`
za{=ChIZ9R8*2C6ATycGAWt`sIt97(_41J(Kgk;o^CW}X)y8pf-2(|r*RLxkQP5kJC
zh3@;~4s)$B-oIzm%%D`VeQ5Y?-9UndWKRoqnoFip7e^do=Awt^aePszN1E3opnFB0
z$-VW1X>ZuEDdTR!vPr++-`{8=<=AXBnJzP0{yOS*#c7-f7Q~A2pma>kV=WM5i&<!d
zLP3)T+drI8ecqv8w;f`Kg~A?kCDI~EydxRX(b@Dh>G5<Bc);3crv~@}gAx7vty%<6
zR+|$-xNEb1e^@pGlIh_56c<Dm{r_kt(()CS@Mu?R$D<4zG5;_`XRgfsx~e{vV`0sd
zZgY_G{`J=Zs{s;?DCs~rlH@(!N0GS_KQFy^VxE6~I&u##Kf;pc3a0~){$G$Jp@X50
zG8_&=%=ie63x7J<xZ4Ab{A9iZ@cAkS+N2oSy#PHx!oL-B2UgBvIKgqq7$_SoP{PA|
z5xZ@YCMu(zfL=?tjhS~(rf!@!7*;BE{Avnnm6PIETr0WvE~fSFR(2hE*m5|5o1E7P
zX-SY4lfq5mF^KKjDtZ|>Y?q**pa^49AgX^A2AplhIr+r6)1NlIR-z{NC%)A)1QhvT
z{Mo2#FqNaAsjXBgq!yhubHYc}*lg)(o0jlH89z!o1Qu=0^rkbT*7a^Xv-*ZfIuW>4
z1#CL%s_K=^ks^*hXT+eN^Q(EVh7PuPZJvwk_t98U(o~K1^FwrBuUE3`&B@cFlM$nV
z*-8m%&;=;Z2|r2)Hthnw54m>^Fo;?cZB)4SW8N1w#qM0D0tpzD4B9`!U(s`*N!a}t
zhST?XX?MgTfB)DYc4Yd6{X^+HEv5vRgJWrel3>xB(MM{uVH86w)}t>{O4yq0m`U5l
z?|zK!b^V30{ptS1r}q$0L7;ZkWzVgr<JW%zAuv7me<|h24jD<^CYdB}4$Kdg8fHJX
zMSo5(E)XRUV8kJHm7CAuSE<rVV4a|AFNc@YVa>t|4;s2z(^Pf1NiMltxkvSdW?=qM
zs_etMV0QZt#0P^<40_nxb&Eyh7~s$8i-o{aq`!Y^#S{DRG}*hgcpG@5MKNCtmzS|U
zY(XJxeG;y079^QID)c^Uwse6b7=g?8PFDKgSdg{x%blcg(=`2)7U7S0$TqMnIu;aa
zZyuKJFCjBtBDK*(vCqvJ-Y%%VxFMUd#g;gB5)7Z|v1Z#cMs1%Deh<5bxG3z+rpDm~
zVgBA$@5bZ>zrZu0k;Y4a*UiJM)H|Bg@wX3vEb3q{*kG!$PNIL&0(mT4AMNq-x<9-R
zw&ts2cjTk^Q|=EVgf}`bI-Lhj0~N6~$7beaB0dwU6WxfRj-lzh&*1j2>NDbLGkT)W
z@12gmewQp<tvjHC$_&d#cmqiEuZ8leWK#UYG^+~Kxb{<4jhh&%jM5|9=e`!42q6k(
zy@q!?8#F<j)oG4e-tN@@l%hAnQ|XXjwYBM{asE|^WF&l0{92Etf6>9;So+_Hv!~!#
zv1$GRi>d_X9mGeBOi7(sN}`|>%0z^%dOXx3gLDY}0v;`gJL5xs@=TxcJc?4ZjE*k-
znu9;ys*GEV(POg6?;&8C?;T>7rOKsn8+sAalw<WFs6d{IyrN}RVEB4i1$lK>d8obH
zvJw)*`i0mQ$7p+Be2p}yBKzj+M*&kbd>bPZxNo)^z;QfVtGO=>-p@Bz6k=XpMdQ9+
z!T}{W3`S=He%=iP{aE18{!V|=Tng2&QMYPIK{JzjNbD<wJ&y10L{U%J{x{(~p}+w1
z%IZqfh*}LTWt~nb11tsfT^4R`Mw0G;mGr~DLKZt2)QL|_7An;~PMggxyl3^VeAi{5
z>^csdU7Kd$Sut(GZ3m{THZ`tJStrmmm@IAjE;PmuAAG~^5+>>Y?FQ35i`NBX_|RUs
zH|Bm=)`P`OWUBMWjOSG&QalSR)T@q12ucNgYpPl#^J7r$k&z&<Zc(+?x`)YWWj-4K
z?~IAVG?=2E#%!smD&5`OL?;JIsz+~9IJEY?GAssjPa3r|aj7K{h_@LJ`TlPJ&aXWC
zAcNRV;O>1t#m_&_j*d6rv#<YraU<Ufl;FX(639&Tv61~7)Xb^6jHvV{AVEE;ZXuLG
z1Dt%EL9!fxtJ)F=-4c{vO*Qv^WspNaL|1=KO-=RF!C4+Zw_!Dv)LbgJV0yhjkF{$c
zabj7N*njTDPoinOH$OHtqeo%+<e)u7<JswbM{r#@HP9(=gG{v3=+U_g9UPsU!6C8&
zgnyc-aG_btQg1%2bb`JlV_v|xTf&R)|6^XaWef>&gRgD7VN9K>GZ{+kGe=Ph+AH+$
zAtcTjHF&Ka2#X?G+fAAVRAj<U>k|8rg6#_oyy-`uvE4j*y`-`%L~L-}rj{qRzMR||
zjAH<9ByITn^1_&~ZDM+5a@ji6blG^VT>dA2k=5ZR^F5>For7v^;q4nui$c(}(h{JF
z8bw6;#SPR#$-bY+jDpNookjfHTy&R7%H6-5oL9#tNhd>&YtT?E>zR+=B>qMbe~cTB
zMWGXxRDhgsnUf2FVi$rD9)+30WzhZf!N!Z>sZD*}rg21l(8fv0scVUuYdMq;*w}G>
ze<tUShup2ZYCfgcQiiQ#kuu2djmd2L-&|>p@$rvCn}C|G`vc(bSX)|fh+VVYA;M<W
zjN8=Ca;i#odI<ME?(u70CjY2Xr^j+757!`XTB$A#$96denENdci41O1n^fhUrOj|E
zFyg*@*ZjBG&FD#uY%@Dq%%x*$i14u>XJXkKvwY`Oi)>t4%9D~}M+o3z9NfET!a~&m
zwToW$v!Bu`-N@ZgCAXvg>S-Pg_U0_O(~Kpc-<ElH+>lk1k|=ZjY;M-)d22kWQ`<|b
zU*7I4$bxoX;p)JA{@dy|UsiD|%E_%x7_e{Jiu^fAWkT-XDbq3et5p||jO4H2hLeSg
z37{0k>b1e0HESAFc7dj+8aK!PQ1w+-@tU}lF|uEO0beYlrL3`ZZbu{S#$Kb&HYb=t
zCxP6d^ViJtVSGS@&zRq(<6?JmTYER^r|PI}rZ7NlnW<i2;6-p}CXvj}RM@JvZs7xG
z^;o{`Oh&;#%j6TIE>gAxnQMq@F+e@{rKqg)2>R+Q1uY)C?^9mvts%$$EY&9i_ho5b
z-+?sucgePP=H?GILyQA9GXR&$bm3ZFE?(wP$<X}V+)MS`if5nsC6BB81ydz^lzx}3
zf$i#O7GzH?wfs@%X31jQ0Kkkji;wq(-sfVY-kuw(*hNm`Hb%TM_BE8HOCAWmP}y$`
z5_aA@OQjB}26(#5(UW-<<2+jTR)x9aj|?@n7=?6_u;_J~ZnBvbdozFHK4@+gc%&zf
zXS1+zC&|(-{ZJ}C+Tk+`jn+=b?jm#bvAAiJWJrWhM{2SbayI-Xk5toifG|XTUz$G@
zz;DU%6XTwo*84P|Cpr${M|fV`YB5e*Q)usCDn=%XNS%55l`NH%7s0}19NFULrBHNe
zcL8eKsQ%By+j(LP3NzQKlUx5~+E!DWo8BP+W`lbu3OM%JsFF)X537iqlr@yrcw;zG
z$_~BLrz9mO<7?t2t36goPep2QoD(tW-7WU5x@dZg#Y6@6&0@z@L5>YYo^vPD{f!3@
zBK%?rc6$yQgAUAIh<EFIC9cX8xQSkvt0Pg0%+F)oSy$_w;#nrBtX^5w<bc|58m>zs
z3%SMWAkJcA+FeU3LhR&eX_>UToZ(R&YbA+VXDf(nxQbf08)nh`^yx?;n{0aNSop++
z%r3iyv)N{6vDL@x)NY-dDp>gRiBpx=PqNPj^NRhf+@aEE55God41Lky^3_Sk`wP3O
zUI*Xhcg_5tp(yh*ZC0#i<zu;2#j-Z_vHC#Q%c&az0;lt5C$$ik*7j7|RA#%sIxc^+
zUwkf<>~N#%+;0Wu$oiahj32f#(t&t)zm3U)W&!+4#eB{5Wv5?$$;J9T*XMbCpR&SS
zO<YN$PcIx1czMX&H`73U@qXN)s;GjKmeTSGRB_ok5a?7Ict1%@sIHY#!G)VFkV1CB
zn>I~BR;{b%XFpxl#EZ-3wfE|7$3SPYda^ze7GdPVz-Ohnbc8#ivVQ8j6`K4IT&Zz4
zzZ+-EL6^=?^9W`1yf*I!&<gPrdIzq@(f=mu5Lhe{tbrfRw+k!sVI$lL$;=Pc0V$5f
zWvETfWxQyG7<G!jm4qeNMwSrHxu?L_K^Pc0Zv{!GsyB4a>|m6KvVyM&b-!#f7p?Nh
z>Vjt2=%&3f94!GHabHfku5>Yax@^3+0vNZ`3+4h>`Mo{?wCpwY)_iXQY_Y443%Kqc
ztV~C@fGg5HI)sJ}SbS@Q3+zYMI@S+Bpxj~HW$QO`=_m7!Up`b64j&C;{xSe*e(n-<
zT*GDdEi@lcal5d(jdNyxNY7YxTC-JB4WGcrNS~bZ`4M!z|Kg?qeEP-Wl~Cbjdb(}E
zr`l@1nZC-bJ3mIH*>N2c-;z1&z+hO@=4ofcUf1*^MVqI3b&(9m$*j}IE+(?+gD|h%
z<a+YehY8|rVm#hwAP4^(+qR1jO5+94ig4G9c2>$u6nCzpc|of0xGTJE^N6F?eD(O3
z_u?@hvo5e~U*<Ex##rF113(a*s<g}wPSrKsK$lypQAh=1-z~jsuBF)GA0d)`Ty7t$
zt9dv&BD}JabRlJ7+bk;Gx}6sGG(XY9q_SII-&m&RVyKjG&Q_bFDeL5-i`Q7BF*T+Q
zJ8jpl%M+Rv(uA%2a6%*%h2Y}XnUFzKjo)*h9fialFVVyU69s}T4XZq-fEHHr<V{jB
zv6MD43?&26QHxYDvtqLTep+eN7JlRvej4NRP4e+Cq%tYqW<vWd4KmBl?e$p~-9=ub
z*_nKe7^8JkcJj+54KA$Y^J}W)HHisx3F+8eR0;-FPa4c*S7pK>enbWBnskVJ^$R5m
zn*q^}I$DwFE_;oTp8~Pyu|Igz=>5;_ymT86H>+<Xp(Q_^<&o-GI~QVKn_tDmP8Pe1
zPVB{>5U1Le>NLyfQyTsf#<;t?hH~b*WmJ|l{jm?za_v98>RIfhJs)r+_9;KLb^O(K
zLEbbr?(eN@FZjzRpZVO-<U}R7GR;Bmde&+>T4+=V$!_W{c~b8#_{iufYwA2?D>JCC
zxg!v-ou!uBAYoM8=d{{?+4V?)@3q@?$iIw>vlShyqC`CmC@EuHN{LHpVL^%4$?mC0
z<)nMHHsd-eS+7EFo7<Ad=YLib1OyBYr=`|#(nzx&JVS#SdLk0VhAa#sKz;t&N_2E-
zW+9hX0m2^wPfw06!Y!L{2cPR<AvHSIqou6qRrjZ_lbovh^6<bXiRoL8orhXghY`~G
z->Z95?FgOqF^>~_FFE`z?qk{HjoCKKu(DY`I&2q3CgEIXow(O?1ayU$fEzT6%3EEz
z7<1*Rsf(rk;zn#OJ<`im7tJlwInTf-q~YrRzq-vEv2(hL?HSom!%<YHX0>rU2Ls(Q
z4Fp!5K4Sa$jYkaoHVcacL5f{4KKHlST;7(Q>4S%tFZzwhnMmoG6Ba#lq_2-^_aT)w
z=Vx?&r)w>!JlxmKa8oJE^je>@2L-%~8vZf}{30&4fjdGk1U|C{b-Q+O{{7Kd$i!y3
zReN&W-#<Aw?s%PJ%pWqE-$d>j_iJ#P1%u*hMIdfLHsi&vDFgUJBVwf+pFY#$b~W!Q
zS}y^|38>d`xp-`A)A>QweW*2Pg#;a+#ZlE%=5U01Sm6G+CYxIqFWYxrMb7l{fWLnD
zP^-MFdiYRV&1UeCA#-K%a_jx3_{zh^7H&RPfyVp}z5Q3ZzcD&)XU}>st}Ot)2ijDQ
z{Xkq)9%ea-qooVD78jp}rqV@o;y*u7_t#!qo3G?~G#YN&rXKY)Yg+lvqvGgyvPWHF
z9QH1Qsss?){5<U2icZJ-+Mmi8G`h&(dPh*9GGM`nISS@7;q`2XmCCeo;RciDFfhLd
z(Bs<~D@fk4P{FP)CeVl{0OD}3#t&`wQ&}e)EJjv_lYAqj!*|+bIca$i-6T?P38*^#
zwh|iQqq5KjiKx5%X>4oS`nS@;O?0e!L!Y*CsbbfXS$tk^a+_3ivwQOW+DNu-_?YOY
z8|$n0gXMtj8x5qQhpr}FH?cN0CUqh~&g6itpu@AtVY-IbwrLWWxuRlz{_B-YxV;d?
z5F6YE({`6cH$RP?*d(iLF8Sx}rRa4sqXMFx+NGP=ek98!BwcK`cK7j?p?pE<u~j%f
zxy#+l8Y8yD8>H;#fveT?9TTh9&Ag47?9Qsw2fXr&N3&)z{&bb~4y}awU8U|MxvTvr
z`nJl~jI?$}`ii_Vx^&dFxKXE3BmSnt1=Sdp%m@2p)F!>>Bi47TuLbjqS7dsQ=f?93
zpS`x%l4Z{hYWK@J-4D~dH59huG;eY=chvgIxgO@T^)7jXqVc@AX!Z)PZ><v(YrpW`
zwQ&odS1S2EU(xA#(P=)v{CVi5SrY&8%Y%1o#?!XjHM*5qCi!bBF7VX0=Qa0tHybD^
z!cz<iFecKotHCcTJAGyS6ZUEPLS{Rvu9J!?==|tg?Q?Bs**<)+BTc$a;pHNHF`@}d
zIt_k<`IuFkuVusXQ>m^1jbrxOM4K;K&&Dkdife1oX(f`Q;r`)y<cms<>iS|D#(2q(
zmV8O-vE*b|y<fYFa&oE*fd1V5J=_s?$IL^H=jY+Pv??i!VYD-Ddyk5aW5%4*sbr|k
zqjc1BBq7_Zt6x8HnFyjAzKcc5j_j{3*1iO8ZP`i%Ulvqd))xg*bRK9}-}kPbwMPuY
zpQ=;$0MKGFpG(d<lyA%gI%$E8PPfYoSu7%9+)dV#y`yVMWn<N&YzS4NzCrW4_m|_<
z=B?xM?VU5vb!a@)w5Gqs1p{-hie^p6W6Y}zC}!^3lGeuwuhp{k`j_s4vOqMD+$yfn
zE~~Y<Si6RkKTWom!4`sK=Gq-t=yTmtr<kfa<s$zp@FHL6iSMzv?X!pJfin@;?qUql
zPB6S6O)H{tqoz#IbsBJ~9blpGyzeO0NwWo6GaZb_l>CRd$C~)ca51a6*H%Hu&8&1+
zq;wv4G!2K2l&B<|lv;mzz);~%A%sXE+4X;~5K;tKJO;5I!YqEBs0J<y!X(b_Mq^d<
z&>AchE}a3=4w=s2LKl?C`FbYYcqW@rY(IC3BLu&96R6%4x`DK7kY`^G4VrUn^FBI^
z!U%4gS5d7#Yr6kn6-#K%UDzaQ$r=^Z<L)iM>9f*It7;7|Ku?;|Z_PI_kYN6OUi3)$
zUMy~HZ6BXj_v-1?i9zRScfV`Yk>v&Q4G-vPgeZ5Nj5WKNQY{6wXC|`9A%6SEFvJ`q
z6353^eeUt_DXm=;=~|u=d|)eQcJ<ZR@$4geyf0T=X>uy9;y~EDFmsJ=H!&sum5C{x
zncYA+-&wQF*QaS!XC{~Sut)6BS$WXF@BFeYHN8*q;|F7D$43gL$n8`k9%vjDvz$ic
z%2%6xyuRuKzWeng$}7k}UCVpV7|^A~aVxJL#qCw6->a_dbnY&U4IVd%Q#!8~hqDFZ
z)@^KKGMRX7?+w>gkx_BZVvQf`W}nw%c%WU?O_F_D5)43iy6yEVxm&G#Xg==b)|s1_
zIWR%{tQb%qHb2(RB91c-ytK_+V&ET4|2xrRJZ~lV5TrFtv9eUa-MYO~6tDT=*85H9
zw^Aa^=UjTSiut~M0o?2A4-q*~F^{O4zFPb~u@IV-s(k5z_h9pqkQc>aGp;r3ql_eS
z+c}mr>+Jc!X<(^z%g|Rtp73p_g&vEehZY^Hn}GnE4+6z8;c`B&OQcl-xwp-0DPx6R
z42xWMc-_@DJ979L!l^eyxWi3ZF9}GF^8pl`v};)0Gw+e8xanpNm7Dk&jwi%BwOdVA
zdFpuRs}olkoh;VE0Qgr8&*Cl*SAMgmb{5sSy_@`4>Vm9hewaR-tEWe`t-eXe>Sa#I
zZL?EA;=uops<#ek<B8fvX=$mTEyb-Z?oix|ySo<+?yfD`q6LZ*ptuE!I}|Sx0tAA)
z1cJMSgs=U*_ult@^Y85L?3|r_&Uwx}Gh?UukX^Ex1(jN&`r{KjKHtRM&MCV$Hx^bV
zk8W`CcYAO=G>IVa*y|<a+^1;|y(`d2bSMA0(b`;{-M+SA%$c{mkN5arK2_7!!8#*@
zu;@d>7Hk?E_wD0H3^||ieZku6Eqrj%?6WB&)n~d4t1%ml0XttC))Wcb#Ny>1dVFQG
zdiV}=q@)*)S7a1=r!St$t5aSG;h|R8=L?x9T?1@8h&LlWyzPNqlTHGQf9@kwmafZ>
zLNaR|zYL~5y1EFcg#2z!^0dCVSU7+sUUfV98j#qS0hxz;@Ku62O3@Ch_}Vq*t09?=
zb|c>Nj8{8>9MIOH!~5EFyj_JF{J|zxPClJ>*zB#JBWOOFTnLFYEJ`e1I_CD1lA1=z
zS$twYW9(lkbk?(fVGjx9yKAFPeFucFZI~ci0E3RsEp($bE;@E>#Fao5V2HYK`_TKN
zyucB&z@~%7i=u~S%ib_+NqBmUb#EcQeT+b=uGJRBKYp6Z^rlYO@NPJAgQ~cWjQiCO
z3!%ct7!O`oO-3B)0(rbx`F|?Y)fJ2Vf2?lf9VG<FUK&dfQ#IiEurG`(ybKJI3S4iP
zXop8oioz&Qwyf8;Zn$Pw!9<b1E+B+Qq>INwku@h-8O+Xo88ee(Yx3*liu_W|AO1G2
z5K)x7s&=u#+t!f5V5=ho53DZAKESj6nxmF?TuE#Yg;CD7Eq5;Oas6^Z!=puYExD8x
z?;ep@nz45b!HI@OclPPaj#iy)18y|;4Zj2v&b7Mx{GLs}55=ReoCUw3<^eB59qTjh
zT1o`Ooc+s?T<OlYi(epZ_bvp1!2vn>$EbC!`S$dYm2K>rHvbh<!>|O`#EbIlh;%Xi
z=m+^P;eo>#zkg*he<Q%Eb?_R*9>My@`EsJJ#eh|AzAbeCfahZm=}Eg;Pw`YhB1e$p
zP@y?hs5iVd^jP#dkzL-^KX4Y;I0nxE(*r)WQ}a?7GD~Q?CCYMyqApcCw%8Le`DSrM
zA!g6-44S&pbJC1Q9^{G(#n09om<37ceXqGBJK;w(dv^vSGFP6QB-9Nw`Ujqe=4;ts
zvjxN`0^*ykz|C$#?ZIvEd$Mc5X{JEH%FT6h@A3?DMo!MoCZb}0sg+2CHbe?SW}r!-
zH5Yu86Zh5vQvAMJ{h(F{*b^M#!APfqXm<IB+OKSfY&<*0zEkDx<`z3od(}j%dzFd6
zQ$s!1cvAA@a{{M$(Pt~_iIlT*qV>T`on-$~^Lo55ZDf-p2<i%Th))Qgs${suaQV@L
zjeK2|9MlF?B#3QR`58A|M;S%y-@4eQi(u4mnqG4xj47;=_EqcfJ$Na$ZYO)BOBq{T
zHaBz~EST?}f_ppG(~3ituEfvuZR)SDP*8z-B*G-Nc!+RLDovrIrdkHJCO)68rgTWH
z>%Ih-WvC6VhJ0?GT=Vtg7x+|eA3Ja|+@t^8-HdM=c5^cAeoF}b7y{&)LHiAN-iDV}
z$E~$UXzf=lFUOr4QV<3k1C0!~s96b8yM7|IdKVw+oP<ULgO`?@hLuw`+Bh3MRNWGb
z&nnIifqWf}09^XhL~+kP<ZLTB-!I}@LhH%KS*I*8;x1|JifHmk1Tu`Dt{bq(djzO%
z?dq*+8$7Da5Rzs)>7v@BY)WS-XUV#xW+q3M{&RDaPF=jo5KA52{TM{{E8NfP;<(4x
z^8;s7%ih%;31sNRKnBp3Z+3L8ht%0KUk>z6)|0Lw<bE0?=#$)i*o#`_Qd_hYXxb`7
zI|Nu<`1|do=|0KBy+kEG8iEK<tX{g_Ju}5;M{RdS=9%Q7vgs0_rBipg<hhI5o0BFx
z6^*gYE28#RS_-_KR@5N`faRiT0l0hsu5x!RLe$qXN}}uIR!AOa0($G`e5Pb4TXp%c
zFQAim^1Rtb#Vfdhi!pGPJ@Z|s6T|T(*)IRtPwy3g+p0q}H3tfm0lQ-Wc5)?<h<z=!
z2n}}9E;IuYBy|Oxu?upcZk;!#x<0f9hHqKd9Nx>Ijwa>w_21EMHqWBH-<K}dy&_Br
zIlpOx<QjUp0;nD5yyt<JQ^_=Kh!@}Xbpr;WHTR)cN0_29A09AovYo(3w!ZayHc$mx
zx<{|2h@L+tj(Mu_h-h>*VS7twT4aB(ybQi)I4Rv9M*aM~YSzVTC8_6o?>e1F7_tvs
znvNfa)8?10wH2jCuFuC(06ZecBF7y(^*r5~zAJ`)jNK_7TR-61c~&LeJ}H0BbR1Ns
zV}T_V%SqC*2RpB|wm+%1h;8pT^1uT<?}G$$^&4e6+ev8}&c=_Aae$wgBR{QU3#=+e
z?vI>drnec&dsX<c5&|@>3=8Nv*&d9aEZQf3mv>#gM|WoHCz@N%y7OCnJL+Wazf=A+
zv2iGi!iM~t<hPnIfk!@d8D*b0*5HRJe|`Cg)InU(Z##-8?JCYo^7kk{mMD8ls-mtD
zxhdV+9rj#W{j)CsCgu{e{k`#8pw1sYylc(SxnS<?(_zg9A9z-^*M?H*t6u9fznoud
z`2&3P5T|kGU5@5zB9n2yFRI1@&@18hKh1v!$6_AQ%9PH&k}8E(S9su=R|nILy?zrx
zueL(7;MQcczP(AbzwCVsT5wg<SIwsRgpNL>y2}HEBZ|qJKKB*kt_UMoX9mc@H(ck3
z0asH*=Q|_39ZiP2{H}s0vUztj%B<qKMt)xv0TSQOIx1uLpU0cVqmmVef;{RHcoW-)
zD0Mv^@2C~KiaX=tGy_)ztk$JGZ{$7IgZzfoY?Nhfv<q{07~HPU+)0f11<V=*xQ()9
z6@1K8vIWH^$peb3I(;j27y#!Tg1qoqFx#WORIzO+;XvA{88(SaZei6dcgIZ}hgPbS
zvT|(wcqJ&p_eh6zpD2QTV=lnMlH0j)wv-;hzCM=ukdeDY2c42Vx(YtxJCwMttcC7J
zJ97}*F28HyZ6ZhvM2F`ua?Ze>i&?#wH3QPuVtwk!h<IkEYM`mFZ&GNtecIP2)5KG7
z)@!QT;=EYnR!V|V{pH65hJY}8v!vei+{OWE;v>u~0dsA>$ja-qd0(<V!snKKK|{|1
zW6JGf2&NVsRv|B;NB0?Oo}z}fTnENXmAt~hm5JJ5ZV2Z^pX|nw=l*!+x&3G}T+eYK
zu@ssk+u<<`F&uA)x|0v+5!<57tO8eG8E*+Wns5=FB6xueoGm%E;<H_&%x&p^h))rS
ziUeIBs4_mhVpg+Rd1%*q<AJ*QSn7|iU1nspt3aW((jGpe_`3+U?6Y6^5(FY9fy*w(
zh&;%wCdJRyEw}criDL5(--8HLz<SFw*_chM$M5gUZ+@{yV<GY89BMl5x(fo@UpXcH
z{@YfuJ)BO>*cNzurBl%-H&83v5nzwlV__`+cb%1FB#M-HXjFWI{w_JCm|nkCX4zCJ
zKP1kf2(~lK+cd$wF#{C2x(<H{DE|r~v^N}(4~5P9;HABNFLWl%vN7Wbq<bebZ%F~U
zx}v6_$+DyQ+aOY6Twy5Wb3O@1lo&sdpONkNg|#c01Lh~L%QdOblL~RSt?BTG>-)VC
zVqm?B4rnB<Y6&i2R^6~~1HVs|a|j!hOEry(Yfe*tW2~m;9qmJ%`!J6d%Dd-=r-<=L
zEY$B-T2b&g!H;_hR-b6+0^u;#VX(A~;v_Kd5P9O~-lenc^~#G}*WG=7@NmdBBzXN7
ztm%yduyOvES&i3<p8f-RN1CFZTv(<o>mcr85mq`YLg&l0cpaD+<kzHBeVoB%HI7dL
zF{HHwYAF{ZG;%AA<d@ljAboqF(XmD`u59Hsu|OkhtDI&z@74O$l<$zZPKEJCe+nSp
zRItF8-)c2#hnP$k$`V`&wZ}6Rw-Si&({H3dzV}`n+Dj%c#ig6BxXTG!6~8;A%Cw3a
zQd??1)iR298Q6Z>N|8H%sg%~pXbV(2;;QKtSQTwDukg-7fZy-)y-l^MfR@srcBj|d
z6530e%HLi`3xt{ko6X=WPY>ib+S0K<Y|nrpz_jURNPY3AS(?NkeXWCx5u;B~<tWgm
zanoU^YoMfLH1@}{uqDn}ms>3a-M6jwPFE^cPEC=${!V7b{n5sDalK3L%^vP*9uB!_
zJPZ?TcJf2z=y8Y3D187UR-1@u;dwt=XBr<T3*wsbp(4$9M?NqhGmL*Fod(gv&&N3}
zFsdt?ymLPukhjqi?_1c|)U>mhCxVotF%udLI<CDr*wbaC5Gg@+i@3H(Oo)DS*bIHR
zclCs~UPju#S|Vc;Rs}&Tbt~SN7sQG_r7Yom_qtdP=G-LNYe!vw_{OkWP;1gD*X;MF
zyS7(~o+QJnFCv`^c*=BRXz9VzKfX{3Nb#m7*z8OC2h=9Cjm8uIvG8G&ZqK5V;dIoY
zwzoT6e<_GQc#JqtIsi)K-Ng+NceGt7cV!PfUzyc&ZHxJyd$ifdK(6bYQoV04NNyUO
zl9G;-dD@@r9&)r&l3n8vSyye8WmdT#Feq9Jp{s3YoOEo+iW#f$LTn^_r&3ZVJwH3m
zAzwdQCT<ok+ITT<W?J<6j6p8oDfjb9gmDp=Uf-I2c}AHg?T^qqkspc($MUe+p#-9D
zVi{~wMs%0(nm<624Di)D;Z#+HuD6~y<#m&ihMqOdh~<mw^?JQx{gVG{8i-&hUY@em
ztAaG%o$Gw8w$u5FWx*=i-KU>_Zl_nVM#oe#JPVNE<sJ;V-;}1^#U*~;J70YWp}53~
z)H7Z#_(3YSMy^Ok5or82UGA$9^RnBHQj>*_Q?z+pUH)}j=posbkesT@3fIt*X{-Y(
z6uq>?MjB{0k9++rJIpM&kCmpBdhmGjkB4PE{ch?XmX#HBkz8qj*9t44G$8Jv*sRn3
zk)9Y%9YpU?l3p+}p$cOytJOkfVsZ*`uIt?_6M5sd9HqDZM3Pf=5245tsdWYLnWQea
zeF7rz3F{Uay_m)ZGCsFSb|T`}+QrvNG3qrFm2iXRj`*yqe@!%|b`pCxS;neqMf;;C
zf7Q(zC|8<AE)?1?-+`J*BEoq4*9`06aBvp+mS=%V-LcX@=Zhri3McUnmoPUD`Cooh
z{V!zaAK0A87%mTmX+^DLzAJuRS2Em7xJy`X`!eTpNZu%DOi|0Y5zc<B1Pbb^X$>WP
z$T(%}rLB*l6KeEmP1xBVPJD6sf^aUtPbl;~CqAov!~zhlZExCD+J1btmsrS;fg$-f
zY9|(mf0f6Qo>TpuE#c*>y3AQb(b9|dv`&ZAIz201lu4W0JX@D(P2I-Hu1#_SJGi9W
zG|KOXblz(Op}kR@+w9INf({SbLSV_u>wptsf3&7i$-Bnz^n%dlHl3^o5>H})w(%Go
zh=`AM|6p;u-2nJttZ6~3<?ej65}R#mVse~^qRj3gqr#4-(ZqPtiGdOE*Ye5toe<rb
z!)`5ic+1I8|6C27@8|BR%M?mWVh_SXQckzGMW!U{t@I!lgPfJJs-zAjS2rJ*l0~nx
zrDLZi5wJ$-M7}-GMd4wz8nM2dxBx3#9Q?jtngfJHDeM5~=Z@qU8r3cjC$Dkk`fy+P
z%G?bkCKt;!3t>DDc!a^`l0cAGSs|jQ*ZEd!%<4mw!ik+ZRvYWGf|I1HqjH}yAS2I&
z#3@8yd=z*ry|Bzrm4ze6GR;RST)8EeTJefv`?&;RH-CL+hPIo!+LzCLsyJRG=cGqo
zSSf0}EZ({)A1dp(&BKX!l2KWA-k_=v8gG0?i;x-=82?<79Pq8UK##tc;(IY26Msjq
z-4mvUg4!_x$~}i~-wc83yZam&JMdS3Y}y)Cot$*bvB}9HN=8|2V_nVVb51K&A2o|a
z7Qv~zUPk3onTF8s)f0x@+jP%Kjo3wgIHUG_H`R#kpMx7P0uJfRwQRtEERytn2fO6v
zobL-qt+lO3^`b&5mSs+M3i=I^<4G$k(4a_h6%whqcy_S8uJ_KVnruKIL*SX%oL5xO
zL<Lljzj4~(kP&8Hk+O-ks*vc<p7kR}0{e{tc43T2;g-R;i*WqurxBwzR;oaX|Hl2K
zon*vKvP?ir(cK`o<zN3hrsVq80Z<qz)eT^X!DLJPvq1M1mHs-|$D(IQ-x`$^6S5dl
zD}ZbY{pcbmQ&v!Z);az1#k+b-tMmYoX_naIhtOfP&e$z-^1NMD5?(b&I{G7OunzOq
z2F7eFE1x}7Ngs9L)E5hV-#h0+e?hE@hYw=m45DrS8O+OZVVeDhe9f;+@eNW$&4v{N
zhZnT+S?P0+jHf-WhsgAmh=hA_X?vnlr3?@QBg_ngG`?5tW->Rd18Y8Msv>*vYxG*<
z>6Prp%W#lno)OFhYH{~HpF5NSJ~QC|$w_7mc|;E}{b6ELAv;XyLM+s+@|);R)l&aJ
zD>HDr;L`e0BUY!45>V5*d3xF}&G%3D4p+;#$8!Stj*^)Uvrn_>ja+uiKR)W=b5Ph#
z+UvR2G?2!F$UCJ9bsfB3HtJd2Endyv)o6hnk+6%|3)jU};#55VSIfkF*+)`d%B+Z^
zf|sW;cc^>c?z{&VFLqH(38(*t_`wq8yK%?wRBRPF6y$=_1tuU#IBB3qUiKTQ0S|vZ
zs@yN$_y&$N?FgOu8E+x5(WCTqg0)SnU?J>Pfs5i_JH9c-$ebAVmtB!Dez+rrLPR7s
zEDX^Sw9}lDcGylV2#QJr7l-gz{JbIWMr$)z75$F>Su)JhIeOdI`*kGZ`Oo$bEm)RH
zWt`R3oY2V)RHqD#R`H|X?vJoBozc((M#hErX5~6{$e5zR-#OIO!hMT+O&`AL$WF`N
zO}1}WHr(lftkxV7bY_IL2`c#>DjQ_$OS;KTS8e4zy(c%}SXfE1u`7|JI!475DIyDL
zoilS`xc|Z3XVt1^`Lgf&Q*;uVW<1j)J8;IX?8?$?DH-4Dzy^WDnC&;-#d&&-`(K7G
zk0L%Inswi?hi;mrUK%d7J!})LW^jd+che)Vu}PkhWE;P_rceDWO(fHPt+V5Eo|4QC
z+I#(C(Uc3)cNtL*cGEL-*IN`d$;(N}`SdcE2$#s!)M>Zxku>0~mwrLYY=}>Z;3%c2
zjXui8N2;pWF2~wDprZdxVr$@l5$5F%qL+j-6EJJ^J<XjtU}^K`PI?pKyynflt4t0t
z0#}YjPS=Qb=ITr3!jW2)chSm3dZ1)A2Zs|@{HV2?ffq;0xD9uUig5JsS`%vVD!s2E
zXswR41maBN$A%TYBYfDmeADwq5?X+(NAcvf;Xua?{b-5k-(U$<d+hQX>pv*^-2;Br
zW1a(9@h!Gg;C)`t8u4cl(9{h5#@lVHx2S3tpjp$ts*~HeS6_l&_Dnr=dGg@S%&PMJ
zhh50?^rPC}>F7wA*-{-NspWc}vNIw)*C9J+`-S<bTl}ud%32i2BpollC4LaEhT0|8
zFT2uQ+|)r|bUS@MS^ae|JIQm=R=IC+1ciB@t$=w~3L%}rF#AZ|sR<^{MR}uNpPilT
z8)sGs3=UbeX)WMi9NQWC)9e?}y8Cy2n5nX*nc-Rq%x;Z?wXokaqEj4he9xLGiI<Bd
znw&p>v3+~-VH<by-W=c7Mg4g$`F0*xK-%(GU*bo9@d?8Y5S9n{EK*t_t4KO#p=G^O
zv0thNZod!t1`=3MP=Du*Y}{3njZZO9Ywm2}cF&2v&3miPk@V*-EoP!BrL|4s2jxxM
z;f|Yni<iKlp^lB+?SYqJ;if*w+uH`q{9Jn6gddu$=PI@c9rn~sFx)#_jHRgYZ!tDt
z4D=on@+7n@mILfU{aL^RkyhABsx$^A#x**;j<_cNRZZV#f*8|=EGJA%e;xPzDVe_I
z$nm=#$Ij-bMy;jNo5%I~jQ`J<Ber6v5f~hZ4003qRUlF^VpMdT%naoq!%6G@Ofm-0
z?#|VmbXRXILCIU%Z;8sui2bVQG}-S^-pIO>dwR;1cb!=vW3{R^qhQm-{r(gwaA!@T
z>%v;jJ|hhNJ-6MxqbXNPI5)W8aH;m%A<K&{V=v?T%RhVfehIa|yN2`<vh&-MagL(k
zT1yiusQtlw5ng=TtT0`@_P}3pKG0ueya&A)z*hfz)AW(?AFyp#@yq#eX3l)6{|QOV
zxXq+I>lYlyyPosSY)j}j?vKV?_czpj6IIShAs`_)Nilawc))?__mb)mfq?y<0pVE$
zqhLJg#j`7RBF*U_)7{hb=PVJ{15wdKQPIj24n8*NRN7vb9$7A|^LAc(IJ0VlzJFSc
z81xD8p@)tO_owd|)l-e~W^BbpHJv&^&SeghCg(F$Y#XSX(DO3lfohJH%a;=n6H|Nq
zlB(c9B%(etRPA|Y8Yw@a_Y}=ixESNJz$ec$RpQMGb@WmdA!*DLfyDp(#b@z*jkYOO
zQ<BA@yfcXUR);7j$I?Bmk}nN}q4>~c0x_)%+%D7PC;R$mhH@BD!!Wbmn-j_;tv7TR
zdk``%n;TMTdsqBsyDM{bIy^t*69);eey(1OwhI0`=q=iHc`eCwa!$d(3R9-3fQ7!B
z-xN7J_`$Av-RxeoK@vY6j>8xeJk;La6gt#6KCx86^5f4LFYT|5<8J%G;M&K+?^q1M
z?W?6!Q*qTaTIP~#RN6GRr~Ql)IXBjE$AXXKd0$>Nei;?MZ!+XGlVGUY6j`5Uf7*^$
zv%JdnT545@2ev9Mr{ZW_7WZ%z#H}2C{_5idHin=!>A{CT>eJ<YC~zrdoW`S+GD`Ke
z%p4~2a`%PKSK<3MhmA3EzTO%7oLYuI^isPtbje8wTHY>k#th{TK8qfzN-9NFDnU25
zz4w0I42pxuJ@s-vA!mEq>a4<1RE*j#;>HWY4SK>j#O`D8QQEeJcI!(KJli0nY2mQ6
zOac0ZaAyaqUPAwPZ`wlkUHXWzI77GEEgijq+v9}bn+sLf#4y_V-}hCM@Xm(&0j{WX
zPQ#-lW!KwbN5}{bJ-Dv}z45xY*Pz5SiX#rshJthS6#60<9>aoxvD;G~Ti<iUr~~M{
z9C0g_tLrcT(pgWGkDdnre<qewc6XW>)*jv5!qs+6lNaCb%w>F%h<?1tsoie#PL09M
zj7+Vv0L80qRoGi@;}9w^`_{Sa?iw8@4FbY3F#Uz#S@@GH9WIY$zRWHw!_k66qk_CI
zm%_g^S?=7p9_8GFfhjPVR-x>VuKO;)UA{is);mCcRaRrv=ng>NL@*g8;xpJ~=!HtT
zdR0?UF4%sc<)0GVPJMOo0}$C4B58c+Lk1*frLI1+d{u#g5w)avJJCM5-yrkYcvdya
zHoh3T8e_|M%Ac1w!SLiP1nlc!FU&lh{nFf#JS%=syx2ppXwY_VXb4K3lgZn6|CkOc
z@(HM7zsof>7fUpm0to0_X2TbqH5>M0HYXc_-+E@?NWEdv)r_svTpd-hSmQr(?5)l$
zP+HNNao}=BRUe58{LDTaqYIiUO82<x)-I&MYaGeQW}*h%seUb)3C%EHa?jj1#_b0O
zF)Kg|sEUm{KGU`Z4r=e?cXl2gMAd%z))n^lz@f9xEcdm7DUXA2%kl7ou92_`0hajr
zLVPScO5AOu)NQzHJ=jravTW*=a!ry)ByseMf5VjP7o0^cSLE^p+jv*e)JdPH$)IGo
zDyd@VWNgy@qrqGQYi8*pu$HGI?U`rQ1Cx4o7R@y~(a+YlJzOgC&xn`VjXYzSt-=zD
zpK8g1#f|sp4WIG+S^j`c<Gu@SjFivXq6g1a5u6Jr1)fb?ku{}tNj>C4j<ZNzS_@>l
z-Ss9!x_{us1~@k!)w6W=7}7hU(cq(+2vOIluCnSz`!}wizgh5<%7P_;p`ZGayd6nG
zw-8Ew)i7)ma|EKaz-gZo_BEYg%HZ~NyiRB|9Q=+O$bAsrpB2C8vd|N|2~BNOrA25y
z`>&^TkP*|pe%*b^`kc(u`5gP78@G~8)roJsbxX@<r}&jQSi(y)HnsIvq?v2uc!YH1
zN2sOp_=aEUJg|dxQ-{W0Ad_P*Rg|<IYiFBkoLN~=d2ej)95oAmMo`U{HC3u0Nb)#N
zOc7?6j<H`_>NZ4oQ*mPkL$CwE@nkc%ADokg8f;&tsVW=UGAByP#|4B-mO=#@cuUju
z&}Pv*nUe{UfY+AT0ofI-dg$NcXKM!MwChgl5K^ry5uCKX>dU%ns4;X~^~|HnL+U8!
z&cIAp&fXMVeL>@@XcTE>2)Xr7tbN*97vod!7UaF4(89gI1aa!e!~7TN!e4~HW%0QW
zwGw$)cxxP54Lb@wEE%gPVt$mCr(d$xWvHEA`6F2mGjX56asil3n)MJ^d&Hterb@_T
zNlUw$fK%h_RBb<V+H&vq1tbEgJFAd0wyCHuUNDe2$2EY3?K40$99k)ffxEtxYN*ae
zY4G({Jyf&g0ncnK??b4GUwZ<riRoU%gsDB7aQLY0vq0pSW7QpeFPYrwywHV4K={E)
z&pcRG!Y9YS642%$3}F~t^<6i})4<0obp}kZW*c*+D_vR7`cfHO@uzC6xzFBw7IdYK
zR%>wZ?$q%SZuju*9K2%>b;-|G8u<3BNvOCTJ1wg6`EmnB8%$ajtO^Tqp`qQHyu`i4
zl^|kvx()PSCOR6YyDKOdS-y0BbRJi>mR%TpFT^W#ap{da#F3ggB*I6_VrTX>og{C3
zg2HG9QJ`d>I@0AHv!fBRsnK40Fu1iMfox5@>bMyi$lDx9=?k)SW82Ryuk6W+QGWT_
z(WUQzP&hXDsO7NpQEq7KH;3CQ_R-YqHU+5~S#8uknGAH=Z{2Uk22NO;0*@N>uv4)a
z9gwzxA9StBV{xin;F6qn_^Aj_?r)t<-Pff=SBwh*MT`YG>6zjrc~glNq1NCCf#d2^
zfhXrnU%OhXHZBsDOj)GnIx=~}ECqQT{o|SfcB6$}za|S>jhlF};^HOc6+d;*aBn`g
z>HLDq>=QhXw24WQa&LV{?gGSqR$+<s7s||)!K%ujB!>5GW{21tRU)P!X7i?RU8xOc
z=?VTVOGSxV%uzm@X7|p^8-sXJ&M#=(LsSMwPgkZn*<ugpQr#EJ-OgvDd3b7AQh&-S
ze=R0r0uS}|4fVMj84d(KeTl4CO3`nf<D4cZY=0@ZlPC<DoU5(3O*3Vin)fRmdOeEp
zES^s{1wo-ErqA7rMWT945<inLFy(bJ_1H=T)UbG)(urvbBlXx4Zf2diEt{`U?hHfG
z62BCDPsbvM5)|)fiOy@30ZZJ2ou>TfC4*wdrYBq1Lah-?-q5&bQk3H*bK>g|PV7>j
z{?9d6_wul-z3S`|qmt>_oe)0;u<(GXX9{>z*Rt=Z^G@}eQSS6WiF89EP-_X5x#xS}
zOp}T|ux`#7Psl?hrxuIvdR-|=^vTXt<w8A!rrYbGn0=Kx_~E!YbiAnk7vRc{(|$O0
zSFz<b`8)FR`ibjJNkf}+<+(BHN@^doGvbF0swV$h<ZhZ-o&A!>bxgUXd+9E^{+Iu7
zN3+Wnw`w6(cH^Fj>f)ofN7aU*nPjk2l;VqTwXx}yF#^XSyAZrtQ>_y5cHUoyOh2z?
zb++ZZ`CS<rNo1d@Yo#48Ff%gqnDMGhU!;uii|Z=Q^yxyl4_oev7pZaK3x}<Ga4KyK
z42gC{8A+WlFSIw(^|yo17rue1LNi}8Rt0h~54-y)YFm<U&cY~#?huK^`U|SZBxmWR
zwwzb)LEo8J1exjgelFBGAJg1ay;wZYmp%snMOpmA*?_X@>Po}`ZXH&*Q6WcVs*|c}
z;Jrae$%^7xa@{>3VF&Ohr*WCD>^v6xp^fvrVTVQAzFXLP&rP*{Va<BufVCBp!{qF&
z&*i^OVV_#(1Kl^Ghlw^~VK&CGv_}BrOQZDmuZ2N-_HwkFlY88c(rDw*{jKMT$0Fx{
zR-lOoVz+BYDV6v|&g?El0Uuu9_(kC*CD>SeJvpPWa$d@>aNeD365{5zAB~wysActj
zW!pQ=9Lk2<eW#-i%9(w)(YSUS4JI+TilOF6i$Z+XE7TXJfZ$R2o(wDEz2YPS45=@9
zbf#~X)pmpIXikMuJySR{A+CRMn270RWvZ@=P(8JAa3X%((8Hf5Q*_1$<hH^_2+Di)
z9<crzkwoErIDXAZdE8e6lrd6uUV|0@gonR$aSacVGdURPI@6+#&CgCn8BZC#Uokli
zOvMxK9m1<8ja@YB@6#9MI~Y`7nl4^U^U73e&m-0=;G`}0tG+cpU;RLj3upqY&O{5X
zfaVg&p`pLm>fv<%@P9C3^U4taa8wvN=16m@6%>}<QGelWP8oznb#(Fgw}xngdEi}C
z_o0|S9~slO_k*~myxoPJ`0VP`L($JR!=%!Q3u6}5(Gu*;?5~v)c@%$2&h4s~#|zrp
zwm4^Oeg?UGG~U5#Ex1fx?w6l}YSr&bNoDa0BmcD5pH4HwN~6_cGyX!4Ry;Gt8j{iQ
zSn$U-0M!gz#~oBs-9I{T>dk8#gWhAwsZ~)h57u`UIzFqPMDXItTE?IQ0c<lWZ+S}F
zg1?{&B{E6unn{>Ac65uBb>6E_Zjsh@XWKNmsVtT3;Z~}R+~p{dQv3RTzvNY1t-E$x
z`Ns0pc*|c~%b-poy<)7_Qn8hySyvtV<kg8?+E}c^O#elu<eT{sJa0A6I%2Pg344{?
zl?(RCZMeRPCJ%$`vjF2hqtl|B;%K}Fhj>M_r!k$bvNsWZwt9LRVopFkF`Gnc`H)st
zzrt8?fBWacAv<u;k83+J6n%AZy0pA0T{8GBrMB`?W?>!@W4VGWFr*VVM}j^UK)M#&
zNOb$G?Zd>0nAy*Z-utxrW4`>_F}&G(>H~q^MS?GDLpWG(z!#v>XN~V1vd@$u)_SIz
zyM{d*^U(M2L0<}Pn));C0g1jjfs3|~waH6iDMyho-e%rOK{FHbn&6L0^`V44x{##7
zO(P#(UCq<h54eo)L=~;$H8D46G6JsF$o4O&*Z1{gZsrqwROg4zIQ9(0Cx52&MeYo4
zb+qKStY%k%dOCB>dFXqZgjo`gN5yrj>8yHP&EH6@n25Hy?LeVBok4eRH_ZMet7#+6
z(r1}QDdu#d5IboIvx|5qCVQ8?D$Z>^xP^1+X1QoFx$(&XXVJL9G-HfWPF|p9A``jM
z=&RENOGzs$^YdwL72~*%TVyov^VRIOi|5sC8WsiZYS?3n#!a_~{PFKI0G)@rN>`;4
z0m!9i1#Ft;(_6LQ7{Q>g$m9`7U%c?-^4an_J~(kof7Ay+^a2%pj9Y@Tb@=JW#D~&w
zSY$(=U-VjQzx4NC837Rm3%-Nyy#UI2P1VhIwfp_)kT_(%_qaTGkp|~o)*j>z1%C)$
z=ghTZH`NYV*f`voA683KR9odjrheNAoM;<Mfv=QPPdQ3xyzvP<!!*{ObrrsvtW@sa
zs=cLUMNsAE?;J>6qi(?k`y!Nrh(t={5Wc<>1l-apd-H2~y6rA2E!X)OyM(!up?Kd?
z%NGg(ZM*wrZUm5u&>1!7^<lq?83*L5Q}$dx<ZP*Q)t07SHDJ4x#v>2qG-)|%{6Y#I
zP(<x`z=rp8@>|7PwmFR$w5k0ZwG?d4J~+;L?f|!&pUfH$YTMi%-SHt9VyrPdt=`tO
z-w+U5IvykPjK8hP$dM@qjUW$ZHevf}@Ndzpw2+k+(a(}uS^MQ1%bv@kh*AcT%Yjgv
zT)xop8j&UeB-+~-^x;}@kFb*sG5sZo5(Dd1mH_7O@1mN&xvHL3a60y835-{6KKc1&
zz4^v~;uoU9d$y^e(D|DDakyB`_qFVlKDeqAMe^Z$Gbtm112(5TSm+pK(&cn<<JmCH
z`3NQBjR*2nd-}XmB`<K08<{h5^}(WZ!xGqc5kHfxu2o6NQ|ky&DZMPRu^&2hX546$
z9qsP&(~u5kH4gmM@AKttf1jv0!^e-N`U}A=y&)X!c<@={^VsKWWYH_+t^k*%!=#W?
zFldTKJt(fw9MZqa!(1!faR>KYU9qP*eSH<vd~EhGGL^X6wm-zD`ATfG^%m0ccptDb
zU3*ToIZ&)9o<^lJ>$dZKJ3!~M*_cH1Vi|~cT51TEk9GiolIQO^Y5>zt3l0f>iz6CP
zt-0;UiAtc=Vgl?09%~GGRkJ7D3olL3ez<vR@YvMikM=iM8ev<Mgwv}UMJw<DF*eT0
zoql)V*13e)s)j>n$=dY+uli_pdG&2(j2~#<Zc|_nCg2Bs505_`Dv8S{`Ab=A5k|~2
ze4s?ZyT$BeGjf_HwlQDHUk?@TNv>m=XA4h45w~S=2gwdP^JRE}BAZtiDa}q9>lS$@
zZEiFyEaU26uzPZS07}RhpkwilI#^A%%xu=KApI!e^|i~-0KP{=rpo3>{`w_uU2QY}
zq(;LCkvwNI^-dy1=nvSIAPGoPqygqF;J7p}u#Nr!pI1=}7#@b+KKm8w6W+W5ekQJV
zO4vkmEPgB3?W0LO$}#4-Z*f(eeyv<?e*B%{{-EnDHb=Tsh*RQvHg`*}^w%Xj=qRYJ
zLBd?91d)f>o20xP$a4v;G}YItm<qRnb;gX=i%QUJZ7ZvO2cCA;cf5qMIN4lpJ~L)D
z_1=w0VVpVVxzYet`KfR^jfcMgXFvMHl$+dlQj&CUu=SuIc=?i~naTarkNOvtN}l2s
z8il=b=XiSNS{`+3j$P^_pFG2YBiF^7p4|<a3Y<erQ0it)OE=%J3#bir^F~ueOw)3D
z9d1R<&j;a0<L?K>OGcPMsf44XW*TF##&;x;ZIFa^faIhz?sfjNfKo!c=~T7BW_~+`
zmTaqWosZ)eIS5&`>3P+o!z1pc5Opzsan=%v+ruR9?ws5ABN9E(Ws-Iy#?r%BwEIxs
zBOO*USydh=<c^EQ6<rWN0#Xhs4HM!|&-$flxC-sv4e8WChMt4dzqM?ZU=P>8Z~#LZ
z-5cBT9@T>2iFh$NHk|9N6S4@q89E__x!_1Uo+ctw?yE*|B=68p0He7spS>a%_?^Y~
z#O6z5{}5;I=159Lyq^xCT2t6@#2a(jsc9{Y=k876|AyJKuveQpJ^|iVCz!QgdR)fw
zd1Hw8zKlk8^S#}mI%M(iMEg(H?(>!G79vYjo10#%biEVbJDbPe$<TE@WHGE^LL3y+
z7=_0Wj1rLquWBc?75J(lyT#($C^q@u`S-$y2@$MHAq&2@L&CwKsOtGaXh4>(nr>mt
zP1eTiyTjN`AJ>)~EB39nMa!z@MqHc5z8&{l=2x@kY(kl49v+*M_%JpT_{K(<w4r?S
z(oF>WhO;nwA#K)y^sg-+vM@#k*FmEpQ|JzlnsVBw5)ZYoSh0$;C`2yFj)Ds;D34Bg
zi@14rcC%eYzbK7C6W6y=-NY`Pg4X{wE}9`a>ILZqNA**0J)yy)>v%k<37FD;S!~v_
zv#n^xkj7&oZM)$j|LDMI;x1yDmxfW~AI)|nB|VL)95Dm2cv`i|2Dtr2^8{9)B&`vs
zL~qh)X-*3yj~l;Xy5vFwn8=I}#Ofs9t2fWP3jtR#mC$45zM~^qcv_hS>mB#d=T*M!
zAALiydl6Ns=zX#~JXq2q08cor3$&bNWxCjVz&7<d8jF!m4tYr0-%B*tU9m761DBvH
zcEzpu#gBctD@$_S82WD(V^~<xzVmJi;G@jjF(O4ZN!1;+vat(gC?lTjVe6{IP@Y4n
zc~{72C7*N8+_-*IKv(yHxMQDHx6SB-eA7y8b5fCpfbHdszEOZ?a#jaQw?297ON08Z
zN%y9R<MY2<q|v!6r*$-BQZx+K<$kWZR9O+fvb;<6Lr7RHf41TdIW1GFzv1;v^4~s2
z1fVRKaxzT0>}yiM`xS|SJp;NungNjJ<c?ftRoO>@sUokNmH51Vt{vfEOm=;R%Vaq~
zo3Ql?QitTj2NGNBaSEDx52ZQ5q{DZ^CtKy_GsmI4((@9f)17Keq<!@i@|$W3q?>|<
zqv_c^*-<k-!#1b-;D9S=aLs+lI#STxBcBNeSs^A-3$k%?hT)Txagqy&cLT3$7Smr=
zrQ4*jFR}@y$Tp^G=;R@6ID3%lx}uG2e;Oid3$;7`KrXHI=!m4?G;E+c)luWyy5ReA
z5QJlx$-eA+IUl*0VlLW^t-YJo@&vB%+~TD3dHJ5ILg4zY1Rj5int2hZvdzT&D~SBB
zA=Z9SV#wW6yE-x}@O`kJEK71hKbArOi#@G07(0a<yCCHURAYa{GLl%Y@J-t7EbI8v
zGQPj2NO2eU{T<qzv(@3$v&B@w-q2;*ZYOOjd)gg3ELuhtk}Z`ZC`9jbnsfl9NT>?j
zgQ@TGVkOwTNiFCh=C;qG{8007F@Knq=wfaWABfsfFK)}90xGY<7LzH*Rm_Qbp%+O&
z-m56^&@CY-yR+)zGiv{0V`y1s`B7S`ONjqNKs3YUPv8D1>_Dw|V%;+$f}Xsgqxp~N
zvwQ^;tw-8f3l7tz%}jjQ)FbnJ|3!tSRL%rfk8x6XP9L!*mmcaR{4CdX%`N#=of?tl
z+B{X7K#~+bmGFW@l~{Y-(Z*1W`-)pvep@$Hy}Bk4nUuZb-)5xd77xt1!8?gx!RXq)
z2i+~LFXD~7#*|rq&mT)*82i4^NUxe~FdLpcjBle~qN@}eD)Mkx$DH`dyzrM{6UX;m
z6-ngSK1`xr_PmEmM+AT^dzqlx=^x*2=y+QlbU4NY<$IRohi)*4;>&>*1DqQ38VzJF
zv-Qw@U6&xSYFsOPmXSSPOyMCPLWM1l{S<V^&DUX(DTohn*TbK=He#u_KjTJhFH>+0
zhsw|96$S=|iL_d-2+>e{H|tjiWj08i_~Q_+_NX@B_i8Ezu-e17`u7Hpi70q!+NKxa
zIuVM5!ptCs_LT*s5hF^ic6lDt#LC6OEN$d8F*cN~A5^B)cIh#bm|p3Vig*0A83fy_
zdZumW=3PwHc5fWQM;-UeUTe`(KhHgJ?C9oo!C*YZd4Aq~^J;f=X`Gt9w(0AdT7ci6
z-zZ>a0Mk1G8J>$klo$UDW~eWD`W)FwjpdR+g}II`OHw}T9j)HN+1B^urxmO#@Op;`
zeY6u=nLRq)v_}&=s2BUsA7E6!C@cp|If@9r`8v+~!DPUs#9smDB;XciXiO5VsXDnu
zFKf&TYtWS~|FBQOyMC@d*0Q*mI-*BO*>+h=hoGWW^?xd1cb7AA2xTNb4&2+R(dy^+
zp(nwLsQKC?M-dR-dzE^)Oy`yN9sXG5mS<@v{?%6?hc}>cy{JvNT6~-S+vW5Ec>hbP
zz)|2{({#Fva`1Xu3MZTB-E20#7+>glOn+&izqExZMc@S%<Iu-mk%0C7RJlTnzplcF
zFJ`^7`Eqr_sE&RWX1%=MwVj>zbYPJoAE~bFw;4_!yBOVp76`uiU1E&74-{#KW|oRv
zfy@JKLcNj9pg<PdD|GpJ`KeP)YK_E++?z!U=;6T(c#yhdWnPff^WfJ}MYF9Y$rx#E
zs3|(+^`bC$;8_fXK)3KU15I@v{?L#xDtd_A3lDm}0PdAw4gjwnFQJd()$3}ujVI)3
zd0nwS|MboJyaUF-44@YO5tq#aKgb#HhRt@KYu7G6AVQB=bKC(D+ZO|1<X*Ikzs!os
zjHMfiN9XbZ^`jZ()rZD|fB9M~L&6gJU;*}@ZAM0uRYZP%a<_9wHAFVVY4a@FnP$bg
zT!TTFu)_o-F*v>4|FMAZB>HBcI*}r<gE-DKv93RnqA9N(J-=VDu<TIMI`xeeDEB;N
zVAH)%I@}e+VS;kD9p_q7aHa0P*D`dlH$_8A#Y5k2-&zE+?XTW*IE$v9Lgr!F)i|eN
zPxII&Mv7kwS)AOMn3cjWa85zZ_YE4WUO}sb*&~$!!p&_|trY{CR7UK7)By*s%^Gk|
z&KjId2Q+2Q$r(w{C<jbIv|~ulKR~E;1vXIJl50{;(Jy2~Qh$;@eZC1$uai-|SFF^h
z(qu*a(<d*qC3{-U!!Fh3ug8t*GGtY|&%PG4cGPKB6t=#&6&DZsX~Wl=c*&RinlcgV
z)zQBL?|*B=b?D*}g(7$R9}OkNM`o&1l}*gYTb$plqotzfbZJem_a-s*Q<XUD8wqW1
zkVppkXY=wMZ4389_@f8dR|9#GKFhF{mrr-}b@qA3Hw@GmXEwY#QdV+ERgi-qmUz8V
zfnDJSLinwGPzSG<L2ho_6t#-sXi>d3HhwV?%@CfI9uPSOAWIPA=J9*~Gw%=ghAwXo
zdN%3@9}@j*_2V9|#oN_N0Z&%ZANogJ1wzi|(Y`bMY4_(JLC$@E7lkGLV?`V4KvP$T
zs&RNfLeky<Q$jxKTsYLls#;p?SBu%zE6zz08uq&V=+>h?0X(YuZpHEK*g~nnAuQRa
z8n0Xz)r>j&u;TywyRkm3bqq?Y>B;&7T))py=zu`7^u?=ak54dt0KQ!gUANoOL0AnW
zuhJWKqPv=vn$~p?PHRl0pQ(hpA1n;4R?rP#Bs3Qd^yv{ur&uJa;y9jM#m-((tx3{n
zVka!~#&RG_^?4yeq0XnLV|32z&FV}Saf=R48Aw~~SO{N4XI%D-?;WG~QJu~y53Pq$
ziTFX;3_p7VEet+R5g2R+&#44I{W4l2s`C8ybSu<aiTTR7{v1H<Z@E}X)_5t*m|a`Z
zba$V&N)zC*ubgQ#WpH+6jz;!=b1(*M3j^!)&qu~?3Smd92g=64OY)g}=#@zG){d_o
zHan#Ieh->0-dH8n*|hwmN#0W=Xl9uglW$1RK^`h7N)~#t+j&eaq*IB|*j^~PA6@Od
z2V<5}?iaV6%nXIxok)4dkI$P-SKJ*enXjU5XXKo~;iFH%&mVm%YrSr!Eqg-BySke8
zPl9_U(Hx%>Y4NKXn;@bk`*?y>iN21|djzc}OF{8AZjI;GYc>}gmiFc?&L+=`$x-sH
zXOxV@9){=$9oFf?mHfqdKNiR*j#vWk%w8`&Y64#bn(3UqumCEbPUEthv343rBJq}=
z^V5EF-Z5&!sJr(1U{v`o?L;E<UaWHCz9nB#efa|RYf+Tu#^LItE$8#g+%*Uh*B)DZ
zwP7_AOBJizLTYU`@`r)Cd5M}w?fb^YYtPe9*}_`0x_pQ>*2L9m0_s(-^3ns#cpEd9
z3yhx>_73F;K}P|f&)>f@=<1RbBRsM~@(O>4A4)4Y7mnPT8sJ|ub9eQqNb$DWmRB*=
z)aJ|YE38=|MAX^Sl`9b1mEM@`Gr`DRO-dNBm*?NPmRzFtGmJ`(>Nrj;!D_nFYPt~5
zS@SZ7mWkmDRbWJ;Dln=&O?324)0)3i{_lV*Slfe)G$V5v{#kE{MMsoA*ywzxX{e56
zoMw0>^L4wQA5qBXW7gpq$3)`mPr^2yh})emi0lH(zP*3NwKrR$mT~dHR_f#ktiq*6
z(_2Wj8&aDcnaW?9>!5acDO@qYdHMYxC}Ut0etXr&Mbn@ov+uJXww6Fhr<&37e)=w`
zK+cq#%~-(rxnMy%Kveemyw6qlYeRY;bBtYzrGGow`DrB#abRHNj1H&&A$gJ^GIdHL
zu;*m0q9>`^R3uyNd3pJIFajBg?Yw-EBO=Nc=-QM{%I|43egBtH^Hch}9&RL>nA*=_
z{ih1SdB`Bq#rQ0SBV5RIC(3>hDf4sNoWXUe<*WN<j*3ADv$uDf#}l?wQz$j0a)b#q
zx(Y-?H;MvlU;HQB!@&6XM<m&KKCrCi{nhP0ZLhzYdA-yYE}sZq$wRw_(1)~i@Y>S<
z%xbj=Y{LJB%o*xqCv&qC4ln*S|B(dnvJ4aBGe-_&tc#UrF?>Iljri*_*^w7RiW+YP
zY(!30&tAI8r-<G4M`-W;O)-+w`SxlT8QP8VWB^7wV_*<mJpS#L;IH`<W3DIv8TmFl
z($h$8xOXMFfu(XIsJOvzH1pEe{Q3<{9OHsA!1ZshNu4!0v2gl|khUUT<eNP|NbIWP
zHe1^{)ARarasJcs!M8}ca2DnTF>s<jYR)+iUlU64Qy~?8x(wFhU}8~D1_aFD=we_<
zpQ${;z{v0PcUR^}&1EhdzF)r?N*oWgv7;GwW5Xn6`L_lo%JNU_BTSB1d;*dKD^e_9
zx4l)|x&J%O5p~tI6NV}LDMfk$8{=~s31-1c&(2_x708Notfyi4l2JVua<*6b;@w|{
zUm$5FVtvg=+=n~l`I|vNCf}b20|#72f`KLT{d+c}y_DGRPc9y)T=?|YRx<k1DP=Vm
z(KB>b<i3eX)CMjGO~^<4;CK8Oq+_ahsS{*N{~fw0TM~gRc>Re>m5S$W_L3M+-g^7D
zmkuBBtzD?Hu&Ei<27^Oz!iD-{88@H(qg^<Rf<6`>p*xR0_&O2jiAMEMRwuq7|H_^@
zx>1vL$@b2A=vFm#YU5o?4Six}c%cwZiK3O3#*4p({WSLpYA>kT&m}0mILJQ|hy6yC
zVl&n|(ZVgt-YRJ5X$+Q~RRSSnX+H;yqTw&)fbKO)JDXE%itw9>J~e(6)kC&1pgnk#
zFd#(tPc+IO)Z!u-WI>wK(@+~P3|Zm&0?uDA#lpIO$q;^ZfJ=zwMXbyqu|@m5>Id<I
zPsk=g)tA&+vEicnU-Z8}T1^?U)&#!3oOrshU(=mngz@>EY=Oqk%%}!F*K^xQtbSxb
zLR0b8T3wjf@?S3S3G?%pO{qV|$|VeucQ^;%s#wLw7<Y7fz7#CvFg^IGU;n)PBfJXk
zQzs2-I^k*qL_AuarS4w8ocsN)?B!qb!j<DjdLa?(CX5%pzVc%7y}%Wd$%&h*WQ)Lq
zjAY~Z=X&?uP$#TI5WO~hAm*8cnN&5&JGvSc+8SXixr=x&aw`R0+3!qj8gBw>w|=bT
z_eDKIB`5+U-1{80-&f1){!;*3$ZZXB*nL;rQqp;fQ+66+);t&yzc7G*O<dIfM2~lb
zwaCK>XCkpg_6F2$b@y>~vyo(7<j7;-Hmjsg&IZ*#K_(573X`!=T(5VqNlh+Ma-yC{
zd@WTX8hpN3`{i3%j7uwNd`mwzhO}Cw3i_-^CtfYF;`5{4A}CtORq6ZZ=!!oTKd8qz
z)Y&VT(?^UOg*zB20z!oq1y$s}Kl4>;S?w@5{e5XN^YmNQ+2LU&pYRN==R?|GXtIMr
zi4?Eqd;8KyAv3(XW!@}dDQ>%$T|2$DEw?WpZ}tzT(C);Y#>=MuC-6U1Y28{a{cwW)
z6w@<C4gDebUWa?d0GrE$a&#UA9e!||zt>L>VqW{I@ebpo<hd;dX+hYNj<=5lIxEy%
zdGq=6X=?2nN=pXRiI6{8FyxX6j>|R{mXz$>y{(w3-Jf6twm|>((Wq4gjT-2PLj>qN
zy@)+Un_a7YB`(zE`K>^wt+cvyl_53L5p;+6F*i^~fbmZb(kEvTexTwHrdJG?sn*%M
z@Y^8>@XV)m5(cLaDV|*Buw2ESyvbL>ko(oJvAK`m2`;LC@IeB)!anLe|EJ)___sf@
z&0l5#c|-a9G{LC-hE`vnoaG6IiRQmL3@%HyJao%fxVBMxw+M=H8tebNDe^yv^(P^j
z?Ie>(pWgoYKZ3u%C5}lGeR{r`PL6YjD-W~xa`@ZSf2jTUuI;oU-7f{dj10G7)0btq
z|1A>ucl7Pc|7M+@X0dJ>j313|!~a_GA13}g_xQ*EjfQfO|IaI6kZSG>uz$OVb#+Na
zyKNBueYgL&m4W|F>7IUrK7FS7zfXJfzp#exkTi>Z`RJb+|Mj#$B|bxh+bc?`aLI2g
zxN!38|F3F^=bWr7yDXpK{Ns+^G}dps&9G6^|Ee}jIV7(4y7d3TdAci-PU8*#&C@bF
zrY;lru6zb%mKt40T?5()<ODcU7VynaJSk*zBaGpxf7DN<e>#M}WzI?ba(?mv_}41w
zE~p4KE1ZQO`nnM@lF;Frt~gx`6w`==OHV|5(9<vf{nP)7Gq3Qa@nT+Cs_uBKd7INn
zUpNnR?99`6Jg9KSQR89anDqa&_nm=kg>U=1l+vnIo6?5bd&X!LEsC14N41n1RXbK|
z*Q^m*gsKt5s1a&JjG|_2LQ#9~5X5+E^#A|z`|y5$f6vF~jOU#DzV2~dCpkxCj!By(
zjD{nFt4hi1XlF?C{6@AkyiFD@WCdY2Zcz=)DeQB2RN-&Q|HSvTqHiiOaj_8k*XN*l
z<kq=~ejbe%rr+Z}voo(eGK1!szKsaI81vD?EZtiRHTO%}c&%&2C^BMhlhP8}2{=e@
zBNX&LA{ovh@G$urWkQI@6Xn_1PM9o1H{GiGM1BvPJl=wK;ZX)?U40_L<YL$Ni`XH^
zu6dgwI}~d8_CdkkPlC>@nn@L3S9Ufb(g5KZkadDMmBsoRY3kw`@_K?==A89w+Gnku
z;r!~6pWCf|?BIFS=hpdi`5)>wew`GkjPZ5Kxehxmgr_ctVk~JQNs=<3i#kf}7Jc}I
z`bx<E!2c1%EbgPqg@?^cZ3pT3lGAjZj$iK<`l4%Uvzq&Zs0m2}MEWb9e%HQtj=1mS
zZx!6p9$~k)mNAY8$ZaEpdJu=2AgQgpo_#Xo1uVbE{mm}6Gv1NvJF79SgaORdmm3p{
zd9l~#7jn)#PK!v;@RLRzXepbFr-mof0LJ?q$wSG>iRzwz0G>Hm09;I$g145NpPH(g
zpPGxUsyS|_z3CuGdi)Q^QlVSXX)LII^KM6pty(46->8TXcs+jl%^QM@F<xc(-`)|V
z5?+yAXlAoCw!}4gN%ejD75B4)&C4?SXvwl`9y-7Jo4kA~4;lZ)`?)jfLMa~HsOH_5
zmH(A($->#+34*eu@;8|?=!=NUw%<f!O}Q@?>9#Wc-K(0FdrkC<T*}@|Po*EK?GsI%
z-BuZT?Z&npK2sXE=EyRbu5<x&V@f_gJ{(N?B-%bYizZFN<n3F-nHpB+U9e||`mHZ_
zcW)Qdk|N7>3qQr{rD_yCd(&K1BPGmUWJJsSltW?c-9BD!ycJaik`SgH{2V^nT(@)k
zA{d4~aiFdD%Q#MdH=Wp*_T(?9x4eU2ceF!Ke%5%kZ0JW>PB%a4KnCR$%^Yn_Y*+V?
zxtb|PkPusZVZJg;#1>$1-^@uDgK;FG|5<Zk-ok?1m~9~y4On8zZx3vMD9O09lhdMh
zityd~$gBtc)yKuJIbnUd`j+3;leE7tlhLhlNx@|rl-2cRTHB$wO`MK{<+!*;uI2D$
zLTGcyrFaE*--yx|YZ0<Ym7JO5?OSw*^`65w|0UJyLViqI*MxWUsLGbW_)w6ErqkN5
zCM3s=zBe=dohEwUHE#QJTO*!CD2Q;fOE$={H^{9We5Paa@)wF7RDyIRJuU1{p9_{b
zYpedAx$ghT#VZmgV?VxTKGf7tW6Qr(f5T8L^@b-ROK#m`E#CxaRg3#Zt-Tg@XJI?9
zt(X~@A@-a_q-#*;=)|pHFMwfDU3)-05ssJKXhwv-AvlK)!ALkI2b#Y6sbjZ;M2R|V
z=oZw!vU)QY9kh5NyG$D=I<$IVJA~vv+WJX;ZP3Enka$24V%mpM=nV63JC5?0G^~f7
z<!Azq1T{~I3NECUVa;@dyDEoVKfLUU$7A$h-1ov-#U&RdE)y85<L8N~pMUiEQ@xC6
zzd9D=qT=<otc$<3-~K`^$H4~p-JYm+Pb#vkhHK|}kFqE$&vC@!r592?36PT*>&_48
zw3wIlnBQ;35M8Zfa}QBX#qxYT+&9|Y?!b%k0(CCs!H<9Z{w?-M|4(mg{if;Z56jdd
zE9pRPNi9a%^UcxQTfI3xO@l_WBI*kYsKvl$5nVeUJ+<r+q1nut_eA1J<&}$CwrY<;
zB%K>C7e(WUcm9C4KD&GyYNeh9adoBUw63&fe=tH`U@G`3J9Njqjbz!aQ1Wvnuc<Vk
zO|Dj4bNykEnA?7D%T*2b0l!^=bWr!>=#C$l<VgV`w%1dxlxuMK-cjGZ2}z%xo(N7B
zzHm|m3nitH&v7OMc<f>NW#bnwtFe6ONLtO_yxw-jN9E`ViaH1U%g_3oeG^=x^BlGI
zT|VU5&Dl$L*&%t3fD|ie;s|Jaimp?bgqdh|SW*1gDegt0E{%5u-;R4bHg0I#mXCp&
zFx+Ya=q~13V2<9cCVSAKeLSlNQ+*U&^J>Jh@3FkWY9hCTJhoGH0APeW*~!cg*HwxL
z^(?w3a!PhiKW^R+1lI>%$d)>WuZraRF42qp92^`0zT3&L<k-2dyD&U-lF%!gS~g4R
z(hd(lSsGx~RcQ-$BQLtvh#PETSQAmK$@W4kaeW*#M|!~r@}LlfSG=`8d9{#FNi^K%
zn9+7-w@{DVU*cP}G!)kQz}#Nk+a}i%DpPk}uzOToXku9U<3!%$`%&srZc(pTPHz1!
zzjWaYuzh{tEbF+YR^`mRb{dX-?5`I}xKY$2GA|h8u)WFRN1~>6i(HU}O^!Ibid+V`
zy1mR+SBuv;CI+gE;Lj$`mJ_lG_Tlo~n{d2I@jMsH3z#p(rSkI@v#0l^@E7@{^w=A_
z=>p0K(SbkSNhZ|h>HxA}Q8bz2Bf2quD`OwC7vXuOd07C!4n$=sQeh<chO)Av&**a#
zzg(TM9W4){Tn`X8H;>eh>HdSx7I9{NuCh1TOh4{gXeFxJD+~kox@+i8Z{(|XVPV;I
z?ZH~4&(0VB0OERW%ErQFr*7awK6@$8t<HG5{uA$_#ghsTu=5dS%o?Q(8FYR2-Np6>
z_18I5kJa`R)j4!iZ)wQ(n#bxcggM&}nc+YT^s5;bV#OuweJp=a_IjED=B9rv2|es?
zpqcu;D5Gy`>24VT(v*KA@GNxv;p;*LeqgolI?yv)ob)IuX=6B6`~?qs?SMVlBMUM(
zJ6p4Dsy_Nn#-)YKDSJbg6SH3*!z5!0>1EFfdrM{Ai#m`zVcc=GU5qj#VfI~bzBU&+
zY$hF^=2pH>Z%5m`ZyLdMmNFLni|)Q%j1M7Vl}kNT$d2C?m|^<yNN=+q4^uB+z8bbJ
z>e=MK<2a7hVoO^VK;jzX6cY$^c`*lyU4RfrY9p~%3HtW}LNTo437g*_n`?vl1B+?U
z5ewCCK-^9gDm8I2h0n19^5><0Eu>p^%@qJ%j1hOMje}=Um>%q|D38)(Rlm3Q_3_cZ
z*e&~H|9ET;%ALADo8GxUKf;lVKX{F7i+$he-r{3=QJEUsSBk@%uc>L}B#ASf<XCeO
zic|^eQ6Wzn63Ec1^iM;&-3NFXSfW+S730@lyh%15=;>#XveJ!5q*d{@T7vbph;wx^
z+2$*%c5_uw6P1PX6LJ=g6DkXzAhk1anUe&`Y%N{JER|ogVPLwDU9hg$13M6F|Lbw-
zpQjA{v=77ST}Vl7Xr|6_QJ`B7g8$RO9`1V7Ni@oQHa?5eTV@r#VaZ-|7!O|cyRk*Y
zD)^y(aC7D3G)Q=6^)dnPBw!Ob!%OiT=SfZXimuDbtS=MrNW{9{w*qv9ai<{gmVu2?
zrVPJD+6@eOI5lG(+2)@g!bClJtv?Cs2WANW6;#1XPh4;jut63ueyDWf=T&tq^;I`I
zk(<xH_OXH>cDUf9YGTK^utl<Gr6fN0vh$UBz%%}-w5#7_*FWt$S0#U_oVmnoT5hmZ
z?SqbrF*MBxZ+vY@rgeG&+uwHVOqHysN3)y-3u@d{CtrZ}v6~fY2BON0ZPwk@pObxV
zbvXw-3OrN52nb|i&l)u8e+^n~@`8&%a#35f<M_#+s(z7b&067x0I$&nU6`XNN6%7j
zmSuGh__iOAJF|E`7i`v0+W`-^8TG;NSufXIJ=UnLp6Xm~yjK<G0j>R<6JFF^<@sk)
z{k#v>;7vPL@O8EN|E-?MhDZ+Amf4Bn#(tyVqd7kV2j9T>k?RpO8c6rdAOYab2XEeG
z^<#B1;ji^@FAXAm@&er`v3Y?Y$)a(NnekeyaqnBAhNGX1QVrM(;%!6>Wt5-!-VGj$
zlO|7QyC6ml!=znDrD^!r=n)wj@s^KqkC7Q6Ixn^Fy(xw+2piZpeH~X;r2;^5K{_&R
zPjX`^8-!cByl)9gDG!WvZY@Eh+!OLN04w%Th9WjXQ4LB;nTbZJ>`L!;xhU`T=e>Ub
zf~=PnkdG;o#IKyLhn0`_JFM9Q5C-@*#^6MyT<6m}MQQuNqsf6(?}%BOf42!O6V>m9
zE-=q&>pelPC(4MqsI(9dAG<4-0F6v3gOhVCxza`(ftgE2#K5wJ)$-z^(oJyhcFGl*
zz(GN36eA{c$<m_ROZbA|v91qh;x8*&AWFgta?QxrQXcD#V^`aq({8wrn+m_&5A;}K
zfo4N)YmI6tHO*Q$kp<)I#1NvAD$CX<In^UDiZOg!%|6e(asePiv4e@SS9TzhM3QNK
z07l^!@!CtI%Sk$%d&RTMaLM04uIn4@*)Fvk(DL`3;sqa+?zV}zu5J3YQYdP6rOMd1
z%fQZ(@zV#OYfVL>U)@b~;c$snvn8?r?Imu<p4fqkBVbe*=MEsGFuu%37vH3uD{5#N
z;d4+fPVr;4WJj5HdpOCc)=zZyM8E%OQ>`@D(klh&;<sv_cHp0j{Z7m$p;7C%OJ)<a
zQ#B$ZzZ@S#vK0b!`$yEZ|DvC1+;23!ic#uQ35$k_;Mb2<JVJ?|m?@^^K1{soDtCz}
zK7B6{U5SFKss1sJ$PE^>SeSAoX5O;m5~4_b77kdwuMbuRm2vh!d?&6&4(z&^Xne;{
zV9xeUO~)+~v%R-_^4~z52r2r2mUk<blG`)aB?r-?>g&nHZ`1dic{*h_Cjr&*ywni2
zNGW=wg|*dbzOwtXieBVe7RvWAo`UvMScPplo}Ls0PCc$|#x9j6WKj3go1zuYwA2UX
zhx3TnljtGtWmi&u*$lXj50^z+Kuv6llxM{7;m?wd<{-IeZ(^KBt<pyj_868hb>~|B
z&#oaJVf&>XzqeTk$J}!D*su{FsXSd}<gR86F|M8S4Q$ilURI3nrcf3(gfh*A$c1pu
zD%Ezoc8fiPu4kCh4If%@_o$#qCComTUe<4Ym>3@tExGn6W@@kaQp~ZAv@(&Vc-e3#
zynF+40_h(Gya>Nq9ttu@)P}iM)xc0P5L+|aTP>%ryvc>ZlTtF5w24=((Jp!>SoOhQ
z>16?K+uyc%OP|{)gCyJAPIj!J1?yZ1cZDuqWCP@X=03Ypap{5_Veiq<B;vJGt%W5Q
zn>f0t7+pfY6WU35?5(+zIih0*fnI9g)wz_9oLWpyx}?YWAel2z(NS6T#C{2%%I}7{
zY1-e9%uvX_na@iL#dtAO7434>j6~~f)XCxa(CYTP&!EX<N5)b9f%4>!HFw<``>Mxz
zdi9S%xicUG`O(MaVCaZ}NpAj;{IUprNYP6(l-J5z=W0ot<s}`Hd7cy3b)Qj^8&X&k
zcK%;STS2^LDnQO>6|)NBdRJ=U^iIlWF#1~b_qzX*r7L1_+ikJ@GM=(?F>rBS+C52p
zm`tP381k7TVnC!ko?3hc`^QyIy4YG8qLQ9pkdQsJRHZs4th}<Oh~*7SD3NYAmS?Hn
z-uCDV$ZyMINtN5vZh0x2+Q;@AIbxMj_`?|Lz{{Xfx677knjLHALyWOAlV$DXQsy8T
z->G>fJhefr$(^M7?j=B`3!}rzh%4XX<mMYPyc4O0n=O4>+Mpb})`pRjH9}m+-@C{s
z$GCy?ir#Pyg#(U!gN@Yia1z#;k80x4*q~cpzv;1DBmUzo#dOBbX+<O9DsOLeCZE@|
zHkl!rPIpRb9WyD9{UJ?*OFzQ@)jK6VzHpT7U}@kdrg;=RK0$xFj$E!R+)I!&SM>75
zJ_i~p5}t#@ic(0oziHD2y!w8(2T1$a@v-T>;5Tyr#H>qYc5RKa{&gxgn^7baH}OG>
zkNVroYePe$T*)3YT+qn{7J#egUt(bctj}J4a@Y`dWi5hcE7UaA>;NTCt#^#THTmGr
z)bP{22?q<ySlqIX{xsA{irrmc;_9Y7oBuk&^Z7el>7kvCHecRl^Uyls^FHRt9lWaI
zCGd)O>C!));MWW{kph5#l<A)wfbMp@Ro8u?Ul5%AartbR|2P#xPiHhk)ZDH6p2k(Z
zlHd=lF1kAmRTs$%HdMT6N=^Fj!rHriGjUZ0$?Hth1G~zea|33kF{G%2n<&dQZqE8g
zZuJpMy%=%XcByDJvWwUumG1y%W|$H-(q9vh;1Qw%j#c?K4^>2y_YrAUy!0-0wom9f
zRz1Nx9N7iunA~k-o)))?+KB1Wu3reUzHk=*&T}3eV|jhra3xsZg0Y1Q?WkbYP2Zog
zQ@ZdOTc6SEQk~V|Is!@ms=6P8tRoM7EI-H=0G6knLrWF(q|0&ywP>f6YM}PkQ?>(t
zJR$6w;<UBm|A8pg`DY9ziw>ZOD*Uz1BGD*gUfBbeT>v(G^N)dS9-m#8q~k#z0QF9s
z)t4BWefg||JVB2mABy?_JQz`Smjjha$pxhEKTj=xpszV8@x(hqe^tlJ?Tx#Vje9>+
zflT#4LDVQ_T5mOCY``@iU*cuOh|{sILI^KgmbvM{9&<~8!}JQi>fF|OPQ)f3s6eEw
zzPs@&4Rl#KqP_*^{7$;{#Bcp*!Yg2YWBQnd;5~=Pu56&mz`Cr!{3k9GUKGeN$8NQB
z*(h&KpZl3R-@wk|;a$YTvxZ~irPX-I_liqQCn9vJNno77*X1!8cxQ3X;J6>vr7O0V
z0#Bj*N00*gMS}lvqL9#ig<(UZ2m}bO;cM#L?E1?UoVG~Lxw$QLkFrr`GVkj{i&EOi
zfmdB4i)!g@Uj(fTRrKTK+4EOB)!d6{K+KZX-Co+g<C=4$RWL}2+v7eAr6>*1mXjlT
z$X0FcGb}#pe#Wzos8;GNb5_digNE?tJ3%FW1s}AO_}f%yVY<0M|4#*o-QE2@>L-P?
zQ0XH5^j0W&^6dvBZycB<tY4Go{a%;D8>et73@5=IQ>Cz-F#FVKrp9d$RW(%Z{i!Q!
z%MMZ2`>XoBj@w_RD~{jx{`~U+z~kCNBId2HOZR7Z1B$`ZG}%bGQ=^XH0~MKWGt;$l
ztDQE)#P#aY{ciT_u=cH+4aIl3nWVt=EB5FX#hdvSd2;3ZaRshOq5V&QD^kfmzH2ZN
zCHHYZj9!%uuFaNpPQ@rih|#FhZ#m);0*uzsIZTFUqlXBka!t&9{#d6WClCreov|(o
z&O3!_vzFtnor?(b*D{C6k|B|6QNQGT_Zz*BmnEC-g`eKS#C1&+vZ^ELSk7-+Wb+2=
z<qWGC2S3d|R2UU86GKK~Eq1-s1YpV<$Y#;wKlbGv>p}VU6iIbTwk&I)7EnfYV7mtz
zYtmUh%GjQBC%n>Y`m4s6u2PT`9N;SC`g_;i8OrhZV6nBVPLlDDOZYM6f@@FsSCTz|
z^5v~d^`VFRaqmA!AWuo`hUe1G8$!A-*iyw-nVxpL2fRGo<%BjZ_g?~RtK?y}#w$M>
z)c^r*qTZQhv9^NoARhx`BTw`lDp;~0K(Z+&(A6*Xi*sSl^E@UB{BF2rlx=y{v?3l=
zh$5#tSm`gWq#Frcqq4!}zho&{W7gAy7U}*h$IQe`N@eB4<8BVP8~#iqW{qPgWXdlc
zZMl^*z&lkSikZ&Rh!o|Bg3L__eEgY-`kW?P`{wmKwXQtQ%4Kz*!Sy{o)|n4nm7QEK
z-?8R*@)BYy*y4pCS`K7%s{^6uf$`_#f}#Os78&Bzx}!uf3g2ju{t;~7nGz)Up`rV?
zfDZ7cTDDlaNUC({bU<&aJ~tQ6*TZuMzdh$3)c6@@WfOMeW)w7i>1~VA^g7ub)51v4
z+XLfSb;X}Gn4WqLLutN69nZz)n_2QiJ9!|##2z}Z+ChNaQKP9l`@%qqG3RdloukIx
z*uF7)F)C(c*~{eymOxkE{wPn$oqp?C_3hYad?SS`BNbEa<x#Q<xn+bI-3x1P6z-m7
zR@x+X-&fhw<LZOJCtehe9FNDkZhCDAKWT{{B=s}{99_rP;y?7sxSk}L$>FO<x^&fo
zq;!ZlvAZcQcGV^h3);B_lfO8S0p)nWb>Y1E0-oUldQCVi&>hBHdThzLDK)jUW-{+Q
zp7g03W3J-x!ew6<MyXo?1OCNg(sD#+Edx2TMB$mJ>D`ewgl$$*P-)>sZoIL6#f2VB
zdF|@62?IW<w2xiW<>^2s$4c_WK>$2{76_|P-0EX|k{)E&7c2V=HJ2owNI101RXYDE
zZsR+w#b_#0GftKvO4Ew??W&8&oz&I^USypnXb*ziA+r}xl*7}Sx4=lFkhP-@Hk#ST
z^06-IA175@zuvx6`k4{#WT;q#GN-UH;qGi(*dI-pjlaV(VpcN`ese-N#=mGq26P(^
zDHPSCSs%{+EV5Zpcc%yEpm#^`MahWGy<Ud)peqo_mFRWM1w@`AO}`H_T4N0JRE<_}
zqq?4h4Y}<lX)4dJw~giXmTf1_b08&In>G#N5{e~wEuAuse{yhCP>YHF&42-|E|#Do
zXCnl$vo)CvgPS|q%u(ncP|8_JmcC<mXZuK0Kxnt+q!?>v1<NE&^1yTNFCo(I%!nHv
z*pY^J860k<M`E^-+>*xmdT2cxmSge{kkedUhYOy+-2Tut@HV)}3*KggJ2cKJ8|Zxw
zM(!Oz8QeQpSH%SPN89_T_^x{(562L*>)EhEPm9A29W`{E=d{q^Va{T*{VcU<yo<H$
zYN`*(VCl8w<6$0uW2Nzt*x(x)DD8EntMaGS^x={#|B?!-w_OR4r=T1jf^_zN$^sPc
zC|-3K7<X-P+cm&?FY)(Lhz3L#X4-ivV|HUQ?(n2m@;vTiis>r>DI^{t#~Wka#5qcF
zg#`166)~H*O+!};12|}6vQG9Y7PD|XJn45~lm-X7$f$}u><mxQXire2OAT_<0`+tG
zWA-TiWVOQYh$CQH#WI;prHcB4$;uCxjJ8d^xP6y{D@WDA?)NUIpTT5iZ(UQ=nj4jn
zN|U^|GSz?S+Ekox%`5~N{!K~3rz&!+xEMJftj~M%h~bMEKuP$!N{gzj{A4Bw0f~%y
zz~>zAj>-&Y24lO~aubdn5S7-xrW!j-bPwHN*Q6Q9$#~h$!Y^8F27%wqns8SJ2RoRL
zj0BCcb1_y^VZ!IUYn%p+3xrhQ{VKW#O-e0FmHoeZSLQPt!a_mg1_yCEL2+DeURx8h
zOvgn-QTb4;icS$9C!1bJVpmYP!KgI%VoAsqcZ|vUQf`)C(MA*keVB8g8}UVA?unz{
zy1UU_z78oSN!$|GM^bqDz2d(K`YC*`Gtn$EO)wXa^V-Qo9o0LP=GNb*vS>@ia|K-`
z>uCces%}leT#(TWsI~b`em@hC_&Xg#TaTZ25mITMaFurBWAqn4T{Ayj#~N`q5lO)s
zYl9zED-YvB06uFO{wK-1y@%#DzFSerNsj#@qK2<n=@EMHL6?wk<Y<K(PGhT9Ax;;G
zN#nK;W@JngWx7?SK0=Wnu|WCGe&|M_6e{w&PX6M-oLv9OuB1IK*5Nz5s8Jjjb*W*X
z^mjjM)-}-zk{*-C=>yp|2p{^2amlLL4jVGnsuhc(9LX|SgaZl}ci0_Snyt(Lu{Fsv
z973rZHJZ2^ZhNfITAkx(FJ%l}<nIrvU_=)=mhP^HHa+)sLRXonsv~x()+)V+tORi)
z@7&kx*N@g`w=R+4MH9?^7cYT;^iZAL{igbxW;!OqWHpv?ozKA*EqOl4MlyVoPsvv0
zL-lg68F`_U-<q@+7U`cHJXwbh3>iRujONONd)x69-ZizY$iW&~@*(W43<99F->L8-
zi<p`33t8^AxL1-E6+uGrZEW!O>mT6C#jw5?tgM3ctfU2quC?BL_0Uv`A{`Re5w(H&
z9YvE@@|P_|Zdkl_G6jJ91AFqGLwNLwoX4P!#!N#;v4vhaw@5xxHS`Y~E>Heo%wP~@
zmC_mAFEcIeae-&2KXoKeMaR-u`h9mpnb!HqoW=0Q=MYorVj^VWTkx6e`2Jicsmlni
zK2wNfn*o7!jRBDJM?dN3-SF9w1O#Nn7dbvs6@2OZY<pIwTKpT||6osvLtV>d1%fyN
z-x;FQmbh-3Q$iq6nLzFx8L|4=qXzG1@Tgqd9CPMBXckHK4`(aywMvZ5<OOqgTsdSn
zNo>0Rqhaw>`g<R(rk>P#zb=g(&}mL#Y;l(LH;OZVS@*hP00WB8t{vW$;RAAR(aqIU
z>z#3fiT30JrBBkIJ?jGQ+$ZZwJ|Nr{2qmb7T8Ex)tp(|9HbrGD@7ptw33L%}Hvd+`
zw5F(QoMrZ44xI0CxZJUJH6k+iJ^ETJHTSy{LEt{{*|jqZ(29vO6!~CbCp?8y2c~;k
zO5Sec{g7e{;(yzHAr&ebAaAG5<}Xr>YNc5a;t%sZybsA~Dd?oYDGu-SwmLF@iDU<!
z8{4DT31{^m>>GG)bfR-&cKe+1oCnf>o;<Hh|9p6Hw*NT!pO^Swtb1{`ucIf||FWDD
sBl~xZa>JOjP0EVu3Wf6K{sjLW{~iAw{~iAw|7XVk00B8n_5i#D0O)F`D*ylh

literal 192
zcmV;x06+g9iwFSxuy|tt1MSbv3c@f92k@Qu6nTQt+HN}w?%+WX@dY}TxjNU*wnO*!
z?gR9sco`z}cli?%LUPE~n_U*Uy9*XW2uV1HG1DZTlC0)>LTL<?GnP<BSjvG=7O&C(
z$Z{vWbk+$ooYGWhR2J2{xnZm<KkS)bfoJ}SV<jzY_r0sM0;L`1TGwzxta%d2wpTe6
u8tvEugV#<P1fhBmMIoKkN?gL$=#vqRjly3)<9VLveeD5yE_j3h2mk;GRaqMV

diff --git a/doc/pub/week46/ipynb/week46.ipynb b/doc/pub/week46/ipynb/week46.ipynb
index 0f0a8be40..d892a5475 100644
--- a/doc/pub/week46/ipynb/week46.ipynb
+++ b/doc/pub/week46/ipynb/week46.ipynb
@@ -2,123 +2,217 @@
  "cells": [
   {
    "cell_type": "markdown",
-   "id": "6606ca1f",
+   "id": "412fa071",
    "metadata": {
     "editable": true
    },
    "source": [
     "<!-- HTML file automatically generated from DocOnce source (https://github.com/doconce/doconce/)\n",
     "doconce format html week46.do.txt --no_mako -->\n",
-    "<!-- dom:TITLE: Week 46: Support Vector Machines and Project 3.  -->"
+    "<!-- dom:TITLE: Week 46: Decision Trees, Ensemble methods  and Random Forests -->"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "2152d350",
+   "id": "06eaa518",
    "metadata": {
     "editable": true
    },
    "source": [
-    "# Week 46: Support Vector Machines and Project 3. \n",
+    "# Week 46: Decision Trees, Ensemble methods  and Random Forests\n",
     "**Morten Hjorth-Jensen**, Department of Physics, University of Oslo and Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University\n",
     "\n",
-    "Date: **Nov 19, 2022**\n",
-    "\n",
-    "Copyright 1999-2022, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license"
+    "Date: **Week 46, November 13-17**"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "7e626509",
+   "id": "f58528ba",
    "metadata": {
     "editable": true
    },
    "source": [
-    "## Overview of week 46\n",
+    "## Plan for week 46\n",
     "\n",
-    "* **Thursday**: Support Vector Machines, discussion of project 3\n",
+    "**Active learning sessions on Tuesday and Wednesday.**\n",
     "\n",
-    "  * [Video of lecture](https://youtu.be/F3CkH-opbdY)\n",
+    "  * Work and Discussion of project 2\n",
     "\n",
-    "* **Friday**: Support vector machines and start Principal Component Analysis\n",
+    "  * Discussion of project 3 as well\n",
     "\n",
-    "  * [Video of lecture](https://youtu.be/BbupEEvMXtg)\n",
+    "  \n",
     "\n",
-    "**Reading.**\n",
+    "**Material for the lecture on Thursday November 16, 2023.**\n",
     "\n",
-    "Reading recommendations:\n",
-    "1. See lecture notes for week 46 at <https://compphysics.github.io/MachineLearning/doc/web/course.html.>\n",
+    "  * Thursday: Basics of decision trees, classification and regression algorithms and ensemble models \n",
     "\n",
-    "2. Hastie et al chapter 12\n",
+    "  * Readings and Videos:\n",
     "\n",
-    "3. Bishop chapter 7.1 and 7.2\n",
+    "    * These lecture notes\n",
     "\n",
-    "**Videos.**\n",
+    "    * [Video of lecture to be added](https://youtu.be/)\n",
     "\n",
-    "1. [Overview video on Support Vector Machines](https://www.youtube.com/watch?v=efR1C6CvhmE&ab_channel=StatQuestwithJoshStarmer)\n",
+    "    * [Video on Decision trees](https://www.youtube.com/watch?v=RmajweUFKvM&ab_channel=Simplilearn)\n",
     "\n",
-    "2. See also [this video](https://www.youtube.com/watch?v=N1vOgolbjSc&ab_channel=AliceZhao)."
+    "    * Decision Trees: Geron's chapter 6 covers decision trees while ensemble models, voting and bagging are discussed in chapter 7. See also lecture from [STK-IN4300, lecture 7](https://www.uio.no/studier/emner/matnat/math/STK-IN4300/h20/slides/lecture_7.pdf). Chapter 9.2 of Hastie et al contains also a good discussion."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "20ff57da",
+   "id": "fe8b81ce",
    "metadata": {
     "editable": true
    },
    "source": [
-    "## Support Vector Machines, overarching aims\n",
+    "## Decision trees, overarching aims\n",
     "\n",
-    "A Support Vector Machine (SVM) is a very powerful and versatile\n",
-    "Machine Learning method, capable of performing linear or nonlinear\n",
-    "classification, regression, and even outlier detection. It is one of\n",
-    "the most popular models in Machine Learning, and anyone interested in\n",
-    "Machine Learning should have it in their toolbox. SVMs are\n",
-    "particularly well suited for classification of complex but small-sized or\n",
-    "medium-sized datasets.  \n",
+    "We start here with the most basic algorithm, the so-called decision\n",
+    "tree. With this basic algorithm we can in turn build more complex\n",
+    "networks, spanning from homogeneous and heterogenous forests (bagging,\n",
+    "random forests and more) to one of the most popular supervised\n",
+    "algorithms nowadays, the extreme gradient boosting, or just\n",
+    "XGBoost. But let us start with the simplest possible ingredient.\n",
     "\n",
-    "The case with two well-separated classes only can be understood in an\n",
-    "intuitive way in terms of lines in a two-dimensional space separating\n",
-    "the two classes (see figure below).\n",
+    "Decision trees are supervised learning algorithms used for both,\n",
+    "classification and regression tasks.\n",
     "\n",
-    "The basic mathematics behind the SVM is however less familiar to most of us. \n",
-    "It relies on the definition of hyperplanes and the\n",
-    "definition of a **margin** which separates classes (in case of\n",
-    "classification problems) of variables. It is also used for regression\n",
-    "problems.\n",
+    "The main idea of decision trees\n",
+    "is to find those descriptive features which contain the most\n",
+    "**information** regarding the target feature and then split the dataset\n",
+    "along the values of these features such that the target feature values\n",
+    "for the resulting underlying datasets are as pure as possible.\n",
     "\n",
-    "With SVMs we distinguish between hard margin and soft margins. The\n",
-    "latter introduces a so-called softening parameter to be discussed\n",
-    "below.  We distinguish also between linear and non-linear\n",
-    "approaches. The latter are the most frequent ones since it is rather\n",
-    "unlikely that we can separate classes easily by say straight lines."
+    "The descriptive features which reproduce best the target/output features are normally  said\n",
+    "to be the most informative ones. The process of finding the **most\n",
+    "informative** feature is done until we accomplish a stopping criteria\n",
+    "where we then finally end up in so called **leaf nodes**."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "3af43c58",
+   "id": "55c9b460",
    "metadata": {
     "editable": true
    },
    "source": [
-    "## Hyperplanes and all that\n",
+    "## Basics of a tree\n",
     "\n",
-    "The theory behind support vector machines (SVM hereafter) is based on\n",
-    "the mathematical description of so-called hyperplanes. Let us start\n",
-    "with a two-dimensional case. This will also allow us to introduce our\n",
-    "first SVM examples. These will be tailored to the case of two specific\n",
-    "classes, as displayed in the figure here based on the usage of the petal data.\n",
+    "A decision tree is typically divided into a **root node**, the **interior nodes**,\n",
+    "and the final **leaf nodes** or just **leaves**. These entities are then connected by so-called **branches**.\n",
     "\n",
-    "We assume here that our data set can be well separated into two\n",
-    "domains, where a straight line does the job in the separating the two\n",
-    "classes. Here the two classes are represented by either squares or\n",
-    "circles."
+    "The leaf nodes\n",
+    "contain the predictions we will make for new query instances presented\n",
+    "to our trained model. This is possible since the model has \n",
+    "learned the underlying structure of the training data and hence can,\n",
+    "given some assumptions, make predictions about the target feature value\n",
+    "(class) of unseen query instances."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "13aa8693",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## A Sketch of a Tree, Regression problem\n",
+    "\n",
+    "[See handwritten notes November 3](https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf)\n",
+    "\n",
+    "<!-- FIGURE: [DataFiles/Regsimpletree.png, width=600 frac=0.8] -->"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3fe97766",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## A Sketch of a Tree, Classification  problem\n",
+    "\n",
+    "[See handwritten notes November 3](https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf)\n",
+    "<!-- FIGURE: [DataFiles/Classimpletree.png, width=600 frac=0.8] -->"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1e40a285",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## A typical Decision Tree with its pertinent Jargon, Classification Problem\n",
+    "\n",
+    "<!-- dom:FIGURE: [DataFiles/cancer.png, width=600 frac=0.8] -->\n",
+    "<!-- begin figure -->\n",
+    "\n",
+    "<img src=\"DataFiles/cancer.png\" width=\"600\"><p style=\"font-size: 0.9em\"><i>Figure 1: </i></p>\n",
+    "<!-- end figure -->\n",
+    "\n",
+    "This tree was produced using the Wisconsin cancer data (discussed here as well, see code examples below) using **Scikit-Learn**'s decision tree classifier. Here we have used the so-called **gini** index (see below) to split the various branches."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "0f039b0c",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## General Features\n",
+    "\n",
+    "The overarching approach to decision trees is a top-down approach.\n",
+    "\n",
+    "* A leaf provides the classification of a given instance.\n",
+    "\n",
+    "* A node specifies a test of some attribute of the instance.\n",
+    "\n",
+    "* A branch corresponds to a possible values of an attribute.\n",
+    "\n",
+    "* An instance is classified by starting at the root node of the tree, testing the attribute specified by this node, then moving down the tree branch corresponding to the value of the attribute in the given example.\n",
+    "\n",
+    "This process is then repeated for the subtree rooted at the new\n",
+    "node."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "9f2b1935",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## How do we set it up?\n",
+    "\n",
+    "In simplified terms, the process of training a decision tree and\n",
+    "predicting the target features of query instances is as follows:\n",
+    "\n",
+    "1. Present a dataset containing of a number of training instances characterized by a number of descriptive features and a target feature\n",
+    "\n",
+    "2. Train the decision tree model by continuously splitting the target feature along the values of the descriptive features using a measure of information gain during the training process\n",
+    "\n",
+    "3. Grow the tree until we accomplish a stopping criteria create leaf nodes which represent the *predictions* we want to make for new query instances\n",
+    "\n",
+    "4. Show query instances to the tree and run down the tree until we arrive at leaf nodes\n",
+    "\n",
+    "Then we are essentially done!"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "667dbac8",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Decision trees and Regression"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": 1,
-   "id": "768f011e",
+   "id": "224b7914",
    "metadata": {
     "collapsed": false,
     "editable": true
@@ -127,390 +221,191 @@
    "source": [
     "%matplotlib inline\n",
     "\n",
-    "from sklearn import datasets\n",
-    "from sklearn.svm import SVC, LinearSVC\n",
-    "from sklearn.linear_model import SGDClassifier\n",
-    "from sklearn.preprocessing import StandardScaler\n",
-    "import matplotlib\n",
+    "import numpy as np\n",
     "import matplotlib.pyplot as plt\n",
-    "plt.rcParams['axes.labelsize'] = 14\n",
-    "plt.rcParams['xtick.labelsize'] = 12\n",
-    "plt.rcParams['ytick.labelsize'] = 12\n",
+    "from sklearn.preprocessing import PolynomialFeatures\n",
+    "from sklearn.linear_model import LinearRegression\n",
+    "\n",
+    "steps=250\n",
+    "\n",
+    "distance=0\n",
+    "x=0\n",
+    "distance_list=[]\n",
+    "steps_list=[]\n",
+    "while x<steps:\n",
+    "    distance+=np.random.randint(-1,2)\n",
+    "    distance_list.append(distance)\n",
+    "    x+=1\n",
+    "    steps_list.append(x)\n",
+    "plt.plot(steps_list,distance_list, color='green', label=\"Random Walk Data\")\n",
+    "\n",
+    "steps_list=np.asarray(steps_list)\n",
+    "distance_list=np.asarray(distance_list)\n",
+    "\n",
+    "X=steps_list[:,np.newaxis]\n",
+    "\n",
+    "#Polynomial fits\n",
+    "\n",
+    "#Degree 2\n",
+    "poly_features=PolynomialFeatures(degree=2, include_bias=False)\n",
+    "X_poly=poly_features.fit_transform(X)\n",
+    "\n",
+    "lin_reg=LinearRegression()\n",
+    "poly_fit=lin_reg.fit(X_poly,distance_list)\n",
+    "b=lin_reg.coef_\n",
+    "c=lin_reg.intercept_\n",
+    "print (\"2nd degree coefficients:\")\n",
+    "print (\"zero power: \",c)\n",
+    "print (\"first power: \", b[0])\n",
+    "print (\"second power: \",b[1])\n",
+    "\n",
+    "z = np.arange(0, steps, .01)\n",
+    "z_mod=b[1]*z**2+b[0]*z+c\n",
+    "\n",
+    "fit_mod=b[1]*X**2+b[0]*X+c\n",
+    "plt.plot(z, z_mod, color='r', label=\"2nd Degree Fit\")\n",
+    "plt.title(\"Polynomial Regression\")\n",
+    "\n",
+    "plt.xlabel(\"Steps\")\n",
+    "plt.ylabel(\"Distance\")\n",
+    "\n",
+    "#Degree 10\n",
+    "poly_features10=PolynomialFeatures(degree=10, include_bias=False)\n",
+    "X_poly10=poly_features10.fit_transform(X)\n",
+    "\n",
+    "poly_fit10=lin_reg.fit(X_poly10,distance_list)\n",
+    "\n",
+    "y_plot=poly_fit10.predict(X_poly10)\n",
+    "plt.plot(X, y_plot, color='black', label=\"10th Degree Fit\")\n",
+    "\n",
+    "plt.legend()\n",
+    "plt.show()\n",
     "\n",
     "\n",
-    "iris = datasets.load_iris()\n",
-    "X = iris[\"data\"][:, (2, 3)]  # petal length, petal width\n",
-    "y = iris[\"target\"]\n",
+    "#Decision Tree Regression\n",
+    "from sklearn.tree import DecisionTreeRegressor\n",
+    "regr_1=DecisionTreeRegressor(max_depth=2)\n",
+    "regr_2=DecisionTreeRegressor(max_depth=5)\n",
+    "regr_3=DecisionTreeRegressor(max_depth=7)\n",
+    "regr_1.fit(X, distance_list)\n",
+    "regr_2.fit(X, distance_list)\n",
+    "regr_3.fit(X, distance_list)\n",
     "\n",
-    "setosa_or_versicolor = (y == 0) | (y == 1)\n",
-    "X = X[setosa_or_versicolor]\n",
-    "y = y[setosa_or_versicolor]\n",
+    "X_test = np.arange(0.0, steps, 0.01)[:, np.newaxis]\n",
+    "y_1 = regr_1.predict(X_test)\n",
+    "y_2 = regr_2.predict(X_test)\n",
+    "y_3=regr_3.predict(X_test)\n",
     "\n",
+    "# Plot the results\n",
+    "plt.figure()\n",
+    "plt.scatter(X, distance_list, s=2.5, c=\"black\", label=\"data\")\n",
+    "plt.plot(X_test, y_1, color=\"red\",\n",
+    "         label=\"max_depth=2\", linewidth=2)\n",
+    "plt.plot(X_test, y_2, color=\"green\", label=\"max_depth=5\", linewidth=2)\n",
+    "plt.plot(X_test, y_3, color=\"m\", label=\"max_depth=7\", linewidth=2)\n",
     "\n",
-    "\n",
-    "C = 5\n",
-    "alpha = 1 / (C * len(X))\n",
-    "\n",
-    "lin_clf = LinearSVC(loss=\"hinge\", C=C, random_state=42)\n",
-    "svm_clf = SVC(kernel=\"linear\", C=C)\n",
-    "sgd_clf = SGDClassifier(loss=\"hinge\", learning_rate=\"constant\", eta0=0.001, alpha=alpha,\n",
-    "                        max_iter=100000, random_state=42)\n",
-    "\n",
-    "scaler = StandardScaler()\n",
-    "X_scaled = scaler.fit_transform(X)\n",
-    "\n",
-    "lin_clf.fit(X_scaled, y)\n",
-    "svm_clf.fit(X_scaled, y)\n",
-    "sgd_clf.fit(X_scaled, y)\n",
-    "\n",
-    "print(\"LinearSVC:                   \", lin_clf.intercept_, lin_clf.coef_)\n",
-    "print(\"SVC:                         \", svm_clf.intercept_, svm_clf.coef_)\n",
-    "print(\"SGDClassifier(alpha={:.5f}):\".format(sgd_clf.alpha), sgd_clf.intercept_, sgd_clf.coef_)\n",
-    "\n",
-    "# Compute the slope and bias of each decision boundary\n",
-    "w1 = -lin_clf.coef_[0, 0]/lin_clf.coef_[0, 1]\n",
-    "b1 = -lin_clf.intercept_[0]/lin_clf.coef_[0, 1]\n",
-    "w2 = -svm_clf.coef_[0, 0]/svm_clf.coef_[0, 1]\n",
-    "b2 = -svm_clf.intercept_[0]/svm_clf.coef_[0, 1]\n",
-    "w3 = -sgd_clf.coef_[0, 0]/sgd_clf.coef_[0, 1]\n",
-    "b3 = -sgd_clf.intercept_[0]/sgd_clf.coef_[0, 1]\n",
-    "\n",
-    "# Transform the decision boundary lines back to the original scale\n",
-    "line1 = scaler.inverse_transform([[-10, -10 * w1 + b1], [10, 10 * w1 + b1]])\n",
-    "line2 = scaler.inverse_transform([[-10, -10 * w2 + b2], [10, 10 * w2 + b2]])\n",
-    "line3 = scaler.inverse_transform([[-10, -10 * w3 + b3], [10, 10 * w3 + b3]])\n",
-    "\n",
-    "# Plot all three decision boundaries\n",
-    "plt.figure(figsize=(11, 4))\n",
-    "plt.plot(line1[:, 0], line1[:, 1], \"k:\", label=\"LinearSVC\")\n",
-    "plt.plot(line2[:, 0], line2[:, 1], \"b--\", linewidth=2, label=\"SVC\")\n",
-    "plt.plot(line3[:, 0], line3[:, 1], \"r-\", label=\"SGDClassifier\")\n",
-    "plt.plot(X[:, 0][y==1], X[:, 1][y==1], \"bs\") # label=\"Iris-Versicolor\"\n",
-    "plt.plot(X[:, 0][y==0], X[:, 1][y==0], \"yo\") # label=\"Iris-Setosa\"\n",
-    "plt.xlabel(\"Petal length\", fontsize=14)\n",
-    "plt.ylabel(\"Petal width\", fontsize=14)\n",
-    "plt.legend(loc=\"upper center\", fontsize=14)\n",
-    "plt.axis([0, 5.5, 0, 2])\n",
-    "\n",
+    "plt.xlabel(\"Data\")\n",
+    "plt.ylabel(\"Darget\")\n",
+    "plt.title(\"Decision Tree Regression\")\n",
+    "plt.legend()\n",
     "plt.show()"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "0a6d2ecd",
+   "id": "61056bcf",
    "metadata": {
     "editable": true
    },
    "source": [
-    "## What is a hyperplane?\n",
+    "## Building a tree, regression\n",
     "\n",
-    "The aim of the SVM algorithm is to find a hyperplane in a\n",
-    "$p$-dimensional space, where $p$ is the number of features that\n",
-    "distinctly classifies the data points.\n",
+    "There are mainly two steps\n",
+    "1. We split the predictor space (the set of possible values $x_1,x_2,\\dots, x_p$) into $J$ distinct and non-non-overlapping regions, $R_1,R_2,\\dots,R_J$.  \n",
     "\n",
-    "In a $p$-dimensional space, a hyperplane is what we call an affine subspace of dimension of $p-1$.\n",
-    "As an example, in two dimension, a hyperplane is simply as straight line while in three dimensions it is \n",
-    "a two-dimensional subspace, or stated simply, a plane. \n",
+    "2. For every observation that falls into the region $R_j$ , we make the same prediction, which is simply the mean of the response values for the training observations in $R_j$.\n",
     "\n",
-    "In two dimensions, with the variables $x_1$ and $x_2$, the hyperplane is defined as"
+    "How do we construct the regions $R_1,\\dots,R_J$?  In theory, the\n",
+    "regions could have any shape. However, we choose to divide the\n",
+    "predictor space into high-dimensional rectangles, or boxes, for\n",
+    "simplicity and for ease of interpretation of the resulting predictive\n",
+    "model. The goal is to find boxes $R_1,\\dots,R_J$ that minimize the\n",
+    "MSE, given by"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "dee583a6",
+   "id": "a9bb3d2c",
    "metadata": {
     "editable": true
    },
    "source": [
     "$$\n",
-    "b+w_1x_1+w_2x_2=0,\n",
+    "\\sum_{j=1}^J\\sum_{i\\in R_j}(y_i-\\overline{y}_{R_j})^2,\n",
     "$$"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "f9c04078",
+   "id": "6aa287a7",
    "metadata": {
     "editable": true
    },
    "source": [
-    "where $b$ is the intercept and $w_1$ and $w_2$ define the elements of a vector orthogonal to the line \n",
-    "$b+w_1x_1+w_2x_2=0$. \n",
-    "In two dimensions we define the vectors $\\boldsymbol{x} =[x1,x2]$ and $\\boldsymbol{w}=[w1,w2]$. \n",
-    "We can then rewrite the above equation as"
+    "where $\\overline{y}_{R_j}$  is the mean response for the training observations \n",
+    "within box $j$."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "cf463940",
+   "id": "a0290f8d",
    "metadata": {
     "editable": true
    },
    "source": [
-    "$$\n",
-    "\\boldsymbol{x}^T\\boldsymbol{w}+b=0.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "bf702f55",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## A $p$-dimensional space of features\n",
+    "## A top-down approach, recursive binary splitting\n",
     "\n",
-    "We limit ourselves to two classes of outputs $y_i$ and assign these classes the values $y_i = \\pm 1$. \n",
-    "In a $p$-dimensional space of say $p$ features we have a hyperplane defines as"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "5e53f7c7",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "b+wx_1+w_2x_2+\\dots +w_px_p=0.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "9398c2c7",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "If we define a \n",
-    "matrix $\\boldsymbol{X}=\\left[\\boldsymbol{x}_1,\\boldsymbol{x}_2,\\dots, \\boldsymbol{x}_p\\right]$\n",
-    "of dimension $n\\times p$, where $n$ represents the observations for each feature and each vector $x_i$ is a column vector of the matrix $\\boldsymbol{X}$,"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "829fae07",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\boldsymbol{x}_i = \\begin{bmatrix} x_{i1} \\\\ x_{i2} \\\\ \\dots \\\\ \\dots \\\\ x_{ip} \\end{bmatrix}.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "91e27f3a",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "If the above condition is not met for a given vector $\\boldsymbol{x}_i$ we have"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "2fb23b0b",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "b+w_1x_{i1}+w_2x_{i2}+\\dots +w_px_{ip} >0,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "8249a87a",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "if our output $y_i=1$.\n",
-    "In this case we say that $\\boldsymbol{x}_i$ lies on one of the sides of the hyperplane and if"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "15e12b59",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "b+w_1x_{i1}+w_2x_{i2}+\\dots +w_px_{ip} < 0,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "c104be97",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "for the class of observations $y_i=-1$, \n",
-    "then $\\boldsymbol{x}_i$ lies on the other side. \n",
+    "Unfortunately, it is computationally infeasible to consider every\n",
+    "possible partition of the feature space into $J$ boxes.  The common\n",
+    "strategy is to take a top-down approach\n",
     "\n",
-    "Equivalently, for the two classes of observations we have"
+    "The approach is top-down because it begins at the top of the tree (all\n",
+    "observations belong to a single region) and then successively splits\n",
+    "the predictor space; each split is indicated via two new branches\n",
+    "further down on the tree. It is greedy because at each step of the\n",
+    "tree-building process, the best split is made at that particular step,\n",
+    "rather than looking ahead and picking a split that will lead to a\n",
+    "better tree in some future step."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "4e509fbe",
+   "id": "1a243e33",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Making a tree\n",
+    "\n",
+    "In order to implement the recursive binary splitting we start by selecting\n",
+    "the predictor $x_j$ and a cutpoint $s$ that splits the predictor space into two regions $R_1$ and $R_2$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "c91c18ba",
    "metadata": {
     "editable": true
    },
    "source": [
     "$$\n",
-    "y_i\\left(b+w_1x_{i1}+w_2x_{i2}+\\dots +w_px_{ip}\\right) > 0.\n",
+    "\\left\\{X\\vert x_j < s\\right\\},\n",
     "$$"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "d11dc99a",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "When we try to separate hyperplanes, if it exists, we can use it to construct a natural classifier: a test observation is assigned a given class depending on which side of the hyperplane it is located."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "90a795e2",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## The two-dimensional case\n",
-    "\n",
-    "Let us try to develop our intuition about SVMs by limiting ourselves to a two-dimensional\n",
-    "plane.  To separate the two classes of data points, there are many\n",
-    "possible lines (hyperplanes if you prefer a more strict naming)  \n",
-    "that could be chosen. Our objective is to find a\n",
-    "plane that has the maximum margin, i.e the maximum distance between\n",
-    "data points of both classes. Maximizing the margin distance provides\n",
-    "some reinforcement so that future data points can be classified with\n",
-    "more confidence.\n",
-    "\n",
-    "What a linear classifier attempts to accomplish is to split the\n",
-    "feature space into two half spaces by placing a hyperplane between the\n",
-    "data points.  This hyperplane will be our decision boundary.  All\n",
-    "points on one side of the plane will belong to class one and all points\n",
-    "on the other side of the plane will belong to the second class two.\n",
-    "\n",
-    "Unfortunately there are many ways in which we can place a hyperplane\n",
-    "to divide the data.  Below is an example of two candidate hyperplanes\n",
-    "for our data sample."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "ab801819",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## Getting into the details\n",
-    "\n",
-    "Let us define the function"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "966f8d0f",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "f(x) = \\boldsymbol{w}^T\\boldsymbol{x}+b = 0,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "5888d347",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "as the function that determines the line $L$ that separates two classes (our two features), see the figure here. \n",
-    "\n",
-    "Any point defined by $\\boldsymbol{x}_i$ and $\\boldsymbol{x}_2$ on the line $L$ will satisfy $\\boldsymbol{w}^T(\\boldsymbol{x}_1-\\boldsymbol{x}_2)=0$. \n",
-    "\n",
-    "The signed distance $\\delta$ from any point defined by a vector $\\boldsymbol{x}$ and a point $\\boldsymbol{x}_0$ on the line $L$ is then"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "8dd10b2b",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\delta = \\frac{1}{\\vert\\vert \\boldsymbol{w}\\vert\\vert}(\\boldsymbol{w}^T\\boldsymbol{x}+b).\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "b487774a",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## First attempt at a minimization approach\n",
-    "\n",
-    "How do we find the parameter $b$ and the vector $\\boldsymbol{w}$? What we could\n",
-    "do is to define a cost function which now contains the set of all\n",
-    "misclassified points $M$ and attempt to minimize this function"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "45280475",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "C(\\boldsymbol{w},b) = -\\sum_{i\\in M} y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b).\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "870dd3f8",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "We could now for example define all values $y_i =1$ as misclassified in case we have $\\boldsymbol{w}^T\\boldsymbol{x}_i+b < 0$ and the opposite if we have $y_i=-1$. Taking the derivatives gives us"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "6d319ebb",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{\\partial C}{\\partial b} = -\\sum_{i\\in M} y_i,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "4961a152",
+   "id": "38e620f7",
    "metadata": {
     "editable": true
    },
@@ -520,1979 +415,2738 @@
   },
   {
    "cell_type": "markdown",
-   "id": "8691427e",
+   "id": "f5bad1b9",
    "metadata": {
     "editable": true
    },
    "source": [
     "$$\n",
-    "\\frac{\\partial C}{\\partial \\boldsymbol{w}} = -\\sum_{i\\in M} y_ix_i.\n",
+    "\\left\\{X\\vert x_j \\geq s\\right\\},\n",
     "$$"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "c4b43b9e",
+   "id": "7df632cb",
    "metadata": {
     "editable": true
    },
    "source": [
-    "## Solving the equations\n",
+    "so that we obtain the lowest MSE, that is"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1ab52a46",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\sum_{i:x_i\\in R_j}(y_i-\\overline{y}_{R_1})^2+\\sum_{i:x_i\\in R_2}(y_i-\\overline{y}_{R_2})^2,\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "fde2c32e",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "which we want to minimize by considering all predictors\n",
+    "$x_1,x_2,\\dots,x_p$.  We consider also all possible values of $s$ for\n",
+    "each predictor. These values could be determined by randomly assigned\n",
+    "numbers or by starting at the midpoint and then proceed till we find\n",
+    "an optimal value.\n",
     "\n",
-    "We can now use the Newton-Raphson method or different variants of the gradient descent family (from plain gradient descent to various stochastic gradient descent approaches) to solve the equations"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "bb15fc5e",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "b \\leftarrow b +\\eta \\frac{\\partial C}{\\partial b},\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "d50023cb",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "and"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "7e2f4b92",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\boldsymbol{w} \\leftarrow \\boldsymbol{w} +\\eta \\frac{\\partial C}{\\partial \\boldsymbol{w}},\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "2b04d613",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "where $\\eta$ is our by now well-known learning rate."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "c9d48cd5",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## Code Example\n",
+    "For any $j$ and $s$, we define the pair of half-planes where\n",
+    "$\\overline{y}_{R_1}$ is the mean response for the training\n",
+    "observations in $R_1(j,s)$, and $\\overline{y}_{R_2}$ is the mean\n",
+    "response for the training observations in $R_2(j,s)$.\n",
     "\n",
-    "The equations we discussed above can be coded rather easily (the\n",
-    "framework is similar to what we developed for logistic\n",
-    "regression). We are going to set up a simple case with two classes only and we want to find a line which separates them the best possible way."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "527e34d1",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## Problems with the Simpler Approach\n",
+    "Finding the values of $j$ and $s$ that minimize the above equation can be\n",
+    "done quite quickly, especially when the number of features $p$ is not\n",
+    "too large.\n",
     "\n",
-    "There are however problems with this approach, although it looks\n",
-    "pretty straightforward to implement. When running the above code, we see that we can easily end up with many diffeent lines which separate the two classes.\n",
+    "Next, we repeat the process, looking\n",
+    "for the best predictor and best cutpoint in order to split the data\n",
+    "further so as to minimize the MSE within each of the resulting\n",
+    "regions. However, this time, instead of splitting the entire predictor\n",
+    "space, we split one of the two previously identified regions. We now\n",
+    "have three regions. Again, we look to split one of these three regions\n",
+    "further, so as to minimize the MSE. The process continues until a\n",
+    "stopping criterion is reached; for instance, we may continue until no\n",
+    "region contains more than five observations."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ba37ac87",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Pruning the tree\n",
     "\n",
-    "For small\n",
-    "gaps between the entries, we may also end up needing many iterations\n",
-    "before the solutions converge and if the data cannot be separated\n",
-    "properly into two distinct classes, we may not experience a converge\n",
-    "at all."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "a6cbeccc",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## A better approach\n",
+    "The above procedure is rather straightforward, but leads often to\n",
+    "overfitting and unnecessarily large and complicated trees. The basic\n",
+    "idea is to grow a large tree $T_0$ and then prune it back in order to\n",
+    "obtain a subtree. A smaller tree with fewer splits (fewer regions) can\n",
+    "lead to smaller variance and better interpretation at the cost of a\n",
+    "little more bias.\n",
     "\n",
-    "A better approach is rather to try to define a large margin between\n",
-    "the two classes (if they are well separated from the beginning).\n",
+    "The so-called Cost complexity pruning algorithm gives us a\n",
+    "way to do just this. Rather than considering every possible subtree,\n",
+    "we consider a sequence of trees indexed by a nonnegative tuning\n",
+    "parameter $\\alpha$.\n",
     "\n",
-    "Thus, we wish to find a margin $M$ with $\\boldsymbol{w}$ normalized to\n",
-    "$\\vert\\vert \\boldsymbol{w}\\vert\\vert =1$ subject to the condition"
+    "Read more at the following [Scikit-Learn link on pruning](https://scikit-learn.org/stable/auto_examples/tree/plot_cost_complexity_pruning.html#sphx-glr-auto-examples-tree-plot-cost-complexity-pruning-py)."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "4d0ffdec",
+   "id": "fc925db6",
    "metadata": {
     "editable": true
    },
    "source": [
-    "$$\n",
-    "y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b) \\geq M \\hspace{0.1cm}\\forall i=1,2,\\dots, p.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "4bcd28aa",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "All points are thus at a signed distance from the decision boundary defined by the line $L$. The parameters $b$ and $w_1$ and $w_2$ define this line. \n",
+    "## Cost complexity pruning\n",
     "\n",
-    "We seek thus the largest value $M$ defined by"
+    "For each value of $\\alpha$  there corresponds a subtree $T \\in T_0$ such that"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "50648a73",
+   "id": "72b12d49",
    "metadata": {
     "editable": true
    },
    "source": [
     "$$\n",
-    "\\frac{1}{\\vert \\vert \\boldsymbol{w}\\vert\\vert}y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b) \\geq M \\hspace{0.1cm}\\forall i=1,2,\\dots, n,\n",
+    "\\sum_{m=1}^{\\overline{T}}\\sum_{i:x_i\\in R_m}(y_i-\\overline{y}_{R_m})^2+\\alpha\\overline{T},\n",
     "$$"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "506eb70b",
+   "id": "2579801c",
    "metadata": {
     "editable": true
    },
    "source": [
-    "or just"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "63b6b7cd",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b) \\geq M\\vert \\vert \\boldsymbol{w}\\vert\\vert \\hspace{0.1cm}\\forall i.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "7733f559",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "If we scale the equation so that $\\vert \\vert \\boldsymbol{w}\\vert\\vert = 1/M$, we have to find the minimum of \n",
-    "$\\boldsymbol{w}^T\\boldsymbol{w}=\\vert \\vert \\boldsymbol{w}\\vert\\vert$ (the norm) subject to the condition"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "9370a738",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b) \\geq 1 \\hspace{0.1cm}\\forall i.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "88a853c9",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "We have thus defined our margin as the invers of the norm of\n",
-    "$\\boldsymbol{w}$. We want to minimize the norm in order to have a as large as\n",
-    "possible margin $M$. Before we proceed, we need to remind ourselves\n",
-    "about Lagrangian multipliers."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "3df0fa2e",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## A quick Reminder on Lagrangian Multipliers\n",
+    "is as small as possible. Here $\\overline{T}$ is \n",
+    "the number of terminal nodes of the tree $T$ , $R_m$ is the\n",
+    "rectangle (i.e. the subset of predictor space)  corresponding to the $m$-th terminal node.\n",
     "\n",
-    "Consider a function of three independent variables $f(x,y,z)$ . For the function $f$ to be an\n",
-    "extreme we have"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "bf6ca034",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "df=0.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "93a081b8",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "A necessary and sufficient condition is"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "a2aec700",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{\\partial f}{\\partial x} =\\frac{\\partial f}{\\partial y}=\\frac{\\partial f}{\\partial z}=0,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "eebc071f",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "due to"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "57101c00",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "df = \\frac{\\partial f}{\\partial x}dx+\\frac{\\partial f}{\\partial y}dy+\\frac{\\partial f}{\\partial z}dz.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "8a6f6dfe",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "In many problems the variables $x,y,z$ are often subject to constraints (such as those above for the margin)\n",
-    "so that they are no longer all independent. It is possible at least in principle to use each \n",
-    "constraint to eliminate one variable\n",
-    "and to proceed with a new and smaller set of independent varables.\n",
+    "The tuning parameter $\\alpha$ controls a trade-off between the subtree’s\n",
+    "complexity and its fit to the training data. When $\\alpha = 0$, then the\n",
+    "subtree $T$ will simply equal $T_0$, \n",
+    "because then the above equation just measures the\n",
+    "training error. \n",
+    "However, as $\\alpha$ increases, there is a price to pay for\n",
+    "having a tree with many terminal nodes. The above equation will\n",
+    "tend to be minimized for a smaller subtree. \n",
     "\n",
-    "The use of so-called Lagrangian  multipliers is an alternative technique  when the elimination\n",
-    "of variables is incovenient or undesirable.  Assume that we have an equation of constraint on \n",
-    "the variables $x,y,z$"
+    "It turns out that as we increase $\\alpha$ from zero\n",
+    "branches get pruned from the tree in a nested and predictable fashion,\n",
+    "so obtaining the whole sequence of subtrees as a function of $\\alpha$ is\n",
+    "easy. We can select a value of $\\alpha$ using a validation set or using\n",
+    "cross-validation. We then return to the full data set and obtain the\n",
+    "subtree corresponding to $\\alpha$."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "44893b93",
+   "id": "9797b152",
    "metadata": {
     "editable": true
    },
    "source": [
-    "$$\n",
-    "\\phi(x,y,z) = 0,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "d162f28a",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "resulting in"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "8b112125",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "d\\phi = \\frac{\\partial \\phi}{\\partial x}dx+\\frac{\\partial \\phi}{\\partial y}dy+\\frac{\\partial \\phi}{\\partial z}dz =0.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "cfdcc304",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "Now we cannot set anymore"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "3136e10b",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{\\partial f}{\\partial x} =\\frac{\\partial f}{\\partial y}=\\frac{\\partial f}{\\partial z}=0,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "a76cda8e",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "if $df=0$ is wanted\n",
-    "because there are now only two independent variables!  Assume $x$ and $y$ are the independent \n",
-    "variables.\n",
-    "Then $dz$ is no longer arbitrary."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "024a5aeb",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## Adding the Multiplier\n",
+    "## Schematic Regression Procedure\n",
     "\n",
-    "However, we can add to"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "408e7e03",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "df = \\frac{\\partial f}{\\partial x}dx+\\frac{\\partial f}{\\partial y}dy+\\frac{\\partial f}{\\partial z}dz,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "33cb3861",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "a multiplum of $d\\phi$, viz. $\\lambda d\\phi$, resulting  in"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "f71f0d77",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "df+\\lambda d\\phi = (\\frac{\\partial f}{\\partial z}+\\lambda\n",
-    "\\frac{\\partial \\phi}{\\partial x})dx+(\\frac{\\partial f}{\\partial y}+\\lambda\\frac{\\partial \\phi}{\\partial y})dy+\n",
-    "(\\frac{\\partial f}{\\partial z}+\\lambda\\frac{\\partial \\phi}{\\partial z})dz =0.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "9c31264d",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "Our multiplier is chosen so that"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "e8466886",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{\\partial f}{\\partial z}+\\lambda\\frac{\\partial \\phi}{\\partial z} =0.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "00dc4276",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "We need to remember that we took $dx$ and $dy$ to be arbitrary and thus we must have"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "594b9dc8",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{\\partial f}{\\partial x}+\\lambda\\frac{\\partial \\phi}{\\partial x} =0,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "369c72df",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "and"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "68c9b2ce",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{\\partial f}{\\partial y}+\\lambda\\frac{\\partial \\phi}{\\partial y} =0.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "160664dd",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "When all these equations are satisfied, $df=0$.  We have four unknowns, $x,y,z$ and\n",
-    "$\\lambda$. Actually we want only $x,y,z$, $\\lambda$ needs not to be determined, \n",
-    "it is therefore often called\n",
-    "Lagrange's undetermined multiplier.\n",
-    "If we have a set of constraints $\\phi_k$ we have the equations"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "316c2f62",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{\\partial f}{\\partial x_i}+\\sum_k\\lambda_k\\frac{\\partial \\phi_k}{\\partial x_i} =0.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "4d5f006e",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## Setting up the Problem\n",
-    "In order to solve the above problem, we define the following Lagrangian function to be minimized"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "9d3be8af",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "{\\cal L}(\\lambda,b,\\boldsymbol{w})=\\frac{1}{2}\\boldsymbol{w}^T\\boldsymbol{w}-\\sum_{i=1}^n\\lambda_i\\left[y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b)-1\\right],\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "bfdd0ba3",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "where $\\lambda_i$ is a so-called Lagrange multiplier subject to the condition $\\lambda_i \\geq 0$.\n",
+    "**Building a Regression Tree.**\n",
     "\n",
-    "Taking the derivatives  with respect to $b$ and $\\boldsymbol{w}$ we obtain"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "26601b00",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{\\partial {\\cal L}}{\\partial b} = -\\sum_{i} \\lambda_iy_i=0,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "2a64189c",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "and"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "1779ea37",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{\\partial {\\cal L}}{\\partial \\boldsymbol{w}} = 0 = \\boldsymbol{w}-\\sum_{i} \\lambda_iy_i\\boldsymbol{x}_i.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "37adea5b",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "Inserting these constraints into the equation for ${\\cal L}$ we obtain"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "98bd44bd",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "{\\cal L}=\\sum_i\\lambda_i-\\frac{1}{2}\\sum_{ij}^n\\lambda_i\\lambda_jy_iy_j\\boldsymbol{x}_i^T\\boldsymbol{x}_j,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "1aeb7c4c",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "subject to the constraints $\\lambda_i\\geq 0$ and $\\sum_i\\lambda_iy_i=0$. \n",
-    "We must in addition satisfy the [Karush-Kuhn-Tucker](https://en.wikipedia.org/wiki/Karush%E2%80%93Kuhn%E2%80%93Tucker_conditions) (KKT) condition"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "93489d2f",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\lambda_i\\left[y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b) -1\\right] \\hspace{0.1cm}\\forall i.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "33e0ca40",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "1. If $\\lambda_i > 0$, then $y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b)=1$ and we say that $x_i$ is on the boundary.\n",
+    "1. Use recursive binary splitting to grow a large tree on the training data, stopping only when each terminal node has fewer than some minimum number of observations.\n",
     "\n",
-    "2. If $y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b)> 1$, we say $x_i$ is not on the boundary and we set $\\lambda_i=0$. \n",
+    "2. Apply cost complexity pruning to the large tree in order to obtain a sequence of best subtrees, as a function of $\\alpha$.\n",
     "\n",
-    "When $\\lambda_i > 0$, the vectors $\\boldsymbol{x}_i$ are called support vectors. They are the vectors closest to the line (or hyperplane) and define the margin $M$."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "ff032710",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## The problem to solve\n",
+    "3. Use for example $K$-fold cross-validation to choose $\\alpha$. Divide the training observations into $K$ folds. For each $k=1,2,\\dots,K$ we: \n",
     "\n",
-    "We can rewrite"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "04764680",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "{\\cal L}=\\sum_i\\lambda_i-\\frac{1}{2}\\sum_{ij}^n\\lambda_i\\lambda_jy_iy_j\\boldsymbol{x}_i^T\\boldsymbol{x}_j,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "5d03db44",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "and its constraints in terms of a matrix-vector problem where we minimize w.r.t. $\\lambda$ the following problem"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "2f357081",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{1}{2} \\boldsymbol{\\lambda}^T\\begin{bmatrix} y_1y_1\\boldsymbol{x}_1^T\\boldsymbol{x}_1 & y_1y_2\\boldsymbol{x}_1^T\\boldsymbol{x}_2 & \\dots & \\dots & y_1y_n\\boldsymbol{x}_1^T\\boldsymbol{x}_n \\\\\n",
-    "y_2y_1\\boldsymbol{x}_2^T\\boldsymbol{x}_1 & y_2y_2\\boldsymbol{x}_2^T\\boldsymbol{x}_2 & \\dots & \\dots & y_1y_n\\boldsymbol{x}_2^T\\boldsymbol{x}_n \\\\\n",
-    "\\dots & \\dots & \\dots & \\dots & \\dots \\\\\n",
-    "\\dots & \\dots & \\dots & \\dots & \\dots \\\\\n",
-    "y_ny_1\\boldsymbol{x}_n^T\\boldsymbol{x}_1 & y_ny_2\\boldsymbol{x}_n^T\\boldsymbol{x}_2 & \\dots & \\dots & y_ny_n\\boldsymbol{x}_n^T\\boldsymbol{x}_n \\\\\n",
-    "\\end{bmatrix}\\boldsymbol{\\lambda}-\\mathbb{1}\\boldsymbol{\\lambda},\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "0c2a6d5f",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "subject to $\\boldsymbol{y}^T\\boldsymbol{\\lambda}=0$. Here we defined the vectors $\\boldsymbol{\\lambda} =[\\lambda_1,\\lambda_2,\\dots,\\lambda_n]$ and \n",
-    "$\\boldsymbol{y}=[y_1,y_2,\\dots,y_n]$."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "21ac0cad",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## The last steps\n",
+    "  * repeat steps 1 and 2 on all but the $k$-th fold of the training data. \n",
     "\n",
-    "Solving the above problem, yields the values of $\\lambda_i$.\n",
-    "To find the coefficients of your hyperplane we need simply to compute"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "9a2eeb8e",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\boldsymbol{w}=\\sum_{i} \\lambda_iy_i\\boldsymbol{x}_i.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "e8380a07",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "With our vector $\\boldsymbol{w}$ we can in turn find the value of the intercept $b$ (here in two dimensions) via"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "cb30be9a",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b)=1,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "8ed52094",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "resulting in"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "6997d3d5",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "b = \\frac{1}{y_i}-\\boldsymbol{w}^T\\boldsymbol{x}_i,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "fa46d07c",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "or if we write it out in terms of the support vectors only, with $N_s$ being their number,  we have"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "219fad4c",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "b = \\frac{1}{N_s}\\sum_{j\\in N_s}\\left(y_j-\\sum_{i=1}^n\\lambda_iy_i\\boldsymbol{x}_i^T\\boldsymbol{x}_j\\right).\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "d515c510",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "With our hyperplane coefficients we can use our classifier to assign any observation by simply using"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "6e4dea54",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "y_i = \\mathrm{sign}(\\boldsymbol{w}^T\\boldsymbol{x}_i+b).\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "72114d67",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "Below we discuss how to find the optimal values of $\\lambda_i$. Before we proceed however, we discuss now the so-called soft classifier."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "d2d64dfe",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## A soft classifier\n",
+    "  * Then we valuate the mean squared prediction error on the data in the left-out $k$-th fold, as a function of $\\alpha$.\n",
     "\n",
-    "Till now, the margin is strictly defined by the support vectors. This defines what is called a hard classifier, that is the margins are well defined.\n",
+    "  * Finally  we average the results for each value of $\\alpha$, and pick $\\alpha$ to minimize the average error.\n",
     "\n",
-    "Suppose now that classes overlap in feature space, as shown in the\n",
-    "figure here. One way to deal with this problem before we define the\n",
-    "so-called **kernel approach**, is to allow a kind of slack in the sense\n",
-    "that we allow some points to be on the wrong side of the margin.\n",
+    "4. Return the subtree from Step 2 that corresponds to the chosen value of $\\alpha$."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4836e23f",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## A Classification Tree\n",
     "\n",
-    "We introduce thus the so-called **slack** variables $\\boldsymbol{\\xi} =[\\xi_1,x_2,\\dots,x_n]$ and \n",
-    "modify our previous equation"
+    "A classification tree is very similar to a regression tree, except\n",
+    "that it is used to predict a qualitative response rather than a\n",
+    "quantitative one. Recall that for a regression tree, the predicted\n",
+    "response for an observation is given by the mean response of the\n",
+    "training observations that belong to the same terminal node. In\n",
+    "contrast, for a classification tree, we predict that each observation\n",
+    "belongs to the most commonly occurring class of training observations\n",
+    "in the region to which it belongs. In interpreting the results of a\n",
+    "classification tree, we are often interested not only in the class\n",
+    "prediction corresponding to a particular terminal node region, but\n",
+    "also in the class proportions among the training observations that\n",
+    "fall into that region."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "f5ea43a7",
+   "id": "2e6f6a44",
    "metadata": {
     "editable": true
    },
    "source": [
-    "$$\n",
-    "y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b)=1,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "c32d613f",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "to"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "fb2d79c8",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b)=1-\\xi_i,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "4f350a35",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "with the requirement $\\xi_i\\geq 0$. The total violation is now $\\sum_i\\xi$. \n",
-    "The value $\\xi_i$ in the constraint the last constraint corresponds to the  amount by which the prediction\n",
-    "$y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b)=1$ is on the wrong side of its margin. Hence by bounding the sum $\\sum_i \\xi_i$,\n",
-    "we bound the total amount by which predictions fall on the wrong side of their margins.\n",
+    "## Growing a classification tree\n",
     "\n",
-    "Misclassifications occur when $\\xi_i > 1$. Thus bounding the total sum by some value $C$ bounds in turn the total number of\n",
-    "misclassifications."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "a7152381",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## Soft optmization problem\n",
+    "The task of growing a\n",
+    "classification tree is quite similar to the task of growing a\n",
+    "regression tree. Just as in the regression setting, we use recursive\n",
+    "binary splitting to grow a classification tree. However, in the\n",
+    "classification setting, the MSE cannot be used as a criterion for making\n",
+    "the binary splits.  A natural alternative to MSE is the **classification\n",
+    "error rate**. Since we plan to assign an observation in a given region\n",
+    "to the most commonly occurring error rate class of training\n",
+    "observations in that region, the classification error rate is simply\n",
+    "the fraction of the training observations in that region that do not\n",
+    "belong to the most common class. \n",
     "\n",
-    "This has in turn the consequences that we change our optmization problem to finding the minimum of"
+    "When building a classification tree, either the Gini index or the\n",
+    "entropy are typically used to evaluate the quality of a particular\n",
+    "split, since these two approaches are more sensitive to node purity\n",
+    "than is the classification error rate."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "6be2b02f",
+   "id": "78fca099",
    "metadata": {
     "editable": true
    },
    "source": [
-    "$$\n",
-    "{\\cal L}=\\frac{1}{2}\\boldsymbol{w}^T\\boldsymbol{w}-\\sum_{i=1}^n\\lambda_i\\left[y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b)-(1-\\xi_)\\right]+C\\sum_{i=1}^n\\xi_i-\\sum_{i=1}^n\\gamma_i\\xi_i,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "66220800",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "subject to"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "ed6c6750",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b)=1-\\xi_i \\hspace{0.1cm}\\forall i,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "48d4f365",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "with the requirement $\\xi_i\\geq 0$.\n",
+    "## Classification tree, how to split nodes\n",
     "\n",
-    "Taking the derivatives  with respect to $b$ and $\\boldsymbol{w}$ we obtain"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "89cf3c1a",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{\\partial {\\cal L}}{\\partial b} = -\\sum_{i} \\lambda_iy_i=0,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "246e9732",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "and"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "bb28975d",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{\\partial {\\cal L}}{\\partial \\boldsymbol{w}} = 0 = \\boldsymbol{w}-\\sum_{i} \\lambda_iy_i\\boldsymbol{x}_i,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "d026242c",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "and"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "edad3470",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\lambda_i = C-\\gamma_i \\hspace{0.1cm}\\forall i.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "8d9ae4ab",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "Inserting these constraints into the equation for ${\\cal L}$ we obtain the same equation as before"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "3cabd2a3",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "{\\cal L}=\\sum_i\\lambda_i-\\frac{1}{2}\\sum_{ij}^n\\lambda_i\\lambda_jy_iy_j\\boldsymbol{x}_i^T\\boldsymbol{x}_j,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "af0b3955",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "but now subject to the constraints $\\lambda_i\\geq 0$, $\\sum_i\\lambda_iy_i=0$ and $0\\leq\\lambda_i \\leq C$. \n",
-    "We must in addition satisfy the Karush-Kuhn-Tucker condition which now reads"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "efce4708",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\lambda_i\\left[y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b) -(1-\\xi_)\\right]=0 \\hspace{0.1cm}\\forall i,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "91b8e3a8",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\gamma_i\\xi_i = 0,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "1b07fe63",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "and"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "e3e5c2e5",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "y_i(\\boldsymbol{w}^T\\boldsymbol{x}_i+b) -(1-\\xi_) \\geq 0 \\hspace{0.1cm}\\forall i.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "a4958010",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## Kernels and non-linearity\n",
+    "If our targets are the outcome of a classification process that takes\n",
+    "for example $k=1,2,\\dots,K$ values, the only thing we need to think of\n",
+    "is to set up the splitting criteria for each node.\n",
     "\n",
-    "The cases we have studied till now, were all characterized by two classes\n",
-    "with a close to linear separability. The classifiers we have described\n",
-    "so far find linear boundaries in our input feature space. It is\n",
-    "possible to make our procedure more flexible by exploring the feature\n",
-    "space using other basis expansions such as higher-order polynomials,\n",
-    "wavelets, splines etc.\n",
+    "We define a PDF $p_{mk}$ that represents the number of observations of\n",
+    "a class $k$ in a region $R_m$ with $N_m$ observations. We represent\n",
+    "this likelihood function in terms of the proportion $I(y_i=k)$ of\n",
+    "observations of this class in the region $R_m$ as"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a85542fc",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "p_{mk} = \\frac{1}{N_m}\\sum_{x_i\\in R_m}I(y_i=k).\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d71ed79a",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "We let $p_{mk}$ represent the majority class of observations in region\n",
+    "$m$. The three most common ways of splitting a node are given by\n",
     "\n",
-    "If our feature space is not easy to separate, as shown in the figure\n",
-    "here, we can achieve a better separation by introducing more complex\n",
-    "basis functions. The ideal would be, as shown in the next figure, to, via a specific transformation to \n",
-    "obtain a separation between the classes which is almost linear. \n",
-    "\n",
-    "The change of basis, from $x\\rightarrow z=\\phi(x)$ leads to the same type of equations to be solved, except that\n",
-    "we need to introduce for example a polynomial transformation to a two-dimensional training set."
+    "* Misclassification error"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f940a8a5",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "p_{mk} = \\frac{1}{N_m}\\sum_{x_i\\in R_m}I(y_i\\ne k) = 1-p_{mk}.\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e514b953",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "* Gini index $g$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "94ee340e",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "g = \\sum_{k=1}^K p_{mk}(1-p_{mk}).\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5258611b",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "* Information entropy or just entropy $s$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "92aff9a1",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "s = -\\sum_{k=1}^K p_{mk}\\log{p_{mk}}.\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3878c118",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Visualizing the Tree, Classification"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": 2,
-   "id": "865d0e1e",
+   "id": "316a9e84",
    "metadata": {
     "collapsed": false,
     "editable": true
    },
    "outputs": [],
    "source": [
-    "import numpy as np\n",
     "import os\n",
+    "from sklearn.datasets import load_breast_cancer\n",
+    "from sklearn.tree import DecisionTreeClassifier\n",
+    "from sklearn.model_selection import train_test_split\n",
+    "from sklearn.metrics import confusion_matrix\n",
+    "from sklearn.tree import export_graphviz\n",
     "\n",
-    "np.random.seed(42)\n",
-    "\n",
-    "# To plot pretty figures\n",
-    "import matplotlib\n",
-    "import matplotlib.pyplot as plt\n",
-    "plt.rcParams['axes.labelsize'] = 14\n",
-    "plt.rcParams['xtick.labelsize'] = 12\n",
-    "plt.rcParams['ytick.labelsize'] = 12\n",
+    "from IPython.display import Image \n",
+    "from pydot import graph_from_dot_data\n",
+    "import pandas as pd\n",
+    "import numpy as np\n",
     "\n",
     "\n",
-    "from sklearn.svm import SVC\n",
-    "from sklearn import datasets\n",
+    "cancer = load_breast_cancer()\n",
+    "X = pd.DataFrame(cancer.data, columns=cancer.feature_names)\n",
+    "print(X)\n",
+    "y = pd.Categorical.from_codes(cancer.target, cancer.target_names)\n",
+    "y = pd.get_dummies(y)\n",
+    "print(y)\n",
+    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=1)\n",
+    "tree_clf = DecisionTreeClassifier(max_depth=5)\n",
+    "tree_clf.fit(X_train, y_train)\n",
     "\n",
-    "\n",
-    "\n",
-    "X1D = np.linspace(-4, 4, 9).reshape(-1, 1)\n",
-    "X2D = np.c_[X1D, X1D**2]\n",
-    "y = np.array([0, 0, 1, 1, 1, 1, 1, 0, 0])\n",
-    "\n",
-    "plt.figure(figsize=(11, 4))\n",
-    "\n",
-    "plt.subplot(121)\n",
-    "plt.grid(True, which='both')\n",
-    "plt.axhline(y=0, color='k')\n",
-    "plt.plot(X1D[:, 0][y==0], np.zeros(4), \"bs\")\n",
-    "plt.plot(X1D[:, 0][y==1], np.zeros(5), \"g^\")\n",
-    "plt.gca().get_yaxis().set_ticks([])\n",
-    "plt.xlabel(r\"$x_1$\", fontsize=20)\n",
-    "plt.axis([-4.5, 4.5, -0.2, 0.2])\n",
-    "\n",
-    "plt.subplot(122)\n",
-    "plt.grid(True, which='both')\n",
-    "plt.axhline(y=0, color='k')\n",
-    "plt.axvline(x=0, color='k')\n",
-    "plt.plot(X2D[:, 0][y==0], X2D[:, 1][y==0], \"bs\")\n",
-    "plt.plot(X2D[:, 0][y==1], X2D[:, 1][y==1], \"g^\")\n",
-    "plt.xlabel(r\"$x_1$\", fontsize=20)\n",
-    "plt.ylabel(r\"$x_2$\", fontsize=20, rotation=0)\n",
-    "plt.gca().get_yaxis().set_ticks([0, 4, 8, 12, 16])\n",
-    "plt.plot([-4.5, 4.5], [6.5, 6.5], \"r--\", linewidth=3)\n",
-    "plt.axis([-4.5, 4.5, -1, 17])\n",
-    "plt.subplots_adjust(right=1)\n",
-    "plt.show()"
+    "export_graphviz(\n",
+    "    tree_clf,\n",
+    "    out_file=\"DataFiles/cancer.dot\",\n",
+    "    feature_names=cancer.feature_names,\n",
+    "    class_names=cancer.target_names,\n",
+    "    rounded=True,\n",
+    "    filled=True\n",
+    ")\n",
+    "cmd = 'dot -Tpng DataFiles/cancer.dot -o DataFiles/cancer.png'\n",
+    "os.system(cmd)"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "d94f8f9d",
+   "id": "d9d502d0",
    "metadata": {
     "editable": true
    },
    "source": [
-    "## The equations\n",
-    "\n",
-    "Suppose we define a polynomial transformation of degree two only (we continue to live in a plane with $x_i$ and $y_i$ as variables)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "4cc02738",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "z = \\phi(x_i) =\\left(x_i^2, y_i^2, \\sqrt{2}x_iy_i\\right).\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "461a4aa7",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "With our new basis, the equations we solved earlier are basically the same, that is we have now (without the slack option for simplicity)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "4650491e",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "{\\cal L}=\\sum_i\\lambda_i-\\frac{1}{2}\\sum_{ij}^n\\lambda_i\\lambda_jy_iy_j\\boldsymbol{z}_i^T\\boldsymbol{z}_j,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "4a9c2c90",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "subject to the constraints $\\lambda_i\\geq 0$, $\\sum_i\\lambda_iy_i=0$, and for the support vectors"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "436211fc",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "y_i(\\boldsymbol{w}^T\\boldsymbol{z}_i+b)= 1 \\hspace{0.1cm}\\forall i,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "87dc972a",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "from which we also find $b$.\n",
-    "To compute $\\boldsymbol{z}_i^T\\boldsymbol{z}_j$ we define the kernel $K(\\boldsymbol{x}_i,\\boldsymbol{x}_j)$ as"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "c787df3e",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "K(\\boldsymbol{x}_i,\\boldsymbol{x}_j)=\\boldsymbol{z}_i^T\\boldsymbol{z}_j= \\phi(\\boldsymbol{x}_i)^T\\phi(\\boldsymbol{x}_j).\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "011c7927",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "For the above example, the kernel reads"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "fd780bfa",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "K(\\boldsymbol{x}_i,\\boldsymbol{x}_j)=[x_i^2, y_i^2, \\sqrt{2}x_iy_i]^T\\begin{bmatrix} x_j^2 \\\\ y_j^2 \\\\ \\sqrt{2}x_jy_j \\end{bmatrix}=x_i^2x_j^2+2x_ix_jy_iy_j+y_i^2y_j^2.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "65043810",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "We note that this is nothing but the dot product of the two original\n",
-    "vectors $(\\boldsymbol{x}_i^T\\boldsymbol{x}_j)^2$. Instead of thus computing the\n",
-    "product in the Lagrangian of $\\boldsymbol{z}_i^T\\boldsymbol{z}_j$ we simply compute\n",
-    "the dot product $(\\boldsymbol{x}_i^T\\boldsymbol{x}_j)^2$.\n",
-    "\n",
-    "This leads to the so-called\n",
-    "kernel trick and the result leads to the same as if we went through\n",
-    "the trouble of performing the transformation\n",
-    "$\\phi(\\boldsymbol{x}_i)^T\\phi(\\boldsymbol{x}_j)$ during the SVM calculations."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "61ba4e3a",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## The problem to solve\n",
-    "Using our definition of the kernel We can rewrite again the Lagrangian"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "57a6689f",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "{\\cal L}=\\sum_i\\lambda_i-\\frac{1}{2}\\sum_{ij}^n\\lambda_i\\lambda_jy_iy_j\\boldsymbol{x}_i^T\\boldsymbol{z}_j,\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "3ecde8d7",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "subject to the constraints $\\lambda_i\\geq 0$, $\\sum_i\\lambda_iy_i=0$ in terms of a convex optimization problem"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "c61046b9",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{1}{2} \\boldsymbol{\\lambda}^T\\begin{bmatrix} y_1y_1K(\\boldsymbol{x}_1,\\boldsymbol{x}_1) & y_1y_2K(\\boldsymbol{x}_1,\\boldsymbol{x}_2) & \\dots & \\dots & y_1y_nK(\\boldsymbol{x}_1,\\boldsymbol{x}_n) \\\\\n",
-    "y_2y_1K(\\boldsymbol{x}_2,\\boldsymbol{x}_1) & y_2y_2(\\boldsymbol{x}_2,\\boldsymbol{x}_2) & \\dots & \\dots & y_1y_nK(\\boldsymbol{x}_2,\\boldsymbol{x}_n) \\\\\n",
-    "\\dots & \\dots & \\dots & \\dots & \\dots \\\\\n",
-    "\\dots & \\dots & \\dots & \\dots & \\dots \\\\\n",
-    "y_ny_1K(\\boldsymbol{x}_n,\\boldsymbol{x}_1) & y_ny_2K(\\boldsymbol{x}_n\\boldsymbol{x}_2) & \\dots & \\dots & y_ny_nK(\\boldsymbol{x}_n,\\boldsymbol{x}_n) \\\\\n",
-    "\\end{bmatrix}\\boldsymbol{\\lambda}-\\mathbb{1}\\boldsymbol{\\lambda},\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "578a400b",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "subject to $\\boldsymbol{y}^T\\boldsymbol{\\lambda}=0$. Here we defined the vectors $\\boldsymbol{\\lambda} =[\\lambda_1,\\lambda_2,\\dots,\\lambda_n]$ and \n",
-    "$\\boldsymbol{y}=[y_1,y_2,\\dots,y_n]$. \n",
-    "If we add the slack constants this leads to the additional constraint $0\\leq \\lambda_i \\leq C$.\n",
-    "\n",
-    "We can rewrite this (see the solutions below) in terms of a convex optimization problem of the type"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "1a4cfb3f",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\begin{align*}\n",
-    "    &\\mathrm{min}_{\\lambda}\\hspace{0.2cm} \\frac{1}{2}\\boldsymbol{\\lambda}^T\\boldsymbol{P}\\boldsymbol{\\lambda}+\\boldsymbol{q}^T\\boldsymbol{\\lambda},\\\\ \\nonumber\n",
-    "    &\\mathrm{subject\\hspace{0.1cm}to} \\hspace{0.2cm} \\boldsymbol{G}\\boldsymbol{\\lambda} \\preceq \\boldsymbol{h} \\hspace{0.2cm} \\wedge \\boldsymbol{A}\\boldsymbol{\\lambda}=f.\n",
-    "\\end{align*}\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "453d00e0",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "Below we discuss how to solve these equations. Here we note that the matrix $\\boldsymbol{P}$ has matrix elements $p_{ij}=y_iy_jK(\\boldsymbol{x}_i,\\boldsymbol{x}_j)$.\n",
-    "Given a kernel $K$ and the targets $y_i$ this matrix is easy to set up. The constraint $\\boldsymbol{y}^T\\boldsymbol{\\lambda}=0$ leads to $f=0$ and $\\boldsymbol{A}=\\boldsymbol{y}$. How to set up the matrix $\\boldsymbol{G}$ is discussed later. Here note that the inequalities $0\\leq \\lambda_i \\leq C$ can be split up into\n",
-    "$0\\leq \\lambda_i$ and $\\lambda_i \\leq C$. These two inequalities define then the matrix $\\boldsymbol{G}$ and the vector $\\boldsymbol{h}$."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "9f1255ac",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## Different kernels and Mercer's theorem\n",
-    "\n",
-    "There are several popular kernels being used. These are\n",
-    "1. Linear: $K(\\boldsymbol{x},\\boldsymbol{y})=\\boldsymbol{x}^T\\boldsymbol{y}$,\n",
-    "\n",
-    "2. Polynomial: $K(\\boldsymbol{x},\\boldsymbol{y})=(\\boldsymbol{x}^T\\boldsymbol{y}+\\gamma)^d$,\n",
-    "\n",
-    "3. Gaussian Radial Basis Function: $K(\\boldsymbol{x},\\boldsymbol{y})=\\exp{\\left(-\\gamma\\vert\\vert\\boldsymbol{x}-\\boldsymbol{y}\\vert\\vert^2\\right)}$,\n",
-    "\n",
-    "4. Tanh: $K(\\boldsymbol{x},\\boldsymbol{y})=\\tanh{(\\boldsymbol{x}^T\\boldsymbol{y}+\\gamma)}$,\n",
-    "\n",
-    "and many other ones.\n",
-    "\n",
-    "An important theorem for us is [Mercer's\n",
-    "theorem](https://en.wikipedia.org/wiki/Mercer%27s_theorem).  The\n",
-    "theorem states that if a kernel function $K$ is symmetric, continuous\n",
-    "and leads to a positive semi-definite matrix $\\boldsymbol{P}$ then there\n",
-    "exists a function $\\phi$ that maps $\\boldsymbol{x}_i$ and $\\boldsymbol{x}_j$ into\n",
-    "another space (possibly with much higher dimensions) such that"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "72c1b4c8",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "K(\\boldsymbol{x}_i,\\boldsymbol{x}_j)=\\phi(\\boldsymbol{x}_i)^T\\phi(\\boldsymbol{x}_j).\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "4f81ee5a",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "So you can use $K$ as a kernel since you know $\\phi$ exists, even if\n",
-    "you don’t know what $\\phi$ is. \n",
-    "\n",
-    "Note that some frequently used kernels (such as the Sigmoid kernel)\n",
-    "don’t respect all of Mercer’s conditions, yet they generally work well\n",
-    "in practice."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "0f53f853",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## The moons example"
+    "## Visualizing the Tree, The Moons"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": 3,
-   "id": "229d93fe",
+   "id": "5682ec34",
    "metadata": {
     "collapsed": false,
     "editable": true
    },
    "outputs": [],
    "source": [
-    "from __future__ import division, print_function, unicode_literals\n",
-    "\n",
+    "# Common imports\n",
     "import numpy as np\n",
+    "from sklearn.model_selection import  train_test_split \n",
+    "from sklearn.tree import DecisionTreeClassifier\n",
+    "from sklearn.datasets import make_moons\n",
+    "from sklearn.tree import export_graphviz\n",
+    "from pydot import graph_from_dot_data\n",
+    "import pandas as pd\n",
+    "import os\n",
+    "\n",
     "np.random.seed(42)\n",
+    "X, y = make_moons(n_samples=100, noise=0.25, random_state=53)\n",
+    "X_train, X_test, y_train, y_test = train_test_split(X,y,random_state=0)\n",
+    "tree_clf = DecisionTreeClassifier(max_depth=5)\n",
+    "tree_clf.fit(X_train, y_train)\n",
     "\n",
-    "import matplotlib\n",
+    "export_graphviz(\n",
+    "    tree_clf,\n",
+    "    out_file=\"DataFiles/moons.dot\",\n",
+    "    rounded=True,\n",
+    "    filled=True\n",
+    ")\n",
+    "cmd = 'dot -Tpng DataFiles/moons.dot -o DataFiles/moons.png'\n",
+    "os.system(cmd)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "c9058975",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Other ways of visualizing the trees\n",
+    "\n",
+    "**Scikit-Learn** has also another way to visualize the trees which is very useful, here with the Iris data."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "0a31a1e8",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "from sklearn.datasets import load_iris\n",
+    "from sklearn import tree\n",
+    "X, y = load_iris(return_X_y=True)\n",
+    "tree_clf = tree.DecisionTreeClassifier()\n",
+    "tree_clf = tree_clf.fit(X, y)\n",
+    "# and then plot the tree\n",
+    "tree.plot_tree(tree_clf)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "c4268d2b",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Printing out as text\n",
+    "\n",
+    "Alternatively, the tree can also be exported in textual format with the function exporttext.\n",
+    "This method doesn’t require the installation of external libraries and is more compact:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "28c43d52",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "from sklearn.datasets import load_iris\n",
+    "from sklearn.tree import DecisionTreeClassifier\n",
+    "from sklearn.tree import export_text\n",
+    "iris = load_iris()\n",
+    "decision_tree = DecisionTreeClassifier(random_state=0, max_depth=2)\n",
+    "decision_tree = decision_tree.fit(iris.data, iris.target)\n",
+    "r = export_text(decision_tree, feature_names=iris['feature_names'])\n",
+    "print(r)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ba7bcbc2",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Algorithms for Setting up Decision Trees\n",
+    "\n",
+    "Two algorithms stand out in the set up of decision trees:\n",
+    "1. The CART (Classification And Regression Tree) algorithm for both classification and regression\n",
+    "\n",
+    "2. The ID3 algorithm based on the computation of the information gain for classification\n",
+    "\n",
+    "We discuss both algorithms with applications here. The popular library\n",
+    "**Scikit-Learn** uses the CART algorithm. For classification problems\n",
+    "you can use either the **gini** index or the **entropy** to split a tree\n",
+    "in two branches."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "bbeb3a76",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## The CART algorithm for Classification\n",
+    "\n",
+    "For classification, the CART algorithm splits the data set in two subsets using a single feature $k$ and a threshold $t_k$.\n",
+    "This could be for example a threshold set by a number below a certain circumference of a malign tumor.\n",
+    "\n",
+    "How do we find these two quantities?\n",
+    "We search for the pair $(k,t_k)$ that produces the purest subset using for example the **gini** factor $G$.\n",
+    "The cost function it tries to minimize is then"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "087645ff",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "C(k,t_k) = \\frac{m_{\\mathrm{left}}}{m}G_{\\mathrm{left}}+ \\frac{m_{\\mathrm{right}}}{m}G_{\\mathrm{right}},\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "0bfe907e",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "where $G_{\\mathrm{left/right}}$ measures the impurity of the left/right subset  and $m_{\\mathrm{left/right}}$\n",
+    " is the number of instances in the left/right subset\n",
+    "\n",
+    "Once it has successfully split the training set in two, it splits the subsets using the same logic, then the subsubsets\n",
+    "and so on, recursively. It stops recursing once it reaches the maximum depth (defined by the\n",
+    "$max\\_depth$ hyperparameter), or if it cannot find a split that will reduce impurity. A few other\n",
+    "hyperparameters control additional stopping conditions such as the $min\\_samples\\_split$,\n",
+    "$min\\_samples\\_leaf$, $min\\_weight\\_fraction\\_leaf$, and $max\\_leaf\\_nodes$."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ed3a24c1",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## The CART algorithm for Regression\n",
+    "\n",
+    "The CART algorithm for regression works is similar to the one for classification except that instead of trying to split the\n",
+    "training set in a way that minimizes say the **gini** or **entropy** impurity, it now tries to split the training set in a way that minimizes our well-known mean-squared error (MSE). The cost function is now"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6d5fa1b2",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "C(k,t_k) = \\frac{m_{\\mathrm{left}}}{m}\\mathrm{MSE}_{\\mathrm{left}}+ \\frac{m_{\\mathrm{right}}}{m}\\mathrm{MSE}_{\\mathrm{right}}.\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ec7a9513",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "Here the MSE for a specific node is defined as"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a20a7523",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\mathrm{MSE}_{\\mathrm{node}}=\\frac{1}{m_\\mathrm{node}}\\sum_{i\\in \\mathrm{node}}(\\overline{y}_{\\mathrm{node}}-y_i)^2,\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ec4a4cab",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "with"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "0013fe85",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\overline{y}_{\\mathrm{node}}=\\frac{1}{m_\\mathrm{node}}\\sum_{i\\in \\mathrm{node}}y_i,\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3ca1f119",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "the mean value of all observations in a specific node.\n",
+    "\n",
+    "Without any regularization, the regression task for decision trees, \n",
+    "just like for classification tasks, is  prone to overfitting."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2053ceb1",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Why binary splits?\n",
+    "\n",
+    "It is custom to split to a tree uising binary splits. The reason is\n",
+    "that multiway splits fragment the data too quickly, leaving\n",
+    "insufficient data at the next level down.  Multiway splits can be\n",
+    "achieved by a series of binary split and this is normally preferred."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "12e5e36e",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Computing a Tree using the Gini Index\n",
+    "\n",
+    "Consider the following example with attributes/features and two\n",
+    "possible outcomes (classes)  for each attribute.  Assume we wish to find some\n",
+    "correlations between the average grade of a student as function of the\n",
+    "number of hours studied and hours slept. We want also to correlate the\n",
+    "grade in a given course with the general trend, whether the students\n",
+    "recently has gotten grades below average or above.\n",
+    "\n",
+    "We have three features/attributes\n",
+    "1. Trend of average grades before present course, classified as either below  or above  the average grade of the whole class  \n",
+    "\n",
+    "2. The number of hours studies, classified again as either higher (more than 3 hours per day) or lower . Here we have used a standard for one $ECTS$ which is scaled to 25-30 hours of work for a semester which lasts 18 weeks, with 15 weeks of lectures and 3 weeks for exams, assuming a total of 30 ECTS per semester. \n",
+    "\n",
+    "3. The number of hours slept as high for more than $8$ hours and below  for less than 8 hours of sleep, classified again as either high or low\n",
+    "\n",
+    "4. The final grade whether it is above or below average"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "44224dcb",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## The Table\n",
+    "\n",
+    "<table class=\"dotable\" border=\"1\">\n",
+    "<thead>\n",
+    "<tr><th align=\"center\">Grade Trend</th> <th align=\"center\">Hours slept</th> <th align=\"center\">Hours Studied</th> <th align=\"center\">Grade</th> </tr>\n",
+    "</thead>\n",
+    "<tbody>\n",
+    "<tr><td align=\"center\">   Above          </td> <td align=\"center\">   Low            </td> <td align=\"center\">   High             </td> <td align=\"center\">   Above    </td> </tr>\n",
+    "<tr><td align=\"center\">   Below          </td> <td align=\"center\">   High           </td> <td align=\"center\">   Low              </td> <td align=\"center\">   Below    </td> </tr>\n",
+    "<tr><td align=\"center\">   Above          </td> <td align=\"center\">   Low            </td> <td align=\"center\">   High             </td> <td align=\"center\">   Above    </td> </tr>\n",
+    "<tr><td align=\"center\">   Above          </td> <td align=\"center\">   High           </td> <td align=\"center\">   High             </td> <td align=\"center\">   Above    </td> </tr>\n",
+    "<tr><td align=\"center\">   Below          </td> <td align=\"center\">   Low            </td> <td align=\"center\">   High             </td> <td align=\"center\">   Below    </td> </tr>\n",
+    "<tr><td align=\"center\">   Above          </td> <td align=\"center\">   Low            </td> <td align=\"center\">   Low              </td> <td align=\"center\">   Below    </td> </tr>\n",
+    "<tr><td align=\"center\">   Below          </td> <td align=\"center\">   High           </td> <td align=\"center\">   High             </td> <td align=\"center\">   Below    </td> </tr>\n",
+    "<tr><td align=\"center\">   Below          </td> <td align=\"center\">   Low            </td> <td align=\"center\">   High             </td> <td align=\"center\">   Below    </td> </tr>\n",
+    "<tr><td align=\"center\">   Above          </td> <td align=\"center\">   Low            </td> <td align=\"center\">   Low              </td> <td align=\"center\">   Below    </td> </tr>\n",
+    "<tr><td align=\"center\">   Above          </td> <td align=\"center\">   High           </td> <td align=\"center\">   High             </td> <td align=\"center\">   Above    </td> </tr>\n",
+    "</tbody>\n",
+    "</table>"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "915c9bcb",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Computing the various Gini Indices\n",
+    "\n",
+    "In computations we will translate all classes into numbers. Being\n",
+    "these binary classes, they can easily be split into ones and zeros.\n",
+    "\n",
+    "**Gini index for Average trend.**\n",
+    "\n",
+    "[See handwritten notes November 3](https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ed1fbcff",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Computing the various Gini Indices, Hours slept\n",
+    "\n",
+    "**Gini index for hour slept.**\n",
+    "\n",
+    "[See handwritten notes November 3](https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "c9d28c84",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Computing the various Gini Indices, Hours studied\n",
+    "\n",
+    "**Gini index for hour studied.**\n",
+    "\n",
+    "[See handwritten notes November 3](https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf)\n",
+    "\n",
+    "For final tree, see the above handwritten notes"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "8627a53a",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## A possible code using Scikit-Learn"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "1aa3c92d",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "# Common imports\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
     "import matplotlib.pyplot as plt\n",
-    "plt.rcParams['axes.labelsize'] = 14\n",
-    "plt.rcParams['xtick.labelsize'] = 12\n",
-    "plt.rcParams['ytick.labelsize'] = 12\n",
+    "from sklearn.tree import DecisionTreeClassifier\n",
+    "from sklearn.model_selection import train_test_split\n",
+    "from sklearn.tree import export_graphviz\n",
+    "from sklearn.preprocessing import StandardScaler, OneHotEncoder\n",
+    "from sklearn.compose import ColumnTransformer\n",
+    "from IPython.display import Image \n",
+    "from pydot import graph_from_dot_data\n",
+    "import os\n",
     "\n",
+    "# Where to save the figures and data files\n",
+    "PROJECT_ROOT_DIR = \"Results\"\n",
+    "FIGURE_ID = \"Results/FigureFiles\"\n",
+    "DATA_ID = \"DataFiles/\"\n",
     "\n",
-    "from sklearn.svm import SVC\n",
-    "from sklearn import datasets\n",
+    "if not os.path.exists(PROJECT_ROOT_DIR):\n",
+    "    os.mkdir(PROJECT_ROOT_DIR)\n",
     "\n",
+    "if not os.path.exists(FIGURE_ID):\n",
+    "    os.makedirs(FIGURE_ID)\n",
     "\n",
+    "if not os.path.exists(DATA_ID):\n",
+    "    os.makedirs(DATA_ID)\n",
     "\n",
-    "from sklearn.pipeline import Pipeline\n",
-    "from sklearn.preprocessing import StandardScaler\n",
-    "from sklearn.svm import LinearSVC\n",
+    "def image_path(fig_id):\n",
+    "    return os.path.join(FIGURE_ID, fig_id)\n",
     "\n",
+    "def data_path(dat_id):\n",
+    "    return os.path.join(DATA_ID, dat_id)\n",
     "\n",
-    "from sklearn.datasets import make_moons\n",
-    "X, y = make_moons(n_samples=100, noise=0.15, random_state=42)\n",
+    "def save_fig(fig_id):\n",
+    "    plt.savefig(image_path(fig_id) + \".png\", format='png')\n",
     "\n",
-    "def plot_dataset(X, y, axes):\n",
-    "    plt.plot(X[:, 0][y==0], X[:, 1][y==0], \"bs\")\n",
-    "    plt.plot(X[:, 0][y==1], X[:, 1][y==1], \"g^\")\n",
+    "infile = open(data_path(\"grades.csv\"),'r')\n",
+    "\n",
+    "# Read the experimental data with Pandas\n",
+    "from IPython.display import display\n",
+    "grades = pd.read_csv(infile)\n",
+    "grades = pd.DataFrame(grades)\n",
+    "display(grades)\n",
+    "# Features and targets\n",
+    "X = grades.loc[:, grades.columns != 'Grade'].values\n",
+    "y = grades.loc[:, grades.columns == 'Grade'].values\n",
+    "print(X)\n",
+    "# Then do a Classification tree\n",
+    "tree_clf = DecisionTreeClassifier(max_depth=2)\n",
+    "tree_clf.fit(X, y)\n",
+    "print(\"Train set accuracy with Decision Tree: {:.2f}\".format(tree_clf.score(X,y)))\n",
+    "#transfer to a decision tree graph\n",
+    "export_graphviz(\n",
+    "    tree_clf,\n",
+    "    out_file=\"DataFiles/grade.dot\",\n",
+    "    rounded=True,\n",
+    "    filled=True\n",
+    ")\n",
+    "cmd = 'dot -Tpng DataFiles/grade.dot -o DataFiles/grades.png'\n",
+    "os.system(cmd)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "dfd62ad1",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Further example: Computing the Gini index\n",
+    "\n",
+    "The next example we will look at is a classical one in many Machine\n",
+    "Learning applications. Based on various meteorological features, we\n",
+    "have several so-called attributes which decide whether we at the end\n",
+    "will do some outdoor activity like skiing, going for a bike ride etc\n",
+    "etc.  The table here contains the feautures **outlook**, **temperature**,\n",
+    "**humidity** and **wind**.  The target or output is whether we ride\n",
+    "(True=1) or whether we do something else that day (False=0). The\n",
+    "attributes for each feature are then sunny, overcast and rain for the\n",
+    "outlook, hot, cold and mild for temperature, high and normal for\n",
+    "humidity and weak and strong for wind.\n",
+    "\n",
+    "The table here summarizes the various attributes and\n",
+    "<table class=\"dotable\" border=\"1\">\n",
+    "<thead>\n",
+    "<tr><th align=\"center\">Day</th> <th align=\"center\">Outlook </th> <th align=\"center\">Temperature</th> <th align=\"center\">Humidity</th> <th align=\"center\"> Wind </th> <th align=\"center\">Ride</th> </tr>\n",
+    "</thead>\n",
+    "<tbody>\n",
+    "<tr><td align=\"center\">   1      </td> <td align=\"center\">   Sunny       </td> <td align=\"center\">   Hot            </td> <td align=\"center\">   High        </td> <td align=\"center\">   Weak      </td> <td align=\"center\">   0       </td> </tr>\n",
+    "<tr><td align=\"center\">   2      </td> <td align=\"center\">   Sunny       </td> <td align=\"center\">   Hot            </td> <td align=\"center\">   High        </td> <td align=\"center\">   Strong    </td> <td align=\"center\">   1       </td> </tr>\n",
+    "<tr><td align=\"center\">   3      </td> <td align=\"center\">   Overcast    </td> <td align=\"center\">   Hot            </td> <td align=\"center\">   High        </td> <td align=\"center\">   Weak      </td> <td align=\"center\">   1       </td> </tr>\n",
+    "<tr><td align=\"center\">   4      </td> <td align=\"center\">   Rain        </td> <td align=\"center\">   Mild           </td> <td align=\"center\">   High        </td> <td align=\"center\">   Weak      </td> <td align=\"center\">   1       </td> </tr>\n",
+    "<tr><td align=\"center\">   5      </td> <td align=\"center\">   Rain        </td> <td align=\"center\">   Cool           </td> <td align=\"center\">   Normal      </td> <td align=\"center\">   Weak      </td> <td align=\"center\">   1       </td> </tr>\n",
+    "<tr><td align=\"center\">   6      </td> <td align=\"center\">   Rain        </td> <td align=\"center\">   Cool           </td> <td align=\"center\">   Normal      </td> <td align=\"center\">   Strong    </td> <td align=\"center\">   0       </td> </tr>\n",
+    "<tr><td align=\"center\">   7      </td> <td align=\"center\">   Overcast    </td> <td align=\"center\">   Cool           </td> <td align=\"center\">   Normal      </td> <td align=\"center\">   Strong    </td> <td align=\"center\">   1       </td> </tr>\n",
+    "<tr><td align=\"center\">   8      </td> <td align=\"center\">   Sunny       </td> <td align=\"center\">   Mild           </td> <td align=\"center\">   High        </td> <td align=\"center\">   Weak      </td> <td align=\"center\">   0       </td> </tr>\n",
+    "<tr><td align=\"center\">   9      </td> <td align=\"center\">   Sunny       </td> <td align=\"center\">   Cool           </td> <td align=\"center\">   Normal      </td> <td align=\"center\">   Weak      </td> <td align=\"center\">   1       </td> </tr>\n",
+    "<tr><td align=\"center\">   10     </td> <td align=\"center\">   Rain        </td> <td align=\"center\">   Mild           </td> <td align=\"center\">   Normal      </td> <td align=\"center\">   Weak      </td> <td align=\"center\">   1       </td> </tr>\n",
+    "<tr><td align=\"center\">   11     </td> <td align=\"center\">   Sunny       </td> <td align=\"center\">   Mild           </td> <td align=\"center\">   Normal      </td> <td align=\"center\">   Strong    </td> <td align=\"center\">   1       </td> </tr>\n",
+    "<tr><td align=\"center\">   12     </td> <td align=\"center\">   Overcast    </td> <td align=\"center\">   Mild           </td> <td align=\"center\">   High        </td> <td align=\"center\">   Strong    </td> <td align=\"center\">   1       </td> </tr>\n",
+    "<tr><td align=\"center\">   13     </td> <td align=\"center\">   Overcast    </td> <td align=\"center\">   Hot            </td> <td align=\"center\">   Normal      </td> <td align=\"center\">   Weak      </td> <td align=\"center\">   1       </td> </tr>\n",
+    "<tr><td align=\"center\">   14     </td> <td align=\"center\">   Rain        </td> <td align=\"center\">   Mild           </td> <td align=\"center\">   High        </td> <td align=\"center\">   Strong    </td> <td align=\"center\">   0       </td> </tr>\n",
+    "</tbody>\n",
+    "</table>"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e6306992",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Simple Python Code to read in Data and perform Classification"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "08cbf1ba",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "# Common imports\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "from sklearn.tree import DecisionTreeClassifier\n",
+    "from sklearn.model_selection import train_test_split\n",
+    "from sklearn.tree import export_graphviz\n",
+    "from sklearn.preprocessing import StandardScaler, OneHotEncoder\n",
+    "from sklearn.compose import ColumnTransformer\n",
+    "from IPython.display import Image \n",
+    "from pydot import graph_from_dot_data\n",
+    "import os\n",
+    "\n",
+    "# Where to save the figures and data files\n",
+    "PROJECT_ROOT_DIR = \"Results\"\n",
+    "FIGURE_ID = \"Results/FigureFiles\"\n",
+    "DATA_ID = \"DataFiles/\"\n",
+    "\n",
+    "if not os.path.exists(PROJECT_ROOT_DIR):\n",
+    "    os.mkdir(PROJECT_ROOT_DIR)\n",
+    "\n",
+    "if not os.path.exists(FIGURE_ID):\n",
+    "    os.makedirs(FIGURE_ID)\n",
+    "\n",
+    "if not os.path.exists(DATA_ID):\n",
+    "    os.makedirs(DATA_ID)\n",
+    "\n",
+    "def image_path(fig_id):\n",
+    "    return os.path.join(FIGURE_ID, fig_id)\n",
+    "\n",
+    "def data_path(dat_id):\n",
+    "    return os.path.join(DATA_ID, dat_id)\n",
+    "\n",
+    "def save_fig(fig_id):\n",
+    "    plt.savefig(image_path(fig_id) + \".png\", format='png')\n",
+    "\n",
+    "infile = open(data_path(\"rideclass.csv\"),'r')\n",
+    "\n",
+    "# Read the experimental data with Pandas\n",
+    "from IPython.display import display\n",
+    "ridedata = pd.read_csv(infile,names = ('Outlook','Temperature','Humidity','Wind','Ride'))\n",
+    "ridedata = pd.DataFrame(ridedata)\n",
+    "\n",
+    "# Features and targets\n",
+    "X = ridedata.loc[:, ridedata.columns != 'Ride'].values\n",
+    "y = ridedata.loc[:, ridedata.columns == 'Ride'].values\n",
+    "\n",
+    "# Create the encoder.\n",
+    "encoder = OneHotEncoder(handle_unknown=\"ignore\")\n",
+    "# Assume for simplicity all features are categorical.\n",
+    "encoder.fit(X)    \n",
+    "# Apply the encoder.\n",
+    "X = encoder.transform(X)\n",
+    "print(X)\n",
+    "# Then do a Classification tree\n",
+    "tree_clf = DecisionTreeClassifier(max_depth=2)\n",
+    "tree_clf.fit(X, y)\n",
+    "print(\"Train set accuracy with Decision Tree: {:.2f}\".format(tree_clf.score(X,y)))\n",
+    "#transfer to a decision tree graph\n",
+    "export_graphviz(\n",
+    "    tree_clf,\n",
+    "    out_file=\"DataFiles/ride.dot\",\n",
+    "    rounded=True,\n",
+    "    filled=True\n",
+    ")\n",
+    "cmd = 'dot -Tpng DataFiles/cancer.dot -o DataFiles/cancer.png'\n",
+    "os.system(cmd)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "badaa462",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Computing the Gini Factor\n",
+    "\n",
+    "The above functions (gini, entropy and misclassification error) are\n",
+    "important components of the so-called CART algorithm. We will discuss\n",
+    "this algorithm below after we have discussed the information gain\n",
+    "algorithm ID3.\n",
+    "\n",
+    "In the example here we have converted all our attributes into numerical values $0,1,2$ etc."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "d7eca4e3",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "# Split a dataset based on an attribute and an attribute value\n",
+    "def test_split(index, value, dataset):\n",
+    "\tleft, right = list(), list()\n",
+    "\tfor row in dataset:\n",
+    "\t\tif row[index] < value:\n",
+    "\t\t\tleft.append(row)\n",
+    "\t\telse:\n",
+    "\t\t\tright.append(row)\n",
+    "\treturn left, right\n",
+    " \n",
+    "# Calculate the Gini index for a split dataset\n",
+    "def gini_index(groups, classes):\n",
+    "\t# count all samples at split point\n",
+    "\tn_instances = float(sum([len(group) for group in groups]))\n",
+    "\t# sum weighted Gini index for each group\n",
+    "\tgini = 0.0\n",
+    "\tfor group in groups:\n",
+    "\t\tsize = float(len(group))\n",
+    "\t\t# avoid divide by zero\n",
+    "\t\tif size == 0:\n",
+    "\t\t\tcontinue\n",
+    "\t\tscore = 0.0\n",
+    "\t\t# score the group based on the score for each class\n",
+    "\t\tfor class_val in classes:\n",
+    "\t\t\tp = [row[-1] for row in group].count(class_val) / size\n",
+    "\t\t\tscore += p * p\n",
+    "\t\t# weight the group score by its relative size\n",
+    "\t\tgini += (1.0 - score) * (size / n_instances)\n",
+    "\treturn gini\n",
+    "\n",
+    "# Select the best split point for a dataset\n",
+    "def get_split(dataset):\n",
+    "\tclass_values = list(set(row[-1] for row in dataset))\n",
+    "\tb_index, b_value, b_score, b_groups = 999, 999, 999, None\n",
+    "\tfor index in range(len(dataset[0])-1):\n",
+    "\t\tfor row in dataset:\n",
+    "\t\t\tgroups = test_split(index, row[index], dataset)\n",
+    "\t\t\tgini = gini_index(groups, class_values)\n",
+    "\t\t\tprint('X%d < %.3f Gini=%.3f' % ((index+1), row[index], gini))\n",
+    "\t\t\tif gini < b_score:\n",
+    "\t\t\t\tb_index, b_value, b_score, b_groups = index, row[index], gini, groups\n",
+    "\treturn {'index':b_index, 'value':b_value, 'groups':b_groups}\n",
+    " \n",
+    "dataset = [[0,0,0,0,0],\n",
+    "            [0,0,0,1,1],\n",
+    "            [1,0,0,0,1],\n",
+    "            [2,1,0,0,1],\n",
+    "            [2,2,1,0,1],\n",
+    "            [2,2,1,1,0],\n",
+    "            [1,2,1,1,1],\n",
+    "            [0,1,0,0,0],\n",
+    "            [0,2,1,0,1],\n",
+    "            [2,1,1,0,1],\n",
+    "            [0,1,1,1,1],\n",
+    "            [1,1,0,1,1],\n",
+    "            [1,0,1,0,1],\n",
+    "            [2,1,0,1,0]]\n",
+    "\n",
+    "split = get_split(dataset)\n",
+    "print('Split: [X%d < %.3f]' % ((split['index']+1), split['value']))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e609f3f0",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Regression trees"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "7783ddc2",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "# Quadratic training set + noise\n",
+    "np.random.seed(42)\n",
+    "m = 200\n",
+    "X = np.random.rand(m, 1)\n",
+    "y = 4 * (X - 0.5) ** 2\n",
+    "y = y + np.random.randn(m, 1) / 10"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "ef4758a4",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "from sklearn.tree import DecisionTreeRegressor\n",
+    "\n",
+    "tree_reg = DecisionTreeRegressor(max_depth=2, random_state=42)\n",
+    "tree_reg.fit(X, y)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a4ecb5a5",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Final regressor code"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "1fd6944c",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "from sklearn.tree import DecisionTreeRegressor\n",
+    "\n",
+    "tree_reg1 = DecisionTreeRegressor(random_state=42, max_depth=2)\n",
+    "tree_reg2 = DecisionTreeRegressor(random_state=42, max_depth=3)\n",
+    "tree_reg1.fit(X, y)\n",
+    "tree_reg2.fit(X, y)\n",
+    "\n",
+    "def plot_regression_predictions(tree_reg, X, y, axes=[0, 1, -0.2, 1], ylabel=\"$y$\"):\n",
+    "    x1 = np.linspace(axes[0], axes[1], 500).reshape(-1, 1)\n",
+    "    y_pred = tree_reg.predict(x1)\n",
     "    plt.axis(axes)\n",
-    "    plt.grid(True, which='both')\n",
-    "    plt.xlabel(r\"$x_1$\", fontsize=20)\n",
-    "    plt.ylabel(r\"$x_2$\", fontsize=20, rotation=0)\n",
+    "    plt.xlabel(\"$x_1$\", fontsize=18)\n",
+    "    if ylabel:\n",
+    "        plt.ylabel(ylabel, fontsize=18, rotation=0)\n",
+    "    plt.plot(X, y, \"b.\")\n",
+    "    plt.plot(x1, y_pred, \"r.-\", linewidth=2, label=r\"$\\hat{y}$\")\n",
     "\n",
-    "plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])\n",
-    "plt.show()\n",
+    "plt.figure(figsize=(11, 4))\n",
+    "plt.subplot(121)\n",
+    "plot_regression_predictions(tree_reg1, X, y)\n",
+    "for split, style in ((0.1973, \"k-\"), (0.0917, \"k--\"), (0.7718, \"k--\")):\n",
+    "    plt.plot([split, split], [-0.2, 1], style, linewidth=2)\n",
+    "plt.text(0.21, 0.65, \"Depth=0\", fontsize=15)\n",
+    "plt.text(0.01, 0.2, \"Depth=1\", fontsize=13)\n",
+    "plt.text(0.65, 0.8, \"Depth=1\", fontsize=13)\n",
+    "plt.legend(loc=\"upper center\", fontsize=18)\n",
+    "plt.title(\"max_depth=2\", fontsize=14)\n",
     "\n",
-    "from sklearn.datasets import make_moons\n",
-    "from sklearn.pipeline import Pipeline\n",
-    "from sklearn.preprocessing import PolynomialFeatures\n",
+    "plt.subplot(122)\n",
+    "plot_regression_predictions(tree_reg2, X, y, ylabel=None)\n",
+    "for split, style in ((0.1973, \"k-\"), (0.0917, \"k--\"), (0.7718, \"k--\")):\n",
+    "    plt.plot([split, split], [-0.2, 1], style, linewidth=2)\n",
+    "for split in (0.0458, 0.1298, 0.2873, 0.9040):\n",
+    "    plt.plot([split, split], [-0.2, 1], \"k:\", linewidth=1)\n",
+    "plt.text(0.3, 0.5, \"Depth=2\", fontsize=13)\n",
+    "plt.title(\"max_depth=3\", fontsize=14)\n",
     "\n",
-    "polynomial_svm_clf = Pipeline([\n",
-    "        (\"poly_features\", PolynomialFeatures(degree=3)),\n",
-    "        (\"scaler\", StandardScaler()),\n",
-    "        (\"svm_clf\", LinearSVC(C=10, loss=\"hinge\", random_state=42))\n",
-    "    ])\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "e9903848",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "tree_reg1 = DecisionTreeRegressor(random_state=42)\n",
+    "tree_reg2 = DecisionTreeRegressor(random_state=42, min_samples_leaf=10)\n",
+    "tree_reg1.fit(X, y)\n",
+    "tree_reg2.fit(X, y)\n",
     "\n",
-    "polynomial_svm_clf.fit(X, y)\n",
-    "\n",
-    "def plot_predictions(clf, axes):\n",
-    "    x0s = np.linspace(axes[0], axes[1], 100)\n",
-    "    x1s = np.linspace(axes[2], axes[3], 100)\n",
-    "    x0, x1 = np.meshgrid(x0s, x1s)\n",
-    "    X = np.c_[x0.ravel(), x1.ravel()]\n",
-    "    y_pred = clf.predict(X).reshape(x0.shape)\n",
-    "    y_decision = clf.decision_function(X).reshape(x0.shape)\n",
-    "    plt.contourf(x0, x1, y_pred, cmap=plt.cm.brg, alpha=0.2)\n",
-    "    plt.contourf(x0, x1, y_decision, cmap=plt.cm.brg, alpha=0.1)\n",
-    "\n",
-    "plot_predictions(polynomial_svm_clf, [-1.5, 2.5, -1, 1.5])\n",
-    "plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])\n",
-    "\n",
-    "plt.show()\n",
-    "\n",
-    "\n",
-    "from sklearn.svm import SVC\n",
-    "\n",
-    "poly_kernel_svm_clf = Pipeline([\n",
-    "        (\"scaler\", StandardScaler()),\n",
-    "        (\"svm_clf\", SVC(kernel=\"poly\", degree=3, coef0=1, C=5))\n",
-    "    ])\n",
-    "poly_kernel_svm_clf.fit(X, y)\n",
-    "\n",
-    "poly100_kernel_svm_clf = Pipeline([\n",
-    "        (\"scaler\", StandardScaler()),\n",
-    "        (\"svm_clf\", SVC(kernel=\"poly\", degree=10, coef0=100, C=5))\n",
-    "    ])\n",
-    "poly100_kernel_svm_clf.fit(X, y)\n",
+    "x1 = np.linspace(0, 1, 500).reshape(-1, 1)\n",
+    "y_pred1 = tree_reg1.predict(x1)\n",
+    "y_pred2 = tree_reg2.predict(x1)\n",
     "\n",
     "plt.figure(figsize=(11, 4))\n",
     "\n",
     "plt.subplot(121)\n",
-    "plot_predictions(poly_kernel_svm_clf, [-1.5, 2.5, -1, 1.5])\n",
-    "plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])\n",
-    "plt.title(r\"$d=3, r=1, C=5$\", fontsize=18)\n",
+    "plt.plot(X, y, \"b.\")\n",
+    "plt.plot(x1, y_pred1, \"r.-\", linewidth=2, label=r\"$\\hat{y}$\")\n",
+    "plt.axis([0, 1, -0.2, 1.1])\n",
+    "plt.xlabel(\"$x_1$\", fontsize=18)\n",
+    "plt.ylabel(\"$y$\", fontsize=18, rotation=0)\n",
+    "plt.legend(loc=\"upper center\", fontsize=18)\n",
+    "plt.title(\"No restrictions\", fontsize=14)\n",
     "\n",
     "plt.subplot(122)\n",
-    "plot_predictions(poly100_kernel_svm_clf, [-1.5, 2.5, -1, 1.5])\n",
-    "plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])\n",
-    "plt.title(r\"$d=10, r=100, C=5$\", fontsize=18)\n",
-    "\n",
-    "plt.show()\n",
-    "\n",
-    "def gaussian_rbf(x, landmark, gamma):\n",
-    "    return np.exp(-gamma * np.linalg.norm(x - landmark, axis=1)**2)\n",
-    "\n",
-    "gamma = 0.3\n",
-    "\n",
-    "x1s = np.linspace(-4.5, 4.5, 200).reshape(-1, 1)\n",
-    "x2s = gaussian_rbf(x1s, -2, gamma)\n",
-    "x3s = gaussian_rbf(x1s, 1, gamma)\n",
-    "\n",
-    "XK = np.c_[gaussian_rbf(X1D, -2, gamma), gaussian_rbf(X1D, 1, gamma)]\n",
-    "yk = np.array([0, 0, 1, 1, 1, 1, 1, 0, 0])\n",
-    "\n",
-    "plt.figure(figsize=(11, 4))\n",
-    "\n",
-    "plt.subplot(121)\n",
-    "plt.grid(True, which='both')\n",
-    "plt.axhline(y=0, color='k')\n",
-    "plt.scatter(x=[-2, 1], y=[0, 0], s=150, alpha=0.5, c=\"red\")\n",
-    "plt.plot(X1D[:, 0][yk==0], np.zeros(4), \"bs\")\n",
-    "plt.plot(X1D[:, 0][yk==1], np.zeros(5), \"g^\")\n",
-    "plt.plot(x1s, x2s, \"g--\")\n",
-    "plt.plot(x1s, x3s, \"b:\")\n",
-    "plt.gca().get_yaxis().set_ticks([0, 0.25, 0.5, 0.75, 1])\n",
-    "plt.xlabel(r\"$x_1$\", fontsize=20)\n",
-    "plt.ylabel(r\"Similarity\", fontsize=14)\n",
-    "plt.annotate(r'$\\mathbf{x}$',\n",
-    "             xy=(X1D[3, 0], 0),\n",
-    "             xytext=(-0.5, 0.20),\n",
-    "             ha=\"center\",\n",
-    "             arrowprops=dict(facecolor='black', shrink=0.1),\n",
-    "             fontsize=18,\n",
-    "            )\n",
-    "plt.text(-2, 0.9, \"$x_2$\", ha=\"center\", fontsize=20)\n",
-    "plt.text(1, 0.9, \"$x_3$\", ha=\"center\", fontsize=20)\n",
-    "plt.axis([-4.5, 4.5, -0.1, 1.1])\n",
-    "\n",
-    "plt.subplot(122)\n",
-    "plt.grid(True, which='both')\n",
-    "plt.axhline(y=0, color='k')\n",
-    "plt.axvline(x=0, color='k')\n",
-    "plt.plot(XK[:, 0][yk==0], XK[:, 1][yk==0], \"bs\")\n",
-    "plt.plot(XK[:, 0][yk==1], XK[:, 1][yk==1], \"g^\")\n",
-    "plt.xlabel(r\"$x_2$\", fontsize=20)\n",
-    "plt.ylabel(r\"$x_3$  \", fontsize=20, rotation=0)\n",
-    "plt.annotate(r'$\\phi\\left(\\mathbf{x}\\right)$',\n",
-    "             xy=(XK[3, 0], XK[3, 1]),\n",
-    "             xytext=(0.65, 0.50),\n",
-    "             ha=\"center\",\n",
-    "             arrowprops=dict(facecolor='black', shrink=0.1),\n",
-    "             fontsize=18,\n",
-    "            )\n",
-    "plt.plot([-0.1, 1.1], [0.57, -0.1], \"r--\", linewidth=3)\n",
-    "plt.axis([-0.1, 1.1, -0.1, 1.1])\n",
-    "    \n",
-    "plt.subplots_adjust(right=1)\n",
-    "\n",
-    "plt.show()\n",
-    "\n",
-    "\n",
-    "x1_example = X1D[3, 0]\n",
-    "for landmark in (-2, 1):\n",
-    "    k = gaussian_rbf(np.array([[x1_example]]), np.array([[landmark]]), gamma)\n",
-    "    print(\"Phi({}, {}) = {}\".format(x1_example, landmark, k))\n",
-    "\n",
-    "rbf_kernel_svm_clf = Pipeline([\n",
-    "        (\"scaler\", StandardScaler()),\n",
-    "        (\"svm_clf\", SVC(kernel=\"rbf\", gamma=5, C=0.001))\n",
-    "    ])\n",
-    "rbf_kernel_svm_clf.fit(X, y)\n",
-    "\n",
-    "\n",
-    "from sklearn.svm import SVC\n",
-    "\n",
-    "gamma1, gamma2 = 0.1, 5\n",
-    "C1, C2 = 0.001, 1000\n",
-    "hyperparams = (gamma1, C1), (gamma1, C2), (gamma2, C1), (gamma2, C2)\n",
-    "\n",
-    "svm_clfs = []\n",
-    "for gamma, C in hyperparams:\n",
-    "    rbf_kernel_svm_clf = Pipeline([\n",
-    "            (\"scaler\", StandardScaler()),\n",
-    "            (\"svm_clf\", SVC(kernel=\"rbf\", gamma=gamma, C=C))\n",
-    "        ])\n",
-    "    rbf_kernel_svm_clf.fit(X, y)\n",
-    "    svm_clfs.append(rbf_kernel_svm_clf)\n",
-    "\n",
-    "plt.figure(figsize=(11, 7))\n",
-    "\n",
-    "for i, svm_clf in enumerate(svm_clfs):\n",
-    "    plt.subplot(221 + i)\n",
-    "    plot_predictions(svm_clf, [-1.5, 2.5, -1, 1.5])\n",
-    "    plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])\n",
-    "    gamma, C = hyperparams[i]\n",
-    "    plt.title(r\"$\\gamma = {}, C = {}$\".format(gamma, C), fontsize=16)\n",
+    "plt.plot(X, y, \"b.\")\n",
+    "plt.plot(x1, y_pred2, \"r.-\", linewidth=2, label=r\"$\\hat{y}$\")\n",
+    "plt.axis([0, 1, -0.2, 1.1])\n",
+    "plt.xlabel(\"$x_1$\", fontsize=18)\n",
+    "plt.title(\"min_samples_leaf={}\".format(tree_reg2.min_samples_leaf), fontsize=14)\n",
     "\n",
     "plt.show()"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "13ea0780",
+   "id": "736c7714",
    "metadata": {
     "editable": true
    },
    "source": [
-    "## Mathematical optimization of convex functions\n",
+    "## Pros and cons of trees, pros\n",
     "\n",
-    "A mathematical (quadratic) optimization problem, or just optimization problem, has the form"
+    "* White box, easy to interpret model. Some people believe that decision trees more closely mirror human decision-making than do the regression and classification approaches discussed earlier (think of support vector machines)\n",
+    "\n",
+    "* Trees are very easy to explain to people. In fact, they are even easier to explain than linear regression!\n",
+    "\n",
+    "* No feature normalization needed\n",
+    "\n",
+    "* Tree models can handle both continuous and categorical data (Classification and Regression Trees)\n",
+    "\n",
+    "* Can model nonlinear relationships\n",
+    "\n",
+    "* Can model interactions between the different descriptive features\n",
+    "\n",
+    "* Trees can be displayed graphically, and are easily interpreted even by a non-expert (especially if they are small)"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "083eee82",
+   "id": "641c4ab7",
    "metadata": {
     "editable": true
    },
    "source": [
-    "$$\n",
-    "\\begin{align*}\n",
-    "    &\\mathrm{min}_{\\lambda}\\hspace{0.2cm} \\frac{1}{2}\\boldsymbol{\\lambda}^T\\boldsymbol{P}\\boldsymbol{\\lambda}+\\boldsymbol{q}^T\\boldsymbol{\\lambda},\\\\ \\nonumber\n",
-    "    &\\mathrm{subject\\hspace{0.1cm}to} \\hspace{0.2cm} \\boldsymbol{G}\\boldsymbol{\\lambda} \\preceq \\boldsymbol{h} \\wedge  \\boldsymbol{A}\\boldsymbol{\\lambda}=f.\n",
-    "\\end{align*}\n",
-    "$$"
+    "## Disadvantages\n",
+    "\n",
+    "* Unfortunately, trees generally do not have the same level of predictive accuracy as some of the other regression and classification approaches\n",
+    "\n",
+    "* If continuous features are used the tree may become quite large and hence less interpretable\n",
+    "\n",
+    "* Decision trees are prone to overfit the training data and hence do not well generalize the data if no stopping criteria or improvements like pruning, boosting or bagging are implemented\n",
+    "\n",
+    "* Small changes in the data may lead to a completely different tree. This issue can be addressed by using ensemble methods like bagging, boosting or random forests\n",
+    "\n",
+    "* Unbalanced datasets where some target feature values occur much more frequently than others may lead to biased trees since the frequently occurring feature values are preferred over the less frequently occurring ones. \n",
+    "\n",
+    "* If the number of features is relatively large (high dimensional) and the number of instances is relatively low, the tree might overfit the data\n",
+    "\n",
+    "* Features with many levels may be preferred over features with less levels since for them it is *more easy* to split the dataset such that the sub datasets only contain pure target feature values. This issue can be addressed by preferring for instance the information gain ratio as splitting criteria over information gain\n",
+    "\n",
+    "However, by aggregating many decision trees, using methods like\n",
+    "bagging, random forests, and boosting, the predictive performance of\n",
+    "trees can be substantially improved."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "cb700221",
+   "id": "f2e72566",
    "metadata": {
     "editable": true
    },
    "source": [
-    "subject to some constraints for say a selected set $i=1,2,\\dots, n$.\n",
-    "In our case we are optimizing with respect to the Lagrangian multipliers $\\lambda_i$, and the\n",
-    "vector $\\boldsymbol{\\lambda}=[\\lambda_1, \\lambda_2,\\dots, \\lambda_n]$ is the optimization variable we are dealing with.\n",
+    "## Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods\n",
     "\n",
-    "In our case we are particularly interested in a class of optimization problems called convex optmization problems. \n",
-    "In our discussion on gradient descent methods we discussed at length the definition of a convex function. \n",
+    "As stated above and seen in many of the examples discussed here about\n",
+    "a single decision tree, we often end up overfitting our training\n",
+    "data. This normally means that we have a high variance. Can we reduce\n",
+    "the variance of a statistical learning method?\n",
     "\n",
-    "Convex optimization problems play a central role in applied mathematics and we recommend strongly [Boyd and Vandenberghe's text on the topics](http://web.stanford.edu/~boyd/cvxbook/)."
+    "This leads us to a set of different methods that can combine different\n",
+    "machine learning algorithms or just use one of them to construct\n",
+    "forests and jungles of trees, homogeneous ones or heterogenous\n",
+    "ones. These methods are recognized by different names which we will\n",
+    "try to explain here. These are\n",
+    "\n",
+    "1. Voting classifiers\n",
+    "\n",
+    "2. Bagging and Pasting\n",
+    "\n",
+    "3. Random forests\n",
+    "\n",
+    "4. Boosting methods, from adaptive to Extreme Gradient Boosting (XGBoost)\n",
+    "\n",
+    "We discuss these methods here."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "77006d30",
+   "id": "ea107cee",
    "metadata": {
     "editable": true
    },
    "source": [
-    "## How do we solve these problems?\n",
+    "## An Overview of Ensemble Methods\n",
     "\n",
-    "If we use Python as programming language and wish to venture beyond\n",
-    "**scikit-learn**, **tensorflow** and similar software which makes our\n",
-    "lives so much easier, we need to dive into the wonderful world of\n",
-    "quadratic programming. We can, if we wish, solve the minimization\n",
-    "problem using say standard gradient methods or conjugate gradient\n",
-    "methods. However, these methods tend to exhibit a rather slow\n",
-    "converge. So, welcome to the promised land of quadratic programming.\n",
+    "<!-- dom:FIGURE: [DataFiles/ensembleoverview.png, width=600 frac=0.8] -->\n",
+    "<!-- begin figure -->\n",
     "\n",
-    "The functions we need are contained in the [quadratic programming library CVXOPT](https://cvxopt.org/userguide/coneprog.html) and we need to import it together with **numpy** as"
+    "<img src=\"DataFiles/ensembleoverview.png\" width=\"600\"><p style=\"font-size: 0.9em\"><i>Figure 1: </i></p>\n",
+    "<!-- end figure -->"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1b1c2c46",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Why Voting?\n",
+    "\n",
+    "The idea behind boosting, and voting as well can be phrased as follows:\n",
+    "**Can a group of people somehow arrive at highly\n",
+    "reasoned decisions, despite the weak judgement of the individual\n",
+    "members?**\n",
+    "\n",
+    "The aim is to create a good classifier by combining several weak classifiers.\n",
+    "**A weak classifier is a classifier which is able to produce results that are only slightly better than guessing at random.**\n",
+    "\n",
+    "The basic approach is to apply repeatedly (in boosting this is done in an iterative way) a weak classifier to modifications of the data.\n",
+    "In voting we simply apply the law of large numbers while in boosting we give more weight to misclassified data in\n",
+    "each iteration.  \n",
+    "\n",
+    "Decision trees play an important role as our weak classifier. They serve as the basic method."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "78671c3e",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Tossing coins\n",
+    "\n",
+    "The simplest case is a so-called voting ensemble. To illustrate this,\n",
+    "think of yourself tossing coins with a biased outcome of 51 per cent\n",
+    "for heads and 49% for tails. With only few tosses,\n",
+    "you may not clearly see this distribution for heads and tails. However, after some\n",
+    "thousands of tosses, there will be a clear majority of heads.  With 2000 tosses\n",
+    "you should see approximately 1020 heads and 980 tails.\n",
+    "\n",
+    "We can then state that the outcome is a clear majority of heads. If\n",
+    "you do this ten thousand times, it is easy to see that there is a 97%\n",
+    "likelihood of a majority of heads.\n",
+    "\n",
+    "Another example would be to collect all polls before an\n",
+    "election. Different polls may show different likelihoods for a\n",
+    "candidate winning with say a majority  of the popular vote. The majority vote\n",
+    "would then consist in many polls indicating that this candidate will\n",
+    "actually win.\n",
+    "\n",
+    "The example here shows how we can implement the coin tossing case,\n",
+    "clealry demostrating that after some tosses we see the [law of large](https://en.wikipedia.org/wiki/Law_of_large_numbers)\n",
+    "numbers kicking in."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2e9572b2",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Standard imports first"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
-   "id": "cc9de138",
+   "execution_count": 13,
+   "id": "87a91216",
    "metadata": {
     "collapsed": false,
     "editable": true
    },
    "outputs": [],
    "source": [
-    "import numpy\n",
-    "import cvxopt"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "5a22f71f",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "This will make our life much easier. You don't need t0 write your own optimizer."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "87a174c1",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "## A simple example\n",
+    "# Common imports\n",
+    "from IPython.display import Image \n",
+    "from pydot import graph_from_dot_data\n",
+    "import pandas as pd\n",
+    "import numpy as np\n",
+    "import matplotlib.pyplot as plt\n",
+    "from sklearn.tree import DecisionTreeClassifier\n",
+    "from sklearn.model_selection import train_test_split\n",
+    "from sklearn.tree import export_graphviz\n",
+    "from sklearn.preprocessing import StandardScaler, OneHotEncoder\n",
+    "from sklearn.compose import ColumnTransformer\n",
+    "from IPython.display import Image \n",
+    "from pydot import graph_from_dot_data\n",
+    "import os\n",
     "\n",
-    "We remind ourselves about the general problem we want to solve"
+    "# Where to save the figures and data files\n",
+    "PROJECT_ROOT_DIR = \"Results\"\n",
+    "FIGURE_ID = \"Results/FigureFiles\"\n",
+    "DATA_ID = \"DataFiles/\"\n",
+    "\n",
+    "if not os.path.exists(PROJECT_ROOT_DIR):\n",
+    "    os.mkdir(PROJECT_ROOT_DIR)\n",
+    "\n",
+    "if not os.path.exists(FIGURE_ID):\n",
+    "    os.makedirs(FIGURE_ID)\n",
+    "\n",
+    "if not os.path.exists(DATA_ID):\n",
+    "    os.makedirs(DATA_ID)\n",
+    "\n",
+    "def image_path(fig_id):\n",
+    "    return os.path.join(FIGURE_ID, fig_id)\n",
+    "\n",
+    "def data_path(dat_id):\n",
+    "    return os.path.join(DATA_ID, dat_id)\n",
+    "\n",
+    "def save_fig(fig_id):\n",
+    "    plt.savefig(image_path(fig_id) + \".png\", format='png')"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "d5476231",
+   "id": "ec21ddcb",
    "metadata": {
     "editable": true
    },
    "source": [
-    "$$\n",
-    "\\begin{align*}\n",
-    "    &\\mathrm{min}_{x}\\hspace{0.2cm} \\frac{1}{2}\\boldsymbol{x}^T\\boldsymbol{P}\\boldsymbol{x}+\\boldsymbol{q}^T\\boldsymbol{x},\\\\ \\nonumber\n",
-    "    &\\mathrm{subject\\hspace{0.1cm} to} \\hspace{0.2cm} \\boldsymbol{G}\\boldsymbol{x} \\preceq \\boldsymbol{h} \\wedge  \\boldsymbol{A}\\boldsymbol{x}=f.\n",
-    "\\end{align*}\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "119e4dd9",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "Let us show how to perform the optmization using a simple case. Assume we want to optimize the following problem"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "0848f2ad",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\begin{align*}\n",
-    "    &\\mathrm{min}_{x}\\hspace{0.2cm} \\frac{1}{2}x^2+5x+3y \\\\ \\nonumber\n",
-    "    &\\mathrm{subject}\\hspace{0.5cm} \\mathrm{to} \\\\ \\nonumber\n",
-    "    &x, y \\geq 0 \\\\ \\nonumber\n",
-    "    &x+3y  \\geq 15 \\\\ \\nonumber\n",
-    "    &2x+5y  \\leq  100 \\\\ \\nonumber\n",
-    "    &3x+4y  \\leq  80.  \\\\ \\nonumber\n",
-    "\\end{align*}\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "5c1f1048",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "The minimization problem can be rewritten in terms of vectors and matrices as (with $x$ and $y$ being the unknowns)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "94c0fa01",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\frac{1}{2}\\begin{bmatrix} x\\\\ y \\end{bmatrix}^T   \\begin{bmatrix} 1 & 0\\\\ 0 & 0  \\end{bmatrix}  \\begin{bmatrix} x \\\\ y \\end{bmatrix}  + \\begin{bmatrix}3\\\\ 4  \\end{bmatrix}^T \\begin{bmatrix}x \\\\ y  \\end{bmatrix}.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "8a98b526",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "Similarly, we can now set up the inequalities (we need to change $\\geq$ to $\\leq$ by multiplying with $-1$ on bot sides) as the following matrix-vector equation"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "e2192fc7",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\begin{bmatrix} -1 & 0 \\\\ 0 & -1 \\\\ -1 & -3 \\\\ 2 & 5 \\\\ 3 & 4\\end{bmatrix}\\begin{bmatrix} x \\\\ y\\end{bmatrix} \\preceq \\begin{bmatrix}0 \\\\ 0\\\\ -15 \\\\ 100 \\\\ 80\\end{bmatrix}.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "9774b659",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "We have collapsed all the inequalities into a single matrix $\\boldsymbol{G}$. We see also that our matrix"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "b0fe573a",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\boldsymbol{P} =\\begin{bmatrix} 1 & 0\\\\ 0 & 0  \\end{bmatrix}\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "40c5c695",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "is clearly positive semi-definite (all eigenvalues larger or equal zero). \n",
-    "Finally, the vector $\\boldsymbol{h}$ is defined as"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "0ee6039e",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "$$\n",
-    "\\boldsymbol{h} = \\begin{bmatrix}0 \\\\ 0\\\\ -15 \\\\ 100 \\\\ 80\\end{bmatrix}.\n",
-    "$$"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "667407ca",
-   "metadata": {
-    "editable": true
-   },
-   "source": [
-    "Since we don't have any equalities the matrix $\\boldsymbol{A}$ is set to zero\n",
-    "The following code solves the equations for us"
+    "## Simple Voting Example, head or tail"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
-   "id": "2375448c",
+   "execution_count": 14,
+   "id": "7efa5983",
    "metadata": {
     "collapsed": false,
     "editable": true
    },
    "outputs": [],
    "source": [
-    "# Import the necessary packages\n",
-    "import numpy\n",
-    "from cvxopt import matrix\n",
-    "from cvxopt import solvers\n",
-    "P = matrix(numpy.diag([1,0]), tc='d')\n",
-    "q = matrix(numpy.array([3,4]), tc='d')\n",
-    "G = matrix(numpy.array([[-1,0],[0,-1],[-1,-3],[2,5],[3,4]]), tc='d')\n",
-    "h = matrix(numpy.array([0,0,-15,100,80]), tc='d')\n",
-    "# Construct the QP, invoke solver\n",
-    "sol = solvers.qp(P,q,G,h)\n",
-    "# Extract optimal value and solution\n",
-    "sol['x'] \n",
-    "sol['primal objective']\n",
-    "print(sol['x'] )"
+    "\n",
+    "# Common imports\n",
+    "import numpy as np\n",
+    "import matplotlib\n",
+    "import matplotlib.pyplot as plt\n",
+    "from matplotlib.colors import ListedColormap\n",
+    "plt.rcParams['axes.labelsize'] = 14\n",
+    "plt.rcParams['xtick.labelsize'] = 12\n",
+    "plt.rcParams['ytick.labelsize'] = 12\n",
+    "\n",
+    "heads_proba = 0.51\n",
+    "coin_tosses = (np.random.rand(10000, 10) < heads_proba).astype(np.int32)\n",
+    "cumulative_heads_ratio = np.cumsum(coin_tosses, axis=0) / np.arange(1, 10001).reshape(-1, 1)\n",
+    "plt.figure(figsize=(8,3.5))\n",
+    "plt.plot(cumulative_heads_ratio)\n",
+    "plt.plot([0, 10000], [0.51, 0.51], \"k--\", linewidth=2, label=\"51%\")\n",
+    "plt.plot([0, 10000], [0.5, 0.5], \"k-\", label=\"50%\")\n",
+    "plt.xlabel(\"Number of coin tosses\")\n",
+    "plt.ylabel(\"Heads ratio\")\n",
+    "plt.legend(loc=\"lower right\")\n",
+    "plt.axis([0, 10000, 0.42, 0.58])\n",
+    "save_fig(\"votingsimple\")\n",
+    "plt.show()"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "815d1235",
+   "id": "c94fdd9e",
    "metadata": {
     "editable": true
    },
    "source": [
-    "## Back to the more realistic cases\n",
+    "## Using the Voting Classifier\n",
     "\n",
-    "We are now ready to return to our setup of the optmization problem for a more realistic case. Introducing the **slack** parameter $C$ we have"
+    "We can use the voting classifier on other data sets, here the exciting binary case of two distinct objects using the make moons functionality of **Scikit-Learn**."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "1ff6b136",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "from sklearn.model_selection import train_test_split\n",
+    "from sklearn.datasets import make_moons\n",
+    "\n",
+    "X, y = make_moons(n_samples=500, noise=0.30, random_state=42)\n",
+    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)\n",
+    "\n",
+    "from sklearn.ensemble import RandomForestClassifier\n",
+    "from sklearn.ensemble import VotingClassifier\n",
+    "from sklearn.linear_model import LogisticRegression\n",
+    "from sklearn.svm import SVC\n",
+    "\n",
+    "log_clf = LogisticRegression(solver=\"liblinear\", random_state=42)\n",
+    "rnd_clf = RandomForestClassifier(n_estimators=10, random_state=42)\n",
+    "svm_clf = SVC(gamma=\"auto\", random_state=42)\n",
+    "\n",
+    "voting_clf = VotingClassifier(\n",
+    "    estimators=[('lr', log_clf), ('rf', rnd_clf), ('svc', svm_clf)],\n",
+    "    voting='hard')\n",
+    "\n",
+    "voting_clf.fit(X_train, y_train)\n",
+    "\n",
+    "from sklearn.metrics import accuracy_score\n",
+    "\n",
+    "for clf in (log_clf, rnd_clf, svm_clf, voting_clf):\n",
+    "    clf.fit(X_train, y_train)\n",
+    "    y_pred = clf.predict(X_test)\n",
+    "    print(clf.__class__.__name__, accuracy_score(y_test, y_pred))\n",
+    "\n",
+    "log_clf = LogisticRegression(solver=\"liblinear\", random_state=42)\n",
+    "rnd_clf = RandomForestClassifier(n_estimators=10, random_state=42)\n",
+    "svm_clf = SVC(gamma=\"auto\", probability=True, random_state=42)\n",
+    "voting_clf = VotingClassifier(\n",
+    "    estimators=[('lr', log_clf), ('rf', rnd_clf), ('svc', svm_clf)],\n",
+    "    voting='soft')\n",
+    "voting_clf.fit(X_train, y_train)\n",
+    "\n",
+    "from sklearn.metrics import accuracy_score\n",
+    "\n",
+    "for clf in (log_clf, rnd_clf, svm_clf, voting_clf):\n",
+    "    clf.fit(X_train, y_train)\n",
+    "    y_pred = clf.predict(X_test)\n",
+    "    print(clf.__class__.__name__, accuracy_score(y_test, y_pred))"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "ec19b899",
+   "id": "baffd70d",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Voting and Bagging"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "2a1f0f0d",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "from sklearn.model_selection import train_test_split\n",
+    "from sklearn.datasets import make_moons\n",
+    "\n",
+    "X, y = make_moons(n_samples=500, noise=0.30, random_state=42)\n",
+    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)\n",
+    "from sklearn.ensemble import RandomForestClassifier\n",
+    "from sklearn.ensemble import VotingClassifier\n",
+    "from sklearn.linear_model import LogisticRegression\n",
+    "from sklearn.svm import SVC\n",
+    "\n",
+    "log_clf = LogisticRegression(random_state=42)\n",
+    "rnd_clf = RandomForestClassifier(random_state=42)\n",
+    "svm_clf = SVC(random_state=42)\n",
+    "\n",
+    "voting_clf = VotingClassifier(\n",
+    "    estimators=[('lr', log_clf), ('rf', rnd_clf), ('svc', svm_clf)],\n",
+    "    voting='hard')\n",
+    "voting_clf.fit(X_train, y_train)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "81cfb49c",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "from sklearn.metrics import accuracy_score\n",
+    "\n",
+    "for clf in (log_clf, rnd_clf, svm_clf, voting_clf):\n",
+    "    clf.fit(X_train, y_train)\n",
+    "    y_pred = clf.predict(X_test)\n",
+    "    print(clf.__class__.__name__, accuracy_score(y_test, y_pred))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "id": "ce146db7",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "log_clf = LogisticRegression(random_state=42)\n",
+    "rnd_clf = RandomForestClassifier(random_state=42)\n",
+    "svm_clf = SVC(probability=True, random_state=42)\n",
+    "\n",
+    "voting_clf = VotingClassifier(\n",
+    "    estimators=[('lr', log_clf), ('rf', rnd_clf), ('svc', svm_clf)],\n",
+    "    voting='soft')\n",
+    "voting_clf.fit(X_train, y_train)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "id": "3e5e5740",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "from sklearn.metrics import accuracy_score\n",
+    "\n",
+    "for clf in (log_clf, rnd_clf, svm_clf, voting_clf):\n",
+    "    clf.fit(X_train, y_train)\n",
+    "    y_pred = clf.predict(X_test)\n",
+    "    print(clf.__class__.__name__, accuracy_score(y_test, y_pred))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "17506d34",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Bagging\n",
+    "\n",
+    "The **plain** decision trees suffer from high\n",
+    "variance. This means that if we split the training data into two parts\n",
+    "at random, and fit a decision tree to both halves, the results that we\n",
+    "get could be quite different. In contrast, a procedure with low\n",
+    "variance will yield similar results if applied repeatedly to distinct\n",
+    "data sets; linear regression tends to have low variance, if the ratio\n",
+    "of $n$ to $p$ is moderately large. \n",
+    "\n",
+    "**Bootstrap aggregation**, or just **bagging**, is a\n",
+    "general-purpose procedure for reducing the variance of a statistical\n",
+    "learning method."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "9bf49762",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## More bagging\n",
+    "\n",
+    "Bagging typically results in improved accuracy\n",
+    "over prediction using a single tree. Unfortunately, however, it can be\n",
+    "difficult to interpret the resulting model. Recall that one of the\n",
+    "advantages of decision trees is the attractive and easily interpreted\n",
+    "diagram that results.\n",
+    "\n",
+    "However, when we bag a large number of trees, it is no longer\n",
+    "possible to represent the resulting statistical learning procedure\n",
+    "using a single tree, and it is no longer clear which variables are\n",
+    "most important to the procedure. Thus, bagging improves prediction\n",
+    "accuracy at the expense of interpretability.  Although the collection\n",
+    "of bagged trees is much more difficult to interpret than a single\n",
+    "tree, one can obtain an overall summary of the importance of each\n",
+    "predictor using the MSE (for bagging regression trees) or the Gini\n",
+    "index (for bagging classification trees). In the case of bagging\n",
+    "regression trees, we can record the total amount that the MSE is\n",
+    "decreased due to splits over a given predictor, averaged over all $B$ possible\n",
+    "trees. A large value indicates an important predictor. Similarly, in\n",
+    "the context of bagging classification trees, we can add up the total\n",
+    "amount that the Gini index  is decreased by splits over a given\n",
+    "predictor, averaged over all $B$ trees."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6f765d43",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Making your own Bootstrap: Changing the Level of the Decision Tree\n",
+    "\n",
+    "Let us bring up our good old boostrap example from the linear regression lectures. We change the linerar regression algorithm with\n",
+    "a decision tree wth different depths and perform a bootstrap aggregate (in this case we perform as many bootstraps as data points $n$)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "95bad7a4",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "\n",
+    "import matplotlib.pyplot as plt\n",
+    "import numpy as np\n",
+    "from sklearn.model_selection import train_test_split\n",
+    "from sklearn.pipeline import make_pipeline\n",
+    "from sklearn.utils import resample\n",
+    "from sklearn.tree import DecisionTreeRegressor\n",
+    "\n",
+    "n = 100\n",
+    "n_boostraps = 100\n",
+    "maxdepth = 8\n",
+    "\n",
+    "# Make data set.\n",
+    "x = np.linspace(-3, 3, n).reshape(-1, 1)\n",
+    "y = np.exp(-x**2) + 1.5 * np.exp(-(x-2)**2)+ np.random.normal(0, 0.1, x.shape)\n",
+    "error = np.zeros(maxdepth)\n",
+    "bias = np.zeros(maxdepth)\n",
+    "variance = np.zeros(maxdepth)\n",
+    "polydegree = np.zeros(maxdepth)\n",
+    "X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.2)\n",
+    "\n",
+    "from sklearn.preprocessing import StandardScaler\n",
+    "scaler = StandardScaler()\n",
+    "scaler.fit(X_train)\n",
+    "X_train_scaled = scaler.transform(X_train)\n",
+    "X_test_scaled = scaler.transform(X_test)\n",
+    "\n",
+    "# we produce a simple tree first as benchmark\n",
+    "simpletree = DecisionTreeRegressor(max_depth=3) \n",
+    "simpletree.fit(X_train_scaled, y_train)\n",
+    "simpleprediction = simpletree.predict(X_test_scaled)\n",
+    "for degree in range(1,maxdepth):\n",
+    "    model = DecisionTreeRegressor(max_depth=degree) \n",
+    "    y_pred = np.empty((y_test.shape[0], n_boostraps))\n",
+    "    for i in range(n_boostraps):\n",
+    "        x_, y_ = resample(X_train_scaled, y_train)\n",
+    "        model.fit(x_, y_)\n",
+    "        y_pred[:, i] = model.predict(X_test_scaled)#.ravel()\n",
+    "\n",
+    "    polydegree[degree] = degree\n",
+    "    error[degree] = np.mean( np.mean((y_test - y_pred)**2, axis=1, keepdims=True) )\n",
+    "    bias[degree] = np.mean( (y_test - np.mean(y_pred, axis=1, keepdims=True))**2 )\n",
+    "    variance[degree] = np.mean( np.var(y_pred, axis=1, keepdims=True) )\n",
+    "    print('Polynomial degree:', degree)\n",
+    "    print('Error:', error[degree])\n",
+    "    print('Bias^2:', bias[degree])\n",
+    "    print('Var:', variance[degree])\n",
+    "    print('{} >= {} + {} = {}'.format(error[degree], bias[degree], variance[degree], bias[degree]+variance[degree]))\n",
+    " \n",
+    "mse_simpletree= np.mean( np.mean((y_test - simpleprediction)**2))\n",
+    "print(\"Simple tree:\",mse_simpletree)\n",
+    "plt.xlim(1,maxdepth)\n",
+    "plt.plot(polydegree, error, label='MSE')\n",
+    "plt.plot(polydegree, bias, label='bias')\n",
+    "plt.plot(polydegree, variance, label='Variance')\n",
+    "plt.legend()\n",
+    "save_fig(\"baggingboot\")\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ce1ddd84",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Random forests\n",
+    "\n",
+    "Random forests provide an improvement over bagged trees by way of a\n",
+    "small tweak that decorrelates the trees. \n",
+    "\n",
+    "As in bagging, we build a\n",
+    "number of decision trees on bootstrapped training samples. But when\n",
+    "building these decision trees, each time a split in a tree is\n",
+    "considered, a random sample of $m$ predictors is chosen as split\n",
+    "candidates from the full set of $p$ predictors. The split is allowed to\n",
+    "use only one of those $m$ predictors. \n",
+    "\n",
+    "A fresh sample of $m$ predictors is\n",
+    "taken at each split, and typically we choose"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d26e0049",
    "metadata": {
     "editable": true
    },
    "source": [
     "$$\n",
-    "\\frac{1}{2} \\boldsymbol{\\lambda}^T\\begin{bmatrix} y_1y_1K(\\boldsymbol{x}_1,\\boldsymbol{x}_1) & y_1y_2K(\\boldsymbol{x}_1,\\boldsymbol{x}_2) & \\dots & \\dots & y_1y_nK(\\boldsymbol{x}_1,\\boldsymbol{x}_n) \\\\\n",
-    "y_2y_1K(\\boldsymbol{x}_2,\\boldsymbol{x}_1) & y_2y_2K(\\boldsymbol{x}_2,\\boldsymbol{x}_2) & \\dots & \\dots & y_1y_nK(\\boldsymbol{x}_2,\\boldsymbol{x}_n) \\\\\n",
-    "\\dots & \\dots & \\dots & \\dots & \\dots \\\\\n",
-    "\\dots & \\dots & \\dots & \\dots & \\dots \\\\\n",
-    "y_ny_1K(\\boldsymbol{x}_n,\\boldsymbol{x}_1) & y_ny_2K(\\boldsymbol{x}_n\\boldsymbol{x}_2) & \\dots & \\dots & y_ny_nK(\\boldsymbol{x}_n,\\boldsymbol{x}_n) \\\\\n",
-    "\\end{bmatrix}\\boldsymbol{\\lambda}-\\mathbb{I}\\boldsymbol{\\lambda},\n",
+    "m\\approx \\sqrt{p}.\n",
     "$$"
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "25eb131c",
+   "id": "fa8e6b26",
    "metadata": {
     "editable": true
    },
    "source": [
-    "subject to $\\boldsymbol{y}^T\\boldsymbol{\\lambda}=0$. Here we defined the vectors $\\boldsymbol{\\lambda} =[\\lambda_1,\\lambda_2,\\dots,\\lambda_n]$ and \n",
-    "$\\boldsymbol{y}=[y_1,y_2,\\dots,y_n]$. \n",
-    "With  the slack constants this leads to the additional constraint $0\\leq \\lambda_i \\leq C$.\n",
+    "In building a random forest, at\n",
+    "each split in the tree, the algorithm is not even allowed to consider\n",
+    "a majority of the available predictors. \n",
     "\n",
-    "Using the **CVXOPT** library, the matrix $P$ would then be defined by the above matrix while the KKT conditions would all be collected by the matrix $G$."
+    "The reason for this is rather clever. Suppose that there is one very\n",
+    "strong predictor in the data set, along with a number of other\n",
+    "moderately strong predictors. Then in the collection of bagged\n",
+    "variable importance random forest trees, most or all of the trees will\n",
+    "use this strong predictor in the top split. Consequently, all of the\n",
+    "bagged trees will look quite similar to each other. Hence the\n",
+    "predictions from the bagged trees will be highly correlated.\n",
+    "Unfortunately, averaging many highly correlated quantities does not\n",
+    "lead to as large of a reduction in variance as averaging many\n",
+    "uncorrelated quantities. In particular, this means that bagging will\n",
+    "not lead to a substantial reduction in variance over a single tree in\n",
+    "this setting."
    ]
   },
   {
    "cell_type": "markdown",
-   "id": "b3377017",
+   "id": "c4a12917",
    "metadata": {
     "editable": true
    },
    "source": [
-    "## Support vector machines for regression\n",
+    "## Random Forest Algorithm\n",
+    "The algorithm described here can be applied to both classification and regression problems.\n",
     "\n",
-    "Material will be added here."
+    "We will grow of forest of say $B$ trees.\n",
+    "1. For $b=1:B$\n",
+    "\n",
+    "  * Draw a bootstrap sample from the training data organized in our $\\boldsymbol{X}$ matrix.\n",
+    "\n",
+    "  * We grow then a random forest tree $T_b$ based on the bootstrapped data by repeating the steps outlined till we reach the maximum node size is reached\n",
+    "\n",
+    "1. we select $m \\le p$ variables at random from the $p$ predictors/features\n",
+    "\n",
+    "2. pick the best split point among the $m$ features using for example the CART algorithm and create a new node\n",
+    "\n",
+    "3. split the node into daughter nodes\n",
+    "\n",
+    "4. Output then the ensemble of trees $\\{T_b\\}_1^{B}$ and make predictions for either a regression type of problem or a classification type of problem."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e115121c",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Random Forests Compared with other Methods on the Cancer Data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "141a1555",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "import matplotlib.pyplot as plt\n",
+    "import numpy as np\n",
+    "from sklearn.model_selection import  train_test_split \n",
+    "from sklearn.datasets import load_breast_cancer\n",
+    "from sklearn.svm import SVC\n",
+    "from sklearn.linear_model import LogisticRegression\n",
+    "from sklearn.tree import DecisionTreeClassifier\n",
+    "from sklearn.ensemble import BaggingClassifier\n",
+    "\n",
+    "# Load the data\n",
+    "cancer = load_breast_cancer()\n",
+    "\n",
+    "X_train, X_test, y_train, y_test = train_test_split(cancer.data,cancer.target,random_state=0)\n",
+    "print(X_train.shape)\n",
+    "print(X_test.shape)\n",
+    "#define methods\n",
+    "# Logistic Regression\n",
+    "logreg = LogisticRegression(solver='lbfgs')\n",
+    "# Support vector machine\n",
+    "svm = SVC(gamma='auto', C=100)\n",
+    "# Decision Trees\n",
+    "deep_tree_clf = DecisionTreeClassifier(max_depth=None)\n",
+    "#Scale the data\n",
+    "from sklearn.preprocessing import StandardScaler\n",
+    "scaler = StandardScaler()\n",
+    "scaler.fit(X_train)\n",
+    "X_train_scaled = scaler.transform(X_train)\n",
+    "X_test_scaled = scaler.transform(X_test)\n",
+    "# Logistic Regression\n",
+    "logreg.fit(X_train_scaled, y_train)\n",
+    "print(\"Test set accuracy Logistic Regression with scaled data: {:.2f}\".format(logreg.score(X_test_scaled,y_test)))\n",
+    "# Support Vector Machine\n",
+    "svm.fit(X_train_scaled, y_train)\n",
+    "print(\"Test set accuracy SVM with scaled data: {:.2f}\".format(logreg.score(X_test_scaled,y_test)))\n",
+    "# Decision Trees\n",
+    "deep_tree_clf.fit(X_train_scaled, y_train)\n",
+    "print(\"Test set accuracy with Decision Trees and scaled data: {:.2f}\".format(deep_tree_clf.score(X_test_scaled,y_test)))\n",
+    "\n",
+    "\n",
+    "from sklearn.ensemble import RandomForestClassifier\n",
+    "from sklearn.preprocessing import LabelEncoder\n",
+    "from sklearn.model_selection import cross_validate\n",
+    "# Data set not specificied\n",
+    "#Instantiate the model with 500 trees and entropy as splitting criteria\n",
+    "Random_Forest_model = RandomForestClassifier(n_estimators=500,criterion=\"entropy\")\n",
+    "Random_Forest_model.fit(X_train_scaled, y_train)\n",
+    "#Cross validation\n",
+    "accuracy = cross_validate(Random_Forest_model,X_test_scaled,y_test,cv=10)['test_score']\n",
+    "print(accuracy)\n",
+    "print(\"Test set accuracy with Random Forests and scaled data: {:.2f}\".format(Random_Forest_model.score(X_test_scaled,y_test)))\n",
+    "\n",
+    "\n",
+    "import scikitplot as skplt\n",
+    "y_pred = Random_Forest_model.predict(X_test_scaled)\n",
+    "skplt.metrics.plot_confusion_matrix(y_test, y_pred, normalize=True)\n",
+    "plt.show()\n",
+    "y_probas = Random_Forest_model.predict_proba(X_test_scaled)\n",
+    "skplt.metrics.plot_roc(y_test, y_probas)\n",
+    "plt.show()\n",
+    "skplt.metrics.plot_cumulative_gain(y_test, y_probas)\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "10a44818",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "Recall that the cumulative gains curve shows the percentage of the\n",
+    "overall number of cases in a given category *gained* by targeting a\n",
+    "percentage of the total number of cases.\n",
+    "\n",
+    "Similarly, the receiver operating characteristic curve, or ROC curve,\n",
+    "displays the diagnostic ability of a binary classifier system as its\n",
+    "discrimination threshold is varied. It plots the true positive rate against the false positive rate."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "bfcef20a",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Compare  Bagging on Trees with Random Forests"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "76fd8c7c",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "bag_clf = BaggingClassifier(\n",
+    "    DecisionTreeClassifier(splitter=\"random\", max_leaf_nodes=16, random_state=42),\n",
+    "    n_estimators=500, max_samples=1.0, bootstrap=True, n_jobs=-1, random_state=42)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "id": "7c8a970d",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "bag_clf.fit(X_train, y_train)\n",
+    "y_pred = bag_clf.predict(X_test)\n",
+    "from sklearn.ensemble import RandomForestClassifier\n",
+    "rnd_clf = RandomForestClassifier(n_estimators=500, max_leaf_nodes=16, n_jobs=-1, random_state=42)\n",
+    "rnd_clf.fit(X_train, y_train)\n",
+    "y_pred_rf = rnd_clf.predict(X_test)\n",
+    "np.sum(y_pred == y_pred_rf) / len(y_pred)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "9dcc9487",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Boosting, a Bird's Eye View\n",
+    "\n",
+    "The basic idea is to combine weak classifiers in order to create a good\n",
+    "classifier. With a weak classifier we often intend a classifier which\n",
+    "produces results which are only slightly better than we would get by\n",
+    "random guesses.\n",
+    "\n",
+    "This is done by applying in an iterative way a weak (or a standard\n",
+    "classifier like decision trees) to modify the data. In each iteration\n",
+    "we emphasize those observations which are misclassified by weighting\n",
+    "them with a factor."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "79076633",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## What is boosting? Additive Modelling/Iterative Fitting\n",
+    "\n",
+    "Boosting is a way of fitting an additive expansion in a set of\n",
+    "elementary basis functions like for example some simple polynomials.\n",
+    "Assume for example that we have a function"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cf4c50af",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "f_M(x) = \\sum_{i=1}^M \\beta_m b(x;\\gamma_m),\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "942127e6",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "where $\\beta_m$ are the expansion parameters to be determined in a\n",
+    "minimization process and $b(x;\\gamma_m)$ are some simple functions of\n",
+    "the multivariable parameter $x$ which is characterized by the\n",
+    "parameters $\\gamma_m$.\n",
+    "\n",
+    "As an example, consider the Sigmoid function we used in logistic\n",
+    "regression. In that case, we can translate the function\n",
+    "$b(x;\\gamma_m)$ into the Sigmoid function"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "52015fd6",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\sigma(t) = \\frac{1}{1+\\exp{(-t)}},\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f4963135",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "where $t=\\gamma_0+\\gamma_1 x$ and the parameters $\\gamma_0$ and\n",
+    "$\\gamma_1$ were determined by the Logistic Regression fitting\n",
+    "algorithm.\n",
+    "\n",
+    "As another example, consider the cost function we defined for linear regression"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "84a3f642",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "C(\\boldsymbol{y},\\boldsymbol{f}) = \\frac{1}{n} \\sum_{i=0}^{n-1}(y_i-f(x_i))^2.\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "d6b120ad",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "In this case the function $f(x)$ was replaced by the design matrix\n",
+    "$\\boldsymbol{X}$ and the unknown linear regression parameters $\\boldsymbol{\\beta}$,\n",
+    "that is $\\boldsymbol{f}=\\boldsymbol{X}\\boldsymbol{\\beta}$. In linear regression we can \n",
+    "simply invert a matrix and obtain the parameters $\\beta$ by"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3057a800",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\boldsymbol{\\beta}=\\left(\\boldsymbol{X}^T\\boldsymbol{X}\\right)^{-1}\\boldsymbol{X}^T\\boldsymbol{y}.\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "9ce41841",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "In iterative fitting or additive modeling, we minimize the cost function with respect to the parameters $\\beta_m$ and $\\gamma_m$."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e629d4f9",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Iterative Fitting, Regression and Squared-error Cost Function\n",
+    "\n",
+    "The way we proceed is as follows (here we specialize to the squared-error cost function)\n",
+    "\n",
+    "1. Establish a cost function, here $C(\\boldsymbol{y},\\boldsymbol{f}) = \\frac{1}{n} \\sum_{i=0}^{n-1}(y_i-f_M(x_i))^2$ with $f_M(x) = \\sum_{i=1}^M \\beta_m b(x;\\gamma_m)$.\n",
+    "\n",
+    "2. Initialize with a guess $f_0(x)$. It could be one or even zero or some random numbers.\n",
+    "\n",
+    "3. For $m=1:M$\n",
+    "\n",
+    "a. minimize $\\sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\\beta b(x;\\gamma))^2$ wrt $\\gamma$ and $\\beta$\n",
+    "\n",
+    "b. This gives the optimal values $\\beta_m$ and $\\gamma_m$\n",
+    "\n",
+    "c. Determine then the new values $f_m(x)=f_{m-1}(x) +\\beta_m b(x;\\gamma_m)$\n",
+    "\n",
+    "We could use any of the algorithms we have discussed till now. If we\n",
+    "use trees, $\\gamma$ parameterizes the split variables and split points\n",
+    "at the internal nodes, and the predictions at the terminal nodes."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4c14092d",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Squared-Error Example and Iterative Fitting\n",
+    "\n",
+    "To better understand what happens, let us develop the steps for the iterative fitting using the above squared error function.\n",
+    "\n",
+    "For simplicity we assume also that our functions $b(x;\\gamma)=1+\\gamma x$. \n",
+    "\n",
+    "This means that for every iteration $m$, we need to optimize"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "9119c6f3",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "(\\beta_m,\\gamma_m) = \\mathrm{argmin}_{\\beta,\\lambda}\\hspace{0.1cm} \\sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\\beta b(x;\\gamma))^2=\\sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\\beta(1+\\gamma x_i))^2.\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "60eb2f52",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "We start our iteration by simply setting $f_0(x)=0$. \n",
+    "Taking the derivatives  with respect to $\\beta$ and $\\gamma$ we obtain"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5f499f04",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\frac{\\partial {\\cal C}}{\\partial \\beta} = -2\\sum_{i}(1+\\gamma x_i)(y_i-\\beta(1+\\gamma x_i))=0,\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5bb9a88f",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "and"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "bdb8a471",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\frac{\\partial {\\cal C}}{\\partial \\gamma} =-2\\sum_{i}\\beta x_i(y_i-\\beta(1+\\gamma x_i))=0.\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "9125265b",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "We can then rewrite these equations as (defining $\\boldsymbol{w}=\\boldsymbol{e}+\\gamma \\boldsymbol{x})$ with $\\boldsymbol{e}$ being the unit vector)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f9bcdf59",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\gamma \\boldsymbol{w}^T(\\boldsymbol{y}-\\beta\\gamma \\boldsymbol{w})=0,\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f64d41da",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "which gives us $\\beta = \\boldsymbol{w}^T\\boldsymbol{y}/(\\boldsymbol{w}^T\\boldsymbol{w})$. Similarly we have"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "60380bf4",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\beta\\gamma \\boldsymbol{x}^T(\\boldsymbol{y}-\\beta(1+\\gamma \\boldsymbol{x}))=0,\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "fb6cd261",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "which leads to $\\gamma =(\\boldsymbol{x}^T\\boldsymbol{y}-\\beta\\boldsymbol{x}^T\\boldsymbol{e})/(\\beta\\boldsymbol{x}^T\\boldsymbol{x})$.  Inserting\n",
+    "for $\\beta$ gives us an equation for $\\gamma$. This is a non-linear equation in the unknown $\\gamma$ and has to be solved numerically. \n",
+    "\n",
+    "The solution to these two equations gives us in turn $\\beta_1$ and $\\gamma_1$ leading to the new expression for $f_1(x)$ as\n",
+    "$f_1(x) = \\beta_1(1+\\gamma_1x)$. Doing this $M$ times results in our final estimate for the function $f$."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "19fa1bc1",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Iterative Fitting, Classification and AdaBoost\n",
+    "\n",
+    "Let us consider a binary classification problem with two outcomes $y_i \\in \\{-1,1\\}$ and $i=0,1,2,\\dots,n-1$ as our set of\n",
+    "observations. We define a classification function $G(x)$ which produces a prediction taking one or the other of the two values \n",
+    "$\\{-1,1\\}$.\n",
+    "\n",
+    "The error rate of the training sample is then"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "14869eaf",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\mathrm{\\overline{err}}=\\frac{1}{n} \\sum_{i=0}^{n-1} I(y_i\\ne G(x_i)).\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "b67a64c3",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "The iterative procedure starts with defining a weak classifier whose\n",
+    "error rate is barely better than random guessing.  The iterative\n",
+    "procedure in boosting is to sequentially apply a  weak\n",
+    "classification algorithm to repeatedly modified versions of the data\n",
+    "producing a sequence of weak classifiers $G_m(x)$.\n",
+    "\n",
+    "Here we will express our  function $f(x)$ in terms of $G(x)$. That is"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "59aca04f",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "f_M(x) = \\sum_{i=1}^M \\beta_m b(x;\\gamma_m),\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "95c0f671",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "will be a function of"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "faf8e906",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "G_M(x) = \\mathrm{sign} \\sum_{i=1}^M \\alpha_m G_m(x).\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "fdef01e7",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Adaptive Boosting, AdaBoost\n",
+    "\n",
+    "In our iterative procedure we define thus"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "83e67d70",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "f_m(x) = f_{m-1}(x)+\\beta_mG_m(x).\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3c167dc0",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "The simplest possible cost function which leads (also simple from a computational point of view) to the AdaBoost algorithm is the\n",
+    "exponential cost/loss function defined as"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a542df10",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "C(\\boldsymbol{y},\\boldsymbol{f}) = \\sum_{i=0}^{n-1}\\exp{(-y_i(f_{m-1}(x_i)+\\beta G(x_i))}.\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "048606f2",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "We optimize $\\beta$ and $G$ for each value of $m=1:M$ as we did in the regression case.\n",
+    "This is normally done in two steps. Let us however first rewrite the cost function as"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "0849517c",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "C(\\boldsymbol{y},\\boldsymbol{f}) = \\sum_{i=0}^{n-1}w_i^{m}\\exp{(-y_i\\beta G(x_i))},\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "22994ad7",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "where we have defined $w_i^m= \\exp{(-y_if_{m-1}(x_i))}$."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "ba76b5fa",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Building up AdaBoost\n",
+    "\n",
+    "First, for any $\\beta > 0$, we optimize $G$ by setting"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "dfa03d62",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "G_m(x) = \\mathrm{sign} \\sum_{i=0}^{n-1} w_i^m I(y_i \\ne G_(x_i)),\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7f8af900",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "which is the classifier that minimizes the weighted error rate in predicting $y$.\n",
+    "\n",
+    "We can do this by rewriting"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "c7bc96a3",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\exp{-(\\beta)}\\sum_{y_i=G(x_i)}w_i^m+\\exp{(\\beta)}\\sum_{y_i\\ne G(x_i)}w_i^m,\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3dd711b0",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "which can be rewritten as"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "362ce9e5",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "(\\exp{(\\beta)}-\\exp{-(\\beta)})\\sum_{i=0}^{n-1}w_i^mI(y_i\\ne G(x_i))+\\exp{(-\\beta)}\\sum_{i=0}^{n-1}w_i^m=0,\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f90182c3",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "which leads to"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "785bd7b4",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\beta_m = \\frac{1}{2}\\log{\\frac{1-\\mathrm{\\overline{err}}}{\\mathrm{\\overline{err}}}},\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "c84248ab",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "where we have redefined the error as"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "bd2cfd71",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\mathrm{\\overline{err}}_m=\\frac{1}{n}\\frac{\\sum_{i=0}^{n-1}w_i^mI(y_i\\ne G(x_i)}{\\sum_{i=0}^{n-1}w_i^m},\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5e30a895",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "which leads to an update of"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "44faac5f",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "f_m(x) = f_{m-1}(x) +\\beta_m G_m(x).\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "7198891c",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "This leads to the new weights"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "4f8608c6",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "w_i^{m+1} = w_i^m \\exp{(-y_i\\beta_m G_m(x_i))}\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "961207e1",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Adaptive boosting: AdaBoost, Basic Algorithm\n",
+    "\n",
+    "The algorithm here is rather straightforward. Assume that our weak\n",
+    "classifier is a decision tree and we consider a binary set of outputs\n",
+    "with $y_i \\in \\{-1,1\\}$ and $i=0,1,2,\\dots,n-1$ as our set of\n",
+    "observations. Our design matrix is given in terms of the\n",
+    "feature/predictor vectors\n",
+    "$\\boldsymbol{X}=[\\boldsymbol{x}_0\\boldsymbol{x}_1\\dots\\boldsymbol{x}_{p-1}]$. Finally, we define also a\n",
+    "classifier determined by our data via a function $G(x)$. This function tells us how well we are able to classify our outputs/targets $\\boldsymbol{y}$. \n",
+    "\n",
+    "We have already defined the misclassification error $\\mathrm{err}$ as"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "1ba60f77",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\mathrm{err}=\\frac{1}{n}\\sum_{i=0}^{n-1}I(y_i\\ne G(x_i)),\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f0c95019",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "where the function $I()$ is one if we misclassify and zero if we classify correctly."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "6a3c8c3e",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## Basic Steps of AdaBoost\n",
+    "\n",
+    "With the above definitions we are now ready to set up the algorithm for AdaBoost.\n",
+    "The basic idea is to set up weights which will be used to scale the correctly classified and the misclassified cases.\n",
+    "1. We start by initializing all weights to $w_i = 1/n$, with $i=0,1,2,\\dots n-1$. It is easy to see that we must have $\\sum_{i=0}^{n-1}w_i = 1$.\n",
+    "\n",
+    "2. We rewrite the misclassification error as"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a7695af1",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "$$\n",
+    "\\mathrm{\\overline{err}}_m=\\frac{\\sum_{i=0}^{n-1}w_i^m I(y_i\\ne G(x_i))}{\\sum_{i=0}^{n-1}w_i},\n",
+    "$$"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "da3b7d37",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "1. Then we start looping over all attempts at classifying, namely we start an iterative process for $m=1:M$, where $M$ is the final number of classifications. Our given classifier could for example be a plain decision tree.\n",
+    "\n",
+    "a. Fit then a given classifier to the training set using the weights $w_i$.\n",
+    "\n",
+    "b. Compute then $\\mathrm{err}$ and figure out which events are classified properly and which are classified wrongly.\n",
+    "\n",
+    "c. Define a quantity $\\alpha_{m} = \\log{(1-\\mathrm{\\overline{err}}_m)/\\mathrm{\\overline{err}}_m}$\n",
+    "\n",
+    "d. Set the new weights to $w_i = w_i\\times \\exp{(\\alpha_m I(y_i\\ne G(x_i)}$.\n",
+    "\n",
+    "5. Compute the new classifier $G(x)= \\sum_{i=0}^{n-1}\\alpha_m I(y_i\\ne G(x_i)$.\n",
+    "\n",
+    "For the iterations with $m \\le 2$ the weights are modified\n",
+    "individually at each steps. The observations which were misclassified\n",
+    "at iteration $m-1$ have a weight which is larger than those which were\n",
+    "classified properly. As this proceeds, the observations which were\n",
+    "difficult to classifiy correctly are given a larger influence. Each\n",
+    "new classification step $m$ is then forced to concentrate on those\n",
+    "observations that are missed in the previous iterations."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "0194b994",
+   "metadata": {
+    "editable": true
+   },
+   "source": [
+    "## AdaBoost Examples\n",
+    "\n",
+    "Using **Scikit-Learn** it is easy to apply the adaptive boosting algorithm, as done here."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "id": "31180ced",
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
+   "outputs": [],
+   "source": [
+    "from sklearn.ensemble import AdaBoostClassifier\n",
+    "\n",
+    "ada_clf = AdaBoostClassifier(\n",
+    "    DecisionTreeClassifier(max_depth=1), n_estimators=200,\n",
+    "    algorithm=\"SAMME.R\", learning_rate=0.5, random_state=42)\n",
+    "ada_clf.fit(X_train, y_train)\n",
+    "\n",
+    "from sklearn.ensemble import AdaBoostClassifier\n",
+    "\n",
+    "ada_clf = AdaBoostClassifier(\n",
+    "    DecisionTreeClassifier(max_depth=1), n_estimators=200,\n",
+    "    algorithm=\"SAMME.R\", learning_rate=0.5, random_state=42)\n",
+    "ada_clf.fit(X_train_scaled, y_train)\n",
+    "y_pred = ada_clf.predict(X_test_scaled)\n",
+    "skplt.metrics.plot_confusion_matrix(y_test, y_pred, normalize=True)\n",
+    "plt.show()\n",
+    "y_probas = ada_clf.predict_proba(X_test_scaled)\n",
+    "skplt.metrics.plot_roc(y_test, y_probas)\n",
+    "plt.show()\n",
+    "skplt.metrics.plot_cumulative_gain(y_test, y_probas)\n",
+    "plt.show()"
    ]
   }
  ],
diff --git a/doc/src/week46/DataFiles/bank.csv b/doc/src/week46/DataFiles/bank.csv
new file mode 100644
index 000000000..930337395
--- /dev/null
+++ b/doc/src/week46/DataFiles/bank.csv
@@ -0,0 +1,1372 @@
+3.6216,8.6661,-2.8073,-0.44699,0
+4.5459,8.1674,-2.4586,-1.4621,0
+3.866,-2.6383,1.9242,0.10645,0
+3.4566,9.5228,-4.0112,-3.5944,0
+0.32924,-4.4552,4.5718,-0.9888,0
+4.3684,9.6718,-3.9606,-3.1625,0
+3.5912,3.0129,0.72888,0.56421,0
+2.0922,-6.81,8.4636,-0.60216,0
+3.2032,5.7588,-0.75345,-0.61251,0
+1.5356,9.1772,-2.2718,-0.73535,0
+1.2247,8.7779,-2.2135,-0.80647,0
+3.9899,-2.7066,2.3946,0.86291,0
+1.8993,7.6625,0.15394,-3.1108,0
+-1.5768,10.843,2.5462,-2.9362,0
+3.404,8.7261,-2.9915,-0.57242,0
+4.6765,-3.3895,3.4896,1.4771,0
+2.6719,3.0646,0.37158,0.58619,0
+0.80355,2.8473,4.3439,0.6017,0
+1.4479,-4.8794,8.3428,-2.1086,0
+5.2423,11.0272,-4.353,-4.1013,0
+5.7867,7.8902,-2.6196,-0.48708,0
+0.3292,-4.4552,4.5718,-0.9888,0
+3.9362,10.1622,-3.8235,-4.0172,0
+0.93584,8.8855,-1.6831,-1.6599,0
+4.4338,9.887,-4.6795,-3.7483,0
+0.7057,-5.4981,8.3368,-2.8715,0
+1.1432,-3.7413,5.5777,-0.63578,0
+-0.38214,8.3909,2.1624,-3.7405,0
+6.5633,9.8187,-4.4113,-3.2258,0
+4.8906,-3.3584,3.4202,1.0905,0
+-0.24811,-0.17797,4.9068,0.15429,0
+1.4884,3.6274,3.308,0.48921,0
+4.2969,7.617,-2.3874,-0.96164,0
+-0.96511,9.4111,1.7305,-4.8629,0
+-1.6162,0.80908,8.1628,0.60817,0
+2.4391,6.4417,-0.80743,-0.69139,0
+2.6881,6.0195,-0.46641,-0.69268,0
+3.6289,0.81322,1.6277,0.77627,0
+4.5679,3.1929,-2.1055,0.29653,0
+3.4805,9.7008,-3.7541,-3.4379,0
+4.1711,8.722,-3.0224,-0.59699,0
+-0.2062,9.2207,-3.7044,-6.8103,0
+-0.0068919,9.2931,-0.41243,-1.9638,0
+0.96441,5.8395,2.3235,0.066365,0
+2.8561,6.9176,-0.79372,0.48403,0
+-0.7869,9.5663,-3.7867,-7.5034,0
+2.0843,6.6258,0.48382,-2.2134,0
+-0.7869,9.5663,-3.7867,-7.5034,0
+3.9102,6.065,-2.4534,-0.68234,0
+1.6349,3.286,2.8753,0.087054,0
+4.3239,-4.8835,3.4356,-0.5776,0
+5.262,3.9834,-1.5572,1.0103,0
+3.1452,5.825,-0.51439,-1.4944,0
+2.549,6.1499,-1.1605,-1.2371,0
+4.9264,5.496,-2.4774,-0.50648,0
+4.8265,0.80287,1.6371,1.1875,0
+2.5635,6.7769,-0.61979,0.38576,0
+5.807,5.0097,-2.2384,0.43878,0
+3.1377,-4.1096,4.5701,0.98963,0
+-0.78289,11.3603,-0.37644,-7.0495,0
+2.888,0.44696,4.5907,-0.24398,0
+0.49665,5.527,1.7785,-0.47156,0
+4.2586,11.2962,-4.0943,-4.3457,0
+1.7939,-1.1174,1.5454,-0.26079,0
+5.4021,3.1039,-1.1536,1.5651,0
+2.5367,2.599,2.0938,0.20085,0
+4.6054,-4.0765,2.7587,0.31981,0
+2.4235,9.5332,-3.0789,-2.7746,0
+1.0009,7.7846,-0.28219,-2.6608,0
+0.12326,8.9848,-0.9351,-2.4332,0
+3.9529,-2.3548,2.3792,0.48274,0
+4.1373,0.49248,1.093,1.8276,0
+4.7181,10.0153,-3.9486,-3.8582,0
+4.1654,-3.4495,3.643,1.0879,0
+4.4069,10.9072,-4.5775,-4.4271,0
+2.3066,3.5364,0.57551,0.41938,0
+3.7935,7.9853,-2.5477,-1.872,0
+0.049175,6.1437,1.7828,-0.72113,0
+0.24835,7.6439,0.9885,-0.87371,0
+1.1317,3.9647,3.3979,0.84351,0
+2.8033,9.0862,-3.3668,-1.0224,0
+4.4682,2.2907,0.95766,0.83058,0
+5.0185,8.5978,-2.9375,-1.281,0
+1.8664,7.7763,-0.23849,-2.9634,0
+3.245,6.63,-0.63435,0.86937,0
+4.0296,2.6756,0.80685,0.71679,0
+-1.1313,1.9037,7.5339,1.022,0
+0.87603,6.8141,0.84198,-0.17156,0
+4.1197,-2.7956,2.0707,0.67412,0
+3.8027,0.81529,2.1041,1.0245,0
+1.4806,7.6377,-2.7876,-1.0341,0
+4.0632,3.584,0.72545,0.39481,0
+4.3064,8.2068,-2.7824,-1.4336,0
+2.4486,-6.3175,7.9632,0.20602,0
+3.2718,1.7837,2.1161,0.61334,0
+-0.64472,-4.6062,8.347,-2.7099,0
+2.9543,1.076,0.64577,0.89394,0
+2.1616,-6.8804,8.1517,-0.081048,0
+3.82,10.9279,-4.0112,-5.0284,0
+-2.7419,11.4038,2.5394,-5.5793,0
+3.3669,-5.1856,3.6935,-1.1427,0
+4.5597,-2.4211,2.6413,1.6168,0
+5.1129,-0.49871,0.62863,1.1189,0
+3.3397,-4.6145,3.9823,-0.23751,0
+4.2027,0.22761,0.96108,0.97282,0
+3.5438,1.2395,1.997,2.1547,0
+2.3136,10.6651,-3.5288,-4.7672,0
+-1.8584,7.886,-1.6643,-1.8384,0
+3.106,9.5414,-4.2536,-4.003,0
+2.9163,10.8306,-3.3437,-4.122,0
+3.9922,-4.4676,3.7304,-0.1095,0
+1.518,5.6946,0.094818,-0.026738,0
+3.2351,9.647,-3.2074,-2.5948,0
+4.2188,6.8162,-1.2804,0.76076,0
+1.7819,6.9176,-1.2744,-1.5759,0
+2.5331,2.9135,-0.822,-0.12243,0
+3.8969,7.4163,-1.8245,0.14007,0
+2.108,6.7955,-0.1708,0.4905,0
+2.8969,0.70768,2.29,1.8663,0
+0.9297,-3.7971,4.6429,-0.2957,0
+3.4642,10.6878,-3.4071,-4.109,0
+4.0713,10.4023,-4.1722,-4.7582,0
+-1.4572,9.1214,1.7425,-5.1241,0
+-1.5075,1.9224,7.1466,0.89136,0
+-0.91718,9.9884,1.1804,-5.2263,0
+2.994,7.2011,-1.2153,0.3211,0
+-2.343,12.9516,3.3285,-5.9426,0
+3.7818,-2.8846,2.2558,-0.15734,0
+4.6689,1.3098,0.055404,1.909,0
+3.4663,1.1112,1.7425,1.3388,0
+3.2697,-4.3414,3.6884,-0.29829,0
+5.1302,8.6703,-2.8913,-1.5086,0
+2.0139,6.1416,0.37929,0.56938,0
+0.4339,5.5395,2.033,-0.40432,0
+-1.0401,9.3987,0.85998,-5.3336,0
+4.1605,11.2196,-3.6136,-4.0819,0
+5.438,9.4669,-4.9417,-3.9202,0
+5.032,8.2026,-2.6256,-1.0341,0
+5.2418,10.5388,-4.1174,-4.2797,0
+-0.2062,9.2207,-3.7044,-6.8103,0
+2.0911,0.94358,4.5512,1.234,0
+1.7317,-0.34765,4.1905,-0.99138,0
+4.1736,3.3336,-1.4244,0.60429,0
+3.9232,-3.2467,3.4579,0.83705,0
+3.8481,10.1539,-3.8561,-4.2228,0
+0.5195,-3.2633,3.0895,-0.9849,0
+3.8584,0.78425,1.1033,1.7008,0
+1.7496,-0.1759,5.1827,1.2922,0
+3.6277,0.9829,0.68861,0.63403,0
+2.7391,7.4018,0.071684,-2.5302,0
+4.5447,8.2274,-2.4166,-1.5875,0
+-1.7599,11.9211,2.6756,-3.3241,0
+5.0691,0.21313,0.20278,1.2095,0
+3.4591,11.112,-4.2039,-5.0931,0
+1.9358,8.1654,-0.023425,-2.2586,0
+2.486,-0.99533,5.3404,-0.15475,0
+2.4226,-4.5752,5.947,0.21507,0
+3.9479,-3.7723,2.883,0.019813,0
+2.2634,-4.4862,3.6558,-0.61251,0
+1.3566,4.2358,2.1341,0.3211,0
+5.0452,3.8964,-1.4304,0.86291,0
+3.5499,8.6165,-3.2794,-1.2009,0
+0.17346,7.8695,0.26876,-3.7883,0
+2.4008,9.3593,-3.3565,-3.3526,0
+4.8851,1.5995,-0.00029081,1.6401,0
+4.1927,-3.2674,2.5839,0.21766,0
+1.1166,8.6496,-0.96252,-1.8112,0
+1.0235,6.901,-2.0062,-2.7125,0
+-1.803,11.8818,2.0458,-5.2728,0
+0.11739,6.2761,-1.5495,-2.4746,0
+0.5706,-0.0248,1.2421,-0.5621,0
+4.0552,-2.4583,2.2806,1.0323,0
+-1.6952,1.0657,8.8294,0.94955,0
+-1.1193,10.7271,2.0938,-5.6504,0
+1.8799,2.4707,2.4931,0.37671,0
+3.583,-3.7971,3.4391,-0.12501,0
+0.19081,9.1297,-3.725,-5.8224,0
+3.6582,5.6864,-1.7157,-0.23751,0
+-0.13144,-1.7775,8.3316,0.35214,0
+2.3925,9.798,-3.0361,-2.8224,0
+1.6426,3.0149,0.22849,-0.147,0
+-0.11783,-1.5789,8.03,-0.028031,0
+-0.69572,8.6165,1.8419,-4.3289,0
+2.9421,7.4101,-0.97709,-0.88406,0
+-1.7559,11.9459,3.0946,-4.8978,0
+-1.2537,10.8803,1.931,-4.3237,0
+3.2585,-4.4614,3.8024,-0.15087,0
+1.8314,6.3672,-0.036278,0.049554,0
+4.5645,-3.6275,2.8684,0.27714,0
+2.7365,-5.0325,6.6608,-0.57889,0
+0.9297,-3.7971,4.6429,-0.2957,0
+3.9663,10.1684,-4.1131,-4.6056,0
+1.4578,-0.08485,4.1785,0.59136,0
+4.8272,3.0687,0.68604,0.80731,0
+-2.341,12.3784,0.70403,-7.5836,0
+-1.8584,7.886,-1.6643,-1.8384,0
+4.1454,7.257,-1.9153,-0.86078,0
+1.9157,6.0816,0.23705,-2.0116,0
+4.0215,-2.1914,2.4648,1.1409,0
+5.8862,5.8747,-2.8167,-0.30087,0
+-2.0897,10.8265,2.3603,-3.4198,0
+4.0026,-3.5943,3.5573,0.26809,0
+-0.78689,9.5663,-3.7867,-7.5034,0
+4.1757,10.2615,-3.8552,-4.3056,0
+0.83292,7.5404,0.65005,-0.92544,0
+4.8077,2.2327,-0.26334,1.5534,0
+5.3063,5.2684,-2.8904,-0.52716,0
+2.5605,9.2683,-3.5913,-1.356,0
+2.1059,7.6046,-0.47755,-1.8461,0
+2.1721,-0.73874,5.4672,-0.72371,0
+4.2899,9.1814,-4.6067,-4.3263,0
+3.5156,10.1891,-4.2759,-4.978,0
+2.614,8.0081,-3.7258,-1.3069,0
+0.68087,2.3259,4.9085,0.54998,0
+4.1962,0.74493,0.83256,0.753,0
+6.0919,2.9673,-1.3267,1.4551,0
+1.3234,3.2964,0.2362,-0.11984,0
+1.3264,1.0326,5.6566,-0.41337,0
+-0.16735,7.6274,1.2061,-3.6241,0
+-1.3,10.2678,-2.953,-5.8638,0
+-2.2261,12.5398,2.9438,-3.5258,0
+2.4196,6.4665,-0.75688,0.228,0
+1.0987,0.6394,5.989,-0.58277,0
+4.6464,10.5326,-4.5852,-4.206,0
+-0.36038,4.1158,3.1143,-0.37199,0
+1.3562,3.2136,4.3465,0.78662,0
+0.5706,-0.0248,1.2421,-0.5621,0
+-2.6479,10.1374,-1.331,-5.4707,0
+3.1219,-3.137,1.9259,-0.37458,0
+5.4944,1.5478,0.041694,1.9284,0
+-1.3389,1.552,7.0806,1.031,0
+-2.3361,11.9604,3.0835,-5.4435,0
+2.2596,-0.033118,4.7355,-0.2776,0
+0.46901,-0.63321,7.3848,0.36507,0
+2.7296,2.8701,0.51124,0.5099,0
+2.0466,2.03,2.1761,-0.083634,0
+-1.3274,9.498,2.4408,-5.2689,0
+3.8905,-2.1521,2.6302,1.1047,0
+3.9994,0.90427,1.1693,1.6892,0
+2.3952,9.5083,-3.1783,-3.0086,0
+3.2704,6.9321,-1.0456,0.23447,0
+-1.3931,1.5664,7.5382,0.78403,0
+1.6406,3.5488,1.3964,-0.36424,0
+2.7744,6.8576,-1.0671,0.075416,0
+2.4287,9.3821,-3.2477,-1.4543,0
+4.2134,-2.806,2.0116,0.67412,0
+1.6472,0.48213,4.7449,1.225,0
+2.0597,-0.99326,5.2119,-0.29312,0
+0.3798,0.7098,0.7572,-0.4444,0
+1.0135,8.4551,-1.672,-2.0815,0
+4.5691,-4.4552,3.1769,0.0042961,0
+0.57461,10.1105,-1.6917,-4.3922,0
+0.5734,9.1938,-0.9094,-1.872,0
+5.2868,3.257,-1.3721,1.1668,0
+4.0102,10.6568,-4.1388,-5.0646,0
+4.1425,-3.6792,3.8281,1.6297,0
+3.0934,-2.9177,2.2232,0.22283,0
+2.2034,5.9947,0.53009,0.84998,0
+3.744,0.79459,0.95851,1.0077,0
+3.0329,2.2948,2.1135,0.35084,0
+3.7731,7.2073,-1.6814,-0.94742,0
+3.1557,2.8908,0.59693,0.79825,0
+1.8114,7.6067,-0.9788,-2.4668,0
+4.988,7.2052,-3.2846,-1.1608,0
+2.483,6.6155,-0.79287,-0.90863,0
+1.594,4.7055,1.3758,0.081882,0
+-0.016103,9.7484,0.15394,-1.6134,0
+3.8496,9.7939,-4.1508,-4.4582,0
+0.9297,-3.7971,4.6429,-0.2957,0
+4.9342,2.4107,-0.17594,1.6245,0
+3.8417,10.0215,-4.2699,-4.9159,0
+5.3915,9.9946,-3.8081,-3.3642,0
+4.4072,-0.070365,2.0416,1.1319,0
+2.6946,6.7976,-0.40301,0.44912,0
+5.2756,0.13863,0.12138,1.1435,0
+3.4312,6.2637,-1.9513,-0.36165,0
+4.052,-0.16555,0.45383,0.51248,0
+1.3638,-4.7759,8.4182,-1.8836,0
+0.89566,7.7763,-2.7473,-1.9353,0
+1.9265,7.7557,-0.16823,-3.0771,0
+0.20977,-0.46146,7.7267,0.90946,0
+4.068,-2.9363,2.1992,0.50084,0
+2.877,-4.0599,3.6259,-0.32544,0
+0.3223,-0.89808,8.0883,0.69222,0
+-1.3,10.2678,-2.953,-5.8638,0
+1.7747,-6.4334,8.15,-0.89828,0
+1.3419,-4.4221,8.09,-1.7349,0
+0.89606,10.5471,-1.4175,-4.0327,0
+0.44125,2.9487,4.3225,0.7155,0
+3.2422,6.2265,0.12224,-1.4466,0
+2.5678,3.5136,0.61406,-0.40691,0
+-2.2153,11.9625,0.078538,-7.7853,0
+4.1349,6.1189,-2.4294,-0.19613,0
+1.934,-9.2828e-06,4.816,-0.33967,0
+2.5068,1.1588,3.9249,0.12585,0
+2.1464,6.0795,-0.5778,-2.2302,0
+0.051979,7.0521,-2.0541,-3.1508,0
+1.2706,8.035,-0.19651,-2.1888,0
+1.143,0.83391,5.4552,-0.56984,0
+2.2928,9.0386,-3.2417,-1.2991,0
+0.3292,-4.4552,4.5718,-0.9888,0
+2.9719,6.8369,-0.2702,0.71291,0
+1.6849,8.7489,-1.2641,-1.3858,0
+-1.9177,11.6894,2.5454,-3.2763,0
+2.3729,10.4726,-3.0087,-3.2013,0
+1.0284,9.767,-1.3687,-1.7853,0
+0.27451,9.2186,-3.2863,-4.8448,0
+1.6032,-4.7863,8.5193,-2.1203,0
+4.616,10.1788,-4.2185,-4.4245,0
+4.2478,7.6956,-2.7696,-1.0767,0
+4.0215,-2.7004,2.4957,0.36636,0
+5.0297,-4.9704,3.5025,-0.23751,0
+1.5902,2.2948,3.2403,0.18404,0
+2.1274,5.1939,-1.7971,-1.1763,0
+1.1811,8.3847,-2.0567,-0.90345,0
+0.3292,-4.4552,4.5718,-0.9888,0
+5.7353,5.2808,-2.2598,0.075416,0
+2.6718,5.6574,0.72974,-1.4892,0
+1.5799,-4.7076,7.9186,-1.5487,0
+2.9499,2.2493,1.3458,-0.037083,0
+0.5195,-3.2633,3.0895,-0.9849,0
+3.7352,9.5911,-3.9032,-3.3487,0
+-1.7344,2.0175,7.7618,0.93532,0
+3.884,10.0277,-3.9298,-4.0819,0
+3.5257,1.2829,1.9276,1.7991,0
+4.4549,2.4976,1.0313,0.96894,0
+-0.16108,-6.4624,8.3573,-1.5216,0
+4.2164,9.4607,-4.9288,-5.2366,0
+3.5152,6.8224,-0.67377,-0.46898,0
+1.6988,2.9094,2.9044,0.11033,0
+1.0607,2.4542,2.5188,-0.17027,0
+2.0421,1.2436,4.2171,0.90429,0
+3.5594,1.3078,1.291,1.6556,0
+3.0009,5.8126,-2.2306,-0.66553,0
+3.9294,1.4112,1.8076,0.89782,0
+3.4667,-4.0724,4.2882,1.5418,0
+3.966,3.9213,0.70574,0.33662,0
+1.0191,2.33,4.9334,0.82929,0
+0.96414,5.616,2.2138,-0.12501,0
+1.8205,6.7562,0.0099913,0.39481,0
+4.9923,7.8653,-2.3515,-0.71984,0
+-1.1804,11.5093,0.15565,-6.8194,0
+4.0329,0.23175,0.89082,1.1823,0
+0.66018,10.3878,-1.4029,-3.9151,0
+3.5982,7.1307,-1.3035,0.21248,0
+-1.8584,7.886,-1.6643,-1.8384,0
+4.0972,0.46972,1.6671,0.91593,0
+3.3299,0.91254,1.5806,0.39352,0
+3.1088,3.1122,0.80857,0.4336,0
+-4.2859,8.5234,3.1392,-0.91639,0
+-1.2528,10.2036,2.1787,-5.6038,0
+0.5195,-3.2633,3.0895,-0.9849,0
+0.3292,-4.4552,4.5718,-0.9888,0
+0.88872,5.3449,2.045,-0.19355,0
+3.5458,9.3718,-4.0351,-3.9564,0
+-0.21661,8.0329,1.8848,-3.8853,0
+2.7206,9.0821,-3.3111,-0.96811,0
+3.2051,8.6889,-2.9033,-0.7819,0
+2.6917,10.8161,-3.3,-4.2888,0
+-2.3242,11.5176,1.8231,-5.375,0
+2.7161,-4.2006,4.1914,0.16981,0
+3.3848,3.2674,0.90967,0.25128,0
+1.7452,4.8028,2.0878,0.62627,0
+2.805,0.57732,1.3424,1.2133,0
+5.7823,5.5788,-2.4089,-0.056479,0
+3.8999,1.734,1.6011,0.96765,0
+3.5189,6.332,-1.7791,-0.020273,0
+3.2294,7.7391,-0.37816,-2.5405,0
+3.4985,3.1639,0.22677,-0.1651,0
+2.1948,1.3781,1.1582,0.85774,0
+2.2526,9.9636,-3.1749,-2.9944,0
+4.1529,-3.9358,2.8633,-0.017686,0
+0.74307,11.17,-1.3824,-4.0728,0
+1.9105,8.871,-2.3386,-0.75604,0
+-1.5055,0.070346,6.8681,-0.50648,0
+0.58836,10.7727,-1.3884,-4.3276,0
+3.2303,7.8384,-3.5348,-1.2151,0
+-1.9922,11.6542,2.6542,-5.2107,0
+2.8523,9.0096,-3.761,-3.3371,0
+4.2772,2.4955,0.48554,0.36119,0
+1.5099,0.039307,6.2332,-0.30346,0
+5.4188,10.1457,-4.084,-3.6991,0
+0.86202,2.6963,4.2908,0.54739,0
+3.8117,10.1457,-4.0463,-4.5629,0
+0.54777,10.3754,-1.5435,-4.1633,0
+2.3718,7.4908,0.015989,-1.7414,0
+-2.4953,11.1472,1.9353,-3.4638,0
+4.6361,-2.6611,2.8358,1.1991,0
+-2.2527,11.5321,2.5899,-3.2737,0
+3.7982,10.423,-4.1602,-4.9728,0
+-0.36279,8.2895,-1.9213,-3.3332,0
+2.1265,6.8783,0.44784,-2.2224,0
+0.86736,5.5643,1.6765,-0.16769,0
+3.7831,10.0526,-3.8869,-3.7366,0
+-2.2623,12.1177,0.28846,-7.7581,0
+1.2616,4.4303,-1.3335,-1.7517,0
+2.6799,3.1349,0.34073,0.58489,0
+-0.39816,5.9781,1.3912,-1.1621,0
+4.3937,0.35798,2.0416,1.2004,0
+2.9695,5.6222,0.27561,-1.1556,0
+1.3049,-0.15521,6.4911,-0.75346,0
+2.2123,-5.8395,7.7687,-0.85302,0
+1.9647,6.9383,0.57722,0.66377,0
+3.0864,-2.5845,2.2309,0.30947,0
+0.3798,0.7098,0.7572,-0.4444,0
+0.58982,7.4266,1.2353,-2.9595,0
+0.14783,7.946,1.0742,-3.3409,0
+-0.062025,6.1975,1.099,-1.131,0
+4.223,1.1319,0.72202,0.96118,0
+0.64295,7.1018,0.3493,-0.41337,0
+1.941,0.46351,4.6472,1.0879,0
+4.0047,0.45937,1.3621,1.6181,0
+3.7767,9.7794,-3.9075,-3.5323,0
+3.4769,-0.15314,2.53,2.4495,0
+1.9818,9.2621,-3.521,-1.872,0
+3.8023,-3.8696,4.044,0.95343,0
+4.3483,11.1079,-4.0857,-4.2539,0
+1.1518,1.3864,5.2727,-0.43536,0
+-1.2576,1.5892,7.0078,0.42455,0
+1.9572,-5.1153,8.6127,-1.4297,0
+-2.484,12.1611,2.8204,-3.7418,0
+-1.1497,1.2954,7.701,0.62627,0
+4.8368,10.0132,-4.3239,-4.3276,0
+-0.12196,8.8068,0.94566,-4.2267,0
+1.9429,6.3961,0.092248,0.58102,0
+1.742,-4.809,8.2142,-2.0659,0
+-1.5222,10.8409,2.7827,-4.0974,0
+-1.3,10.2678,-2.953,-5.8638,0
+3.4246,-0.14693,0.80342,0.29136,0
+2.5503,-4.9518,6.3729,-0.41596,0
+1.5691,6.3465,-0.1828,-2.4099,0
+1.3087,4.9228,2.0013,0.22024,0
+5.1776,8.2316,-3.2511,-1.5694,0
+2.229,9.6325,-3.1123,-2.7164,0
+5.6272,10.0857,-4.2931,-3.8142,0
+1.2138,8.7986,-2.1672,-0.74182,0
+0.3798,0.7098,0.7572,-0.4444,0
+0.5415,6.0319,1.6825,-0.46122,0
+4.0524,5.6802,-1.9693,0.026279,0
+4.7285,2.1065,-0.28305,1.5625,0
+3.4359,0.66216,2.1041,1.8922,0
+0.86816,10.2429,-1.4912,-4.0082,0
+3.359,9.8022,-3.8209,-3.7133,0
+3.6702,2.9942,0.85141,0.30688,0
+1.3349,6.1189,0.46497,0.49826,0
+3.1887,-3.4143,2.7742,-0.2026,0
+2.4527,2.9653,0.20021,-0.056479,0
+3.9121,2.9735,0.92852,0.60558,0
+3.9364,10.5885,-3.725,-4.3133,0
+3.9414,-3.2902,3.1674,1.0866,0
+3.6922,-3.9585,4.3439,1.3517,0
+5.681,7.795,-2.6848,-0.92544,0
+0.77124,9.0862,-1.2281,-1.4996,0
+3.5761,9.7753,-3.9795,-3.4638,0
+1.602,6.1251,0.52924,0.47886,0
+2.6682,10.216,-3.4414,-4.0069,0
+2.0007,1.8644,2.6491,0.47369,0
+0.64215,3.1287,4.2933,0.64696,0
+4.3848,-3.0729,3.0423,1.2741,0
+0.77445,9.0552,-2.4089,-1.3884,0
+0.96574,8.393,-1.361,-1.4659,0
+3.0948,8.7324,-2.9007,-0.96682,0
+4.9362,7.6046,-2.3429,-0.85302,0
+-1.9458,11.2217,1.9079,-3.4405,0
+5.7403,-0.44284,0.38015,1.3763,0
+-2.6989,12.1984,0.67661,-8.5482,0
+1.1472,3.5985,1.9387,-0.43406,0
+2.9742,8.96,-2.9024,-1.0379,0
+4.5707,7.2094,-3.2794,-1.4944,0
+0.1848,6.5079,2.0133,-0.87242,0
+0.87256,9.2931,-0.7843,-2.1978,0
+0.39559,6.8866,1.0588,-0.67587,0
+3.8384,6.1851,-2.0439,-0.033204,0
+2.8209,7.3108,-0.81857,-1.8784,0
+2.5817,9.7546,-3.1749,-2.9957,0
+3.8213,0.23175,2.0133,2.0564,0
+0.3798,0.7098,0.7572,-0.4444,0
+3.4893,6.69,-1.2042,-0.38751,0
+-1.7781,0.8546,7.1303,0.027572,0
+2.0962,2.4769,1.9379,-0.040962,0
+0.94732,-0.57113,7.1903,-0.67587,0
+2.8261,9.4007,-3.3034,-1.0509,0
+0.0071249,8.3661,0.50781,-3.8155,0
+0.96788,7.1907,1.2798,-2.4565,0
+4.7432,2.1086,0.1368,1.6543,0
+3.6575,7.2797,-2.2692,-1.144,0
+3.8832,6.4023,-2.432,-0.98363,0
+3.4776,8.811,-3.1886,-0.92285,0
+1.1315,7.9212,1.093,-2.8444,0
+2.8237,2.8597,0.19678,0.57196,0
+1.9321,6.0423,0.26019,-2.053,0
+3.0632,-3.3315,5.1305,0.8267,0
+-1.8411,10.8306,2.769,-3.0901,0
+2.8084,11.3045,-3.3394,-4.4194,0
+2.5698,-4.4076,5.9856,0.078002,0
+-0.12624,10.3216,-3.7121,-6.1185,0
+3.3756,-4.0951,4.367,1.0698,0
+-0.048008,-1.6037,8.4756,0.75558,0
+0.5706,-0.0248,1.2421,-0.5621,0
+0.88444,6.5906,0.55837,-0.44182,0
+3.8644,3.7061,0.70403,0.35214,0
+1.2999,2.5762,2.0107,-0.18967,0
+2.0051,-6.8638,8.132,-0.2401,0
+4.9294,0.27727,0.20792,0.33662,0
+2.8297,6.3485,-0.73546,-0.58665,0
+2.565,8.633,-2.9941,-1.3082,0
+2.093,8.3061,0.022844,-3.2724,0
+4.6014,5.6264,-2.1235,0.19309,0
+5.0617,-0.35799,0.44698,0.99868,0
+-0.2951,9.0489,-0.52725,-2.0789,0
+3.577,2.4004,1.8908,0.73231,0
+3.9433,2.5017,1.5215,0.903,0
+2.6648,10.754,-3.3994,-4.1685,0
+5.9374,6.1664,-2.5905,-0.36553,0
+2.0153,1.8479,3.1375,0.42843,0
+5.8782,5.9409,-2.8544,-0.60863,0
+-2.3983,12.606,2.9464,-5.7888,0
+1.762,4.3682,2.1384,0.75429,0
+4.2406,-2.4852,1.608,0.7155,0
+3.4669,6.87,-1.0568,-0.73147,0
+3.1896,5.7526,-0.18537,-0.30087,0
+0.81356,9.1566,-2.1492,-4.1814,0
+0.52855,0.96427,4.0243,-1.0483,0
+2.1319,-2.0403,2.5574,-0.061652,0
+0.33111,4.5731,2.057,-0.18967,0
+1.2746,8.8172,-1.5323,-1.7957,0
+2.2091,7.4556,-1.3284,-3.3021,0
+2.5328,7.528,-0.41929,-2.6478,0
+3.6244,1.4609,1.3501,1.9284,0
+-1.3885,12.5026,0.69118,-7.5487,0
+5.7227,5.8312,-2.4097,-0.24527,0
+3.3583,10.3567,-3.7301,-3.6991,0
+2.5227,2.2369,2.7236,0.79438,0
+0.045304,6.7334,1.0708,-0.9332,0
+4.8278,7.7598,-2.4491,-1.2216,0
+1.9476,-4.7738,8.527,-1.8668,0
+2.7659,0.66216,4.1494,-0.28406,0
+-0.10648,-0.76771,7.7575,0.64179,0
+0.72252,-0.053811,5.6703,-1.3509,0
+4.2475,1.4816,-0.48355,0.95343,0
+3.9772,0.33521,2.2566,2.1625,0
+3.6667,4.302,0.55923,0.33791,0
+2.8232,10.8513,-3.1466,-3.9784,0
+-1.4217,11.6542,-0.057699,-7.1025,0
+4.2458,1.1981,0.66633,0.94696,0
+4.1038,-4.8069,3.3491,-0.49225,0
+1.4507,8.7903,-2.2324,-0.65259,0
+3.4647,-3.9172,3.9746,0.36119,0
+1.8533,6.1458,1.0176,-2.0401,0
+3.5288,0.71596,1.9507,1.9375,0
+3.9719,1.0367,0.75973,1.0013,0
+3.534,9.3614,-3.6316,-1.2461,0
+3.6894,9.887,-4.0788,-4.3664,0
+3.0672,-4.4117,3.8238,-0.81682,0
+2.6463,-4.8152,6.3549,0.003003,0
+2.2893,3.733,0.6312,-0.39786,0
+1.5673,7.9274,-0.056842,-2.1694,0
+4.0405,0.51524,1.0279,1.106,0
+4.3846,-4.8794,3.3662,-0.029324,0
+2.0165,-0.25246,5.1707,1.0763,0
+4.0446,11.1741,-4.3582,-4.7401,0
+-0.33729,-0.64976,7.6659,0.72326,0
+-2.4604,12.7302,0.91738,-7.6418,0
+4.1195,10.9258,-3.8929,-4.1802,0
+2.0193,0.82356,4.6369,1.4202,0
+1.5701,7.9129,0.29018,-2.1953,0
+2.6415,7.586,-0.28562,-1.6677,0
+5.0214,8.0764,-3.0515,-1.7155,0
+4.3435,3.3295,0.83598,0.64955,0
+1.8238,-6.7748,8.3873,-0.54139,0
+3.9382,0.9291,0.78543,0.6767,0
+2.2517,-5.1422,4.2916,-1.2487,0
+5.504,10.3671,-4.413,-4.0211,0
+2.8521,9.171,-3.6461,-1.2047,0
+1.1676,9.1566,-2.0867,-0.80647,0
+2.6104,8.0081,-0.23592,-1.7608,0
+0.32444,10.067,-1.1982,-4.1284,0
+3.8962,-4.7904,3.3954,-0.53751,0
+2.1752,-0.8091,5.1022,-0.67975,0
+1.1588,8.9331,-2.0807,-1.1272,0
+4.7072,8.2957,-2.5605,-1.4905,0
+-1.9667,11.8052,-0.40472,-7.8719,0
+4.0552,0.40143,1.4563,0.65343,0
+2.3678,-6.839,8.4207,-0.44829,0
+0.33565,6.8369,0.69718,-0.55691,0
+4.3398,-5.3036,3.8803,-0.70432,0
+1.5456,8.5482,0.4187,-2.1784,0
+1.4276,8.3847,-2.0995,-1.9677,0
+-0.27802,8.1881,-3.1338,-2.5276,0
+0.93611,8.6413,-1.6351,-1.3043,0
+4.6352,-3.0087,2.6773,1.212,0
+1.5268,-5.5871,8.6564,-1.722,0
+0.95626,2.4728,4.4578,0.21636,0
+-2.7914,1.7734,6.7756,-0.39915,0
+5.2032,3.5116,-1.2538,1.0129,0
+3.1836,7.2321,-1.0713,-2.5909,0
+0.65497,5.1815,1.0673,-0.42113,0
+5.6084,10.3009,-4.8003,-4.3534,0
+1.105,7.4432,0.41099,-3.0332,0
+3.9292,-2.9156,2.2129,0.30817,0
+1.1558,6.4003,1.5506,0.6961,0
+2.5581,2.6218,1.8513,0.40257,0
+2.7831,10.9796,-3.557,-4.4039,0
+3.7635,2.7811,0.66119,0.34179,0
+-2.6479,10.1374,-1.331,-5.4707,0
+1.0652,8.3682,-1.4004,-1.6509,0
+-1.4275,11.8797,0.41613,-6.9978,0
+5.7456,10.1808,-4.7857,-4.3366,0
+5.086,3.2798,-1.2701,1.1189,0
+3.4092,5.4049,-2.5228,-0.89958,0
+-0.2361,9.3221,2.1307,-4.3793,0
+3.8197,8.9951,-4.383,-4.0327,0
+-1.1391,1.8127,6.9144,0.70127,0
+4.9249,0.68906,0.77344,1.2095,0
+2.5089,6.841,-0.029423,0.44912,0
+-0.2062,9.2207,-3.7044,-6.8103,0
+3.946,6.8514,-1.5443,-0.5582,0
+-0.278,8.1881,-3.1338,-2.5276,0
+1.8592,3.2074,-0.15966,-0.26208,0
+0.56953,7.6294,1.5754,-3.2233,0
+3.4626,-4.449,3.5427,0.15429,0
+3.3951,1.1484,2.1401,2.0862,0
+5.0429,-0.52974,0.50439,1.106,0
+3.7758,7.1783,-1.5195,0.40128,0
+4.6562,7.6398,-2.4243,-1.2384,0
+4.0948,-2.9674,2.3689,0.75429,0
+1.8384,6.063,0.54723,0.51248,0
+2.0153,0.43661,4.5864,-0.3151,0
+3.5251,0.7201,1.6928,0.64438,0
+3.757,-5.4236,3.8255,-1.2526,0
+2.5989,3.5178,0.7623,0.81119,0
+1.8994,0.97462,4.2265,0.81377,0
+3.6941,-3.9482,4.2625,1.1577,0
+4.4295,-2.3507,1.7048,0.90946,0
+6.8248,5.2187,-2.5425,0.5461,0
+1.8967,-2.5163,2.8093,-0.79742,0
+2.1526,-6.1665,8.0831,-0.34355,0
+3.3004,7.0811,-1.3258,0.22283,0
+2.7213,7.05,-0.58808,0.41809,0
+3.8846,-3.0336,2.5334,0.20214,0
+4.1665,-0.4449,0.23448,0.27843,0
+0.94225,5.8561,1.8762,-0.32544,0
+5.1321,-0.031048,0.32616,1.1151,0
+0.38251,6.8121,1.8128,-0.61251,0
+3.0333,-2.5928,2.3183,0.303,0
+2.9233,6.0464,-0.11168,-0.58665,0
+1.162,10.2926,-1.2821,-4.0392,0
+3.7791,2.5762,1.3098,0.5655,0
+0.77765,5.9781,1.1941,-0.3526,0
+-0.38388,-1.0471,8.0514,0.49567,0
+0.21084,9.4359,-0.094543,-1.859,0
+2.9571,-4.5938,5.9068,0.57196,0
+4.6439,-3.3729,2.5976,0.55257,0
+3.3577,-4.3062,6.0241,0.18274,0
+3.5127,2.9073,1.0579,0.40774,0
+2.6562,10.7044,-3.3085,-4.0767,0
+-1.3612,10.694,1.7022,-2.9026,0
+-0.278,8.1881,-3.1338,-2.5276,0
+1.04,-6.9321,8.2888,-1.2991,0
+2.1881,2.7356,1.3278,-0.1832,0
+4.2756,-2.6528,2.1375,0.94437,0
+-0.11996,6.8741,0.91995,-0.6694,0
+2.9736,8.7944,-3.6359,-1.3754,0
+3.7798,-3.3109,2.6491,0.066365,0
+5.3586,3.7557,-1.7345,1.0789,0
+1.8373,6.1292,0.84027,0.55257,0
+1.2262,0.89599,5.7568,-0.11596,0
+-0.048008,-0.56078,7.7215,0.453,0
+0.5706,-0.024841,1.2421,-0.56208,0
+4.3634,0.46351,1.4281,2.0202,0
+3.482,-4.1634,3.5008,-0.078462,0
+0.51947,-3.2633,3.0895,-0.98492,0
+2.3164,-2.628,3.1529,-0.08622,0
+-1.8348,11.0334,3.1863,-4.8888,0
+1.3754,8.8793,-1.9136,-0.53751,0
+-0.16682,5.8974,0.49839,-0.70044,0
+0.29961,7.1328,-0.31475,-1.1828,0
+0.25035,9.3262,-3.6873,-6.2543,0
+2.4673,1.3926,1.7125,0.41421,0
+0.77805,6.6424,-1.1425,-1.0573,0
+3.4465,2.9508,1.0271,0.5461,0
+2.2429,-4.1427,5.2333,-0.40173,0
+3.7321,-3.884,3.3577,-0.0060486,0
+4.3365,-3.584,3.6884,0.74912,0
+-2.0759,10.8223,2.6439,-4.837,0
+4.0715,7.6398,-2.0824,-1.1698,0
+0.76163,5.8209,1.1959,-0.64613,0
+-0.53966,7.3273,0.46583,-1.4543,0
+2.6213,5.7919,0.065686,-1.5759,0
+3.0242,-3.3378,2.5865,-0.54785,0
+5.8519,5.3905,-2.4037,-0.061652,0
+0.5706,-0.0248,1.2421,-0.5621,0
+3.9771,11.1513,-3.9272,-4.3444,0
+1.5478,9.1814,-1.6326,-1.7375,0
+0.74054,0.36625,2.1992,0.48403,0
+0.49571,10.2243,-1.097,-4.0159,0
+1.645,7.8612,-0.87598,-3.5569,0
+3.6077,6.8576,-1.1622,0.28231,0
+3.2403,-3.7082,5.2804,0.41291,0
+3.9166,10.2491,-4.0926,-4.4659,0
+3.9262,6.0299,-2.0156,-0.065531,0
+5.591,10.4643,-4.3839,-4.3379,0
+3.7522,-3.6978,3.9943,1.3051,0
+1.3114,4.5462,2.2935,0.22541,0
+3.7022,6.9942,-1.8511,-0.12889,0
+4.364,-3.1039,2.3757,0.78532,0
+3.5829,1.4423,1.0219,1.4008,0
+4.65,-4.8297,3.4553,-0.25174,0
+5.1731,3.9606,-1.983,0.40774,0
+3.2692,3.4184,0.20706,-0.066824,0
+2.4012,1.6223,3.0312,0.71679,0
+1.7257,-4.4697,8.2219,-1.8073,0
+4.7965,6.9859,-1.9967,-0.35001,0
+4.0962,10.1891,-3.9323,-4.1827,0
+2.5559,3.3605,2.0321,0.26809,0
+3.4916,8.5709,-3.0326,-0.59182,0
+0.5195,-3.2633,3.0895,-0.9849,0
+2.9856,7.2673,-0.409,-2.2431,0
+4.0932,5.4132,-1.8219,0.23576,0
+1.7748,-0.76978,5.5854,1.3039,0
+5.2012,0.32694,0.17965,1.1797,0
+-0.45062,-1.3678,7.0858,-0.40303,0
+4.8451,8.1116,-2.9512,-1.4724,0
+0.74841,7.2756,1.1504,-0.5388,0
+5.1213,8.5565,-3.3917,-1.5474,0
+3.6181,-3.7454,2.8273,-0.71208,0
+0.040498,8.5234,1.4461,-3.9306,0
+-2.6479,10.1374,-1.331,-5.4707,0
+0.37984,0.70975,0.75716,-0.44441,0
+-0.95923,0.091039,6.2204,-1.4828,0
+2.8672,10.0008,-3.2049,-3.1095,0
+1.0182,9.109,-0.62064,-1.7129,0
+-2.7143,11.4535,2.1092,-3.9629,0
+3.8244,-3.1081,2.4537,0.52024,0
+2.7961,2.121,1.8385,0.38317,0
+3.5358,6.7086,-0.81857,0.47886,0
+-0.7056,8.7241,2.2215,-4.5965,0
+4.1542,7.2756,-2.4766,-1.2099,0
+0.92703,9.4318,-0.66263,-1.6728,0
+1.8216,-6.4748,8.0514,-0.41855,0
+-2.4473,12.6247,0.73573,-7.6612,0
+3.5862,-3.0957,2.8093,0.24481,0
+0.66191,9.6594,-0.28819,-1.6638,0
+4.7926,1.7071,-0.051701,1.4926,0
+4.9852,8.3516,-2.5425,-1.2823,0
+0.75736,3.0294,2.9164,-0.068117,0
+4.6499,7.6336,-1.9427,-0.37458,0
+-0.023579,7.1742,0.78457,-0.75734,0
+0.85574,0.0082678,6.6042,-0.53104,0
+0.88298,0.66009,6.0096,-0.43277,0
+4.0422,-4.391,4.7466,1.137,0
+2.2546,8.0992,-0.24877,-3.2698,0
+0.38478,6.5989,-0.3336,-0.56466,0
+3.1541,-5.1711,6.5991,0.57455,0
+2.3969,0.23589,4.8477,1.437,0
+4.7114,2.0755,-0.2702,1.2379,0
+4.0127,10.1477,-3.9366,-4.0728,0
+2.6606,3.1681,1.9619,0.18662,0
+3.931,1.8541,-0.023425,1.2314,0
+0.01727,8.693,1.3989,-3.9668,0
+3.2414,0.40971,1.4015,1.1952,0
+2.2504,3.5757,0.35273,0.2836,0
+-1.3971,3.3191,-1.3927,-1.9948,1
+0.39012,-0.14279,-0.031994,0.35084,1
+-1.6677,-7.1535,7.8929,0.96765,1
+-3.8483,-12.8047,15.6824,-1.281,1
+-3.5681,-8.213,10.083,0.96765,1
+-2.2804,-0.30626,1.3347,1.3763,1
+-1.7582,2.7397,-2.5323,-2.234,1
+-0.89409,3.1991,-1.8219,-2.9452,1
+0.3434,0.12415,-0.28733,0.14654,1
+-0.9854,-6.661,5.8245,0.5461,1
+-2.4115,-9.1359,9.3444,-0.65259,1
+-1.5252,-6.2534,5.3524,0.59912,1
+-0.61442,-0.091058,-0.31818,0.50214,1
+-0.36506,2.8928,-3.6461,-3.0603,1
+-5.9034,6.5679,0.67661,-6.6797,1
+-1.8215,2.7521,-0.72261,-2.353,1
+-0.77461,-1.8768,2.4023,1.1319,1
+-1.8187,-9.0366,9.0162,-0.12243,1
+-3.5801,-12.9309,13.1779,-2.5677,1
+-1.8219,-6.8824,5.4681,0.057313,1
+-0.3481,-0.38696,-0.47841,0.62627,1
+0.47368,3.3605,-4.5064,-4.0431,1
+-3.4083,4.8587,-0.76888,-4.8668,1
+-1.6662,-0.30005,1.4238,0.024986,1
+-2.0962,-7.1059,6.6188,-0.33708,1
+-2.6685,-10.4519,9.1139,-1.7323,1
+-0.47465,-4.3496,1.9901,0.7517,1
+1.0552,1.1857,-2.6411,0.11033,1
+1.1644,3.8095,-4.9408,-4.0909,1
+-4.4779,7.3708,-0.31218,-6.7754,1
+-2.7338,0.45523,2.4391,0.21766,1
+-2.286,-5.4484,5.8039,0.88231,1
+-1.6244,-6.3444,4.6575,0.16981,1
+0.50813,0.47799,-1.9804,0.57714,1
+1.6408,4.2503,-4.9023,-2.6621,1
+0.81583,4.84,-5.2613,-6.0823,1
+-5.4901,9.1048,-0.38758,-5.9763,1
+-3.2238,2.7935,0.32274,-0.86078,1
+-2.0631,-1.5147,1.219,0.44524,1
+-0.91318,-2.0113,-0.19565,0.066365,1
+0.6005,1.9327,-3.2888,-0.32415,1
+0.91315,3.3377,-4.0557,-1.6741,1
+-0.28015,3.0729,-3.3857,-2.9155,1
+-3.6085,3.3253,-0.51954,-3.5737,1
+-6.2003,8.6806,0.0091344,-3.703,1
+-4.2932,3.3419,0.77258,-0.99785,1
+-3.0265,-0.062088,0.68604,-0.055186,1
+-1.7015,-0.010356,-0.99337,-0.53104,1
+-0.64326,2.4748,-2.9452,-1.0276,1
+-0.86339,1.9348,-2.3729,-1.0897,1
+-2.0659,1.0512,-0.46298,-1.0974,1
+-2.1333,1.5685,-0.084261,-1.7453,1
+-1.2568,-1.4733,2.8718,0.44653,1
+-3.1128,-6.841,10.7402,-1.0172,1
+-4.8554,-5.9037,10.9818,-0.82199,1
+-2.588,3.8654,-0.3336,-1.2797,1
+0.24394,1.4733,-1.4192,-0.58535,1
+-1.5322,-5.0966,6.6779,0.17498,1
+-4.0025,-13.4979,17.6772,-3.3202,1
+-4.0173,-8.3123,12.4547,-1.4375,1
+-3.0731,-0.53181,2.3877,0.77627,1
+-1.979,3.2301,-1.3575,-2.5819,1
+-0.4294,-0.14693,0.044265,-0.15605,1
+-2.234,-7.0314,7.4936,0.61334,1
+-4.211,-12.4736,14.9704,-1.3884,1
+-3.8073,-8.0971,10.1772,0.65084,1
+-2.5912,-0.10554,1.2798,1.0414,1
+-2.2482,3.0915,-2.3969,-2.6711,1
+-1.4427,3.2922,-1.9702,-3.4392,1
+-0.39416,-0.020702,-0.066267,-0.44699,1
+-1.522,-6.6383,5.7491,-0.10691,1
+-2.8267,-9.0407,9.0694,-0.98233,1
+-1.7263,-6.0237,5.2419,0.29524,1
+-0.94255,0.039307,-0.24192,0.31593,1
+-0.89569,3.0025,-3.6067,-3.4457,1
+-6.2815,6.6651,0.52581,-7.0107,1
+-2.3211,3.166,-1.0002,-2.7151,1
+-1.3414,-2.0776,2.8093,0.60688,1
+-2.258,-9.3263,9.3727,-0.85949,1
+-3.8858,-12.8461,12.7957,-3.1353,1
+-1.8969,-6.7893,5.2761,-0.32544,1
+-0.52645,-0.24832,-0.45613,0.41938,1
+0.0096613,3.5612,-4.407,-4.4103,1
+-3.8826,4.898,-0.92311,-5.0801,1
+-2.1405,-0.16762,1.321,-0.20906,1
+-2.4824,-7.3046,6.839,-0.59053,1
+-2.9098,-10.0712,8.4156,-1.9948,1
+-0.60975,-4.002,1.8471,0.6017,1
+0.83625,1.1071,-2.4706,-0.062945,1
+0.60731,3.9544,-4.772,-4.4853,1
+-4.8861,7.0542,-0.17252,-6.959,1
+-3.1366,0.42212,2.6225,-0.064238,1
+-2.5754,-5.6574,6.103,0.65214,1
+-1.8782,-6.5865,4.8486,-0.021566,1
+0.24261,0.57318,-1.9402,0.44007,1
+1.296,4.2855,-4.8457,-2.9013,1
+0.25943,5.0097,-5.0394,-6.3862,1
+-5.873,9.1752,-0.27448,-6.0422,1
+-3.4605,2.6901,0.16165,-1.0224,1
+-2.3797,-1.4402,1.1273,0.16076,1
+-1.2424,-1.7175,-0.52553,-0.21036,1
+0.20216,1.9182,-3.2828,-0.61768,1
+0.59823,3.5012,-3.9795,-1.7841,1
+-0.77995,3.2322,-3.282,-3.1004,1
+-4.1409,3.4619,-0.47841,-3.8879,1
+-6.5084,8.7696,0.23191,-3.937,1
+-4.4996,3.4288,0.56265,-1.1672,1
+-3.3125,0.10139,0.55323,-0.2957,1
+-1.9423,0.3766,-1.2898,-0.82458,1
+-0.75793,2.5349,-3.0464,-1.2629,1
+-0.95403,1.9824,-2.3163,-1.1957,1
+-2.2173,1.4671,-0.72689,-1.1724,1
+-2.799,1.9679,-0.42357,-2.1125,1
+-1.8629,-0.84841,2.5377,0.097399,1
+-3.5916,-6.2285,10.2389,-1.1543,1
+-5.1216,-5.3118,10.3846,-1.0612,1
+-3.2854,4.0372,-0.45356,-1.8228,1
+-0.56877,1.4174,-1.4252,-1.1246,1
+-2.3518,-4.8359,6.6479,-0.060358,1
+-4.4861,-13.2889,17.3087,-3.2194,1
+-4.3876,-7.7267,11.9655,-1.4543,1
+-3.3604,-0.32696,2.1324,0.6017,1
+-1.0112,2.9984,-1.1664,-1.6185,1
+0.030219,-1.0512,1.4024,0.77369,1
+-1.6514,-8.4985,9.1122,1.2379,1
+-3.2692,-12.7406,15.5573,-0.14182,1
+-2.5701,-6.8452,8.9999,2.1353,1
+-1.3066,0.25244,0.7623,1.7758,1
+-1.6637,3.2881,-2.2701,-2.2224,1
+-0.55008,2.8659,-1.6488,-2.4319,1
+0.21431,-0.69529,0.87711,0.29653,1
+-0.77288,-7.4473,6.492,0.36119,1
+-1.8391,-9.0883,9.2416,-0.10432,1
+-0.63298,-5.1277,4.5624,1.4797,1
+0.0040545,0.62905,-0.64121,0.75817,1
+-0.28696,3.1784,-3.5767,-3.1896,1
+-5.2406,6.6258,-0.19908,-6.8607,1
+-1.4446,2.1438,-0.47241,-1.6677,1
+-0.65767,-2.8018,3.7115,0.99739,1
+-1.5449,-10.1498,9.6152,-1.2332,1
+-2.8957,-12.0205,11.9149,-2.7552,1
+-0.81479,-5.7381,4.3919,0.3211,1
+0.50225,0.65388,-1.1793,0.39998,1
+0.74521,3.6357,-4.4044,-4.1414,1
+-2.9146,4.0537,-0.45699,-4.0327,1
+-1.3907,-1.3781,2.3055,-0.021566,1
+-1.786,-8.1157,7.0858,-1.2112,1
+-1.7322,-9.2828,7.719,-1.7168,1
+0.55298,-3.4619,1.7048,1.1008,1
+2.031,1.852,-3.0121,0.003003,1
+1.2279,4.0309,-4.6435,-3.9125,1
+-4.2249,6.2699,0.15822,-5.5457,1
+-2.5346,-0.77392,3.3602,0.00171,1
+-1.749,-6.332,6.0987,0.14266,1
+-0.539,-5.167,3.4399,0.052141,1
+1.5631,0.89599,-1.9702,0.65472,1
+2.3917,4.5565,-4.9888,-2.8987,1
+0.89512,4.7738,-4.8431,-5.5909,1
+-5.4808,8.1819,0.27818,-5.0323,1
+-2.8833,1.7713,0.68946,-0.4638,1
+-1.4174,-2.2535,1.518,0.61981,1
+0.4283,-0.94981,-1.0731,0.3211,1
+1.5904,2.2121,-3.1183,-0.11725,1
+1.7425,3.6833,-4.0129,-1.7207,1
+-0.23356,3.2405,-3.0669,-2.7784,1
+-3.6227,3.9958,-0.35845,-3.9047,1
+-6.1536,7.9295,0.61663,-3.2646,1
+-3.9172,2.6652,0.78886,-0.7819,1
+-2.2214,-0.23798,0.56008,0.05602,1
+-0.49241,0.89392,-1.6283,-0.56854,1
+0.26517,2.4066,-2.8416,-0.59958,1
+-0.10234,1.8189,-2.2169,-0.56725,1
+-1.6176,1.0926,-0.35502,-0.59958,1
+-1.8448,1.254,0.27218,-1.0728,1
+-1.2786,-2.4087,4.5735,0.47627,1
+-2.902,-7.6563,11.8318,-0.84268,1
+-4.3773,-5.5167,10.939,-0.4082,1
+-2.0529,3.8385,-0.79544,-1.2138,1
+0.18868,0.70148,-0.51182,0.0055892,1
+-1.7279,-6.841,8.9494,0.68058,1
+-3.3793,-13.7731,17.9274,-2.0323,1
+-3.1273,-7.1121,11.3897,-0.083634,1
+-2.121,-0.05588,1.949,1.353,1
+-1.7697,3.4329,-1.2144,-2.3789,1
+-0.0012852,0.13863,-0.19651,0.0081754,1
+-1.682,-6.8121,7.1398,1.3323,1
+-3.4917,-12.1736,14.3689,-0.61639,1
+-3.1158,-8.6289,10.4403,0.97153,1
+-2.0891,-0.48422,1.704,1.7435,1
+-1.6936,2.7852,-2.1835,-1.9276,1
+-1.2846,3.2715,-1.7671,-3.2608,1
+-0.092194,0.39315,-0.32846,-0.13794,1
+-1.0292,-6.3879,5.5255,0.79955,1
+-2.2083,-9.1069,8.9991,-0.28406,1
+-1.0744,-6.3113,5.355,0.80472,1
+-0.51003,-0.23591,0.020273,0.76334,1
+-0.36372,3.0439,-3.4816,-2.7836,1
+-6.3979,6.4479,1.0836,-6.6176,1
+-2.2501,3.3129,-0.88369,-2.8974,1
+-1.1859,-1.2519,2.2635,0.77239,1
+-1.8076,-8.8131,8.7086,-0.21682,1
+-3.3863,-12.9889,13.0545,-2.7202,1
+-1.4106,-7.108,5.6454,0.31335,1
+-0.21394,-0.68287,0.096532,1.1965,1
+0.48797,3.5674,-4.3882,-3.8116,1
+-3.8167,5.1401,-0.65063,-5.4306,1
+-1.9555,0.20692,1.2473,-0.3707,1
+-2.1786,-6.4479,6.0344,-0.20777,1
+-2.3299,-9.9532,8.4756,-1.8733,1
+0.0031201,-4.0061,1.7956,0.91722,1
+1.3518,1.0595,-2.3437,0.39998,1
+1.2309,3.8923,-4.8277,-4.0069,1
+-5.0301,7.5032,-0.13396,-7.5034,1
+-3.0799,0.60836,2.7039,-0.23751,1
+-2.2987,-5.227,5.63,0.91722,1
+-1.239,-6.541,4.8151,-0.033204,1
+0.75896,0.29176,-1.6506,0.83834,1
+1.6799,4.2068,-4.5398,-2.3931,1
+0.63655,5.2022,-5.2159,-6.1211,1
+-6.0598,9.2952,-0.43642,-6.3694,1
+-3.518,2.8763,0.1548,-1.2086,1
+-2.0336,-1.4092,1.1582,0.36507,1
+-0.69745,-1.7672,-0.34474,-0.12372,1
+0.75108,1.9161,-3.1098,-0.20518,1
+0.84546,3.4826,-3.6307,-1.3961,1
+-0.55648,3.2136,-3.3085,-2.7965,1
+-3.6817,3.2239,-0.69347,-3.4004,1
+-6.7526,8.8172,-0.061983,-3.725,1
+-4.577,3.4515,0.66719,-0.94742,1
+-2.9883,0.31245,0.45041,0.068951,1
+-1.4781,0.14277,-1.1622,-0.48579,1
+-0.46651,2.3383,-2.9812,-1.0431,1
+-0.8734,1.6533,-2.1964,-0.78061,1
+-2.1234,1.1815,-0.55552,-0.81165,1
+-2.3142,2.0838,-0.46813,-1.6767,1
+-1.4233,-0.98912,2.3586,0.39481,1
+-3.0866,-6.6362,10.5405,-0.89182,1
+-4.7331,-6.1789,11.388,-1.0741,1
+-2.8829,3.8964,-0.1888,-1.1672,1
+-0.036127,1.525,-1.4089,-0.76121,1
+-1.7104,-4.778,6.2109,0.3974,1
+-3.8203,-13.0551,16.9583,-2.3052,1
+-3.7181,-8.5089,12.363,-0.95518,1
+-2.899,-0.60424,2.6045,1.3776,1
+-0.98193,2.7956,-1.2341,-1.5668,1
+-0.17296,-1.1816,1.3818,0.7336,1
+-1.9409,-8.6848,9.155,0.94049,1
+-3.5713,-12.4922,14.8881,-0.47027,1
+-2.9915,-6.6258,8.6521,1.8198,1
+-1.8483,0.31038,0.77344,1.4189,1
+-2.2677,3.2964,-2.2563,-2.4642,1
+-0.50816,2.868,-1.8108,-2.2612,1
+0.14329,-1.0885,1.0039,0.48791,1
+-0.90784,-7.9026,6.7807,0.34179,1
+-2.0042,-9.3676,9.3333,-0.10303,1
+-0.93587,-5.1008,4.5367,1.3866,1
+-0.40804,0.54214,-0.52725,0.6586,1
+-0.8172,3.3812,-3.6684,-3.456,1
+-4.8392,6.6755,-0.24278,-6.5775,1
+-1.2792,2.1376,-0.47584,-1.3974,1
+-0.66008,-3.226,3.8058,1.1836,1
+-1.7713,-10.7665,10.2184,-1.0043,1
+-3.0061,-12.2377,11.9552,-2.1603,1
+-1.1022,-5.8395,4.5641,0.68705,1
+0.11806,0.39108,-0.98223,0.42843,1
+0.11686,3.735,-4.4379,-4.3741,1
+-2.7264,3.9213,-0.49212,-3.6371,1
+-1.2369,-1.6906,2.518,0.51636,1
+-1.8439,-8.6475,7.6796,-0.66682,1
+-1.8554,-9.6035,7.7764,-0.97716,1
+0.16358,-3.3584,1.3749,1.3569,1
+1.5077,1.9596,-3.0584,-0.12243,1
+0.67886,4.1199,-4.569,-4.1414,1
+-3.9934,5.8333,0.54723,-4.9379,1
+-2.3898,-0.78427,3.0141,0.76205,1
+-1.7976,-6.7686,6.6753,0.89912,1
+-0.70867,-5.5602,4.0483,0.903,1
+1.0194,1.1029,-2.3,0.59395,1
+1.7875,4.78,-5.1362,-3.2362,1
+0.27331,4.8773,-4.9194,-5.8198,1
+-5.1661,8.0433,0.044265,-4.4983,1
+-2.7028,1.6327,0.83598,-0.091393,1
+-1.4904,-2.2183,1.6054,0.89394,1
+-0.014902,-1.0243,-0.94024,0.64955,1
+0.88992,2.2638,-3.1046,-0.11855,1
+1.0637,3.6957,-4.1594,-1.9379,1
+-0.8471,3.1329,-3.0112,-2.9388,1
+-3.9594,4.0289,-0.35845,-3.8957,1
+-5.8818,7.6584,0.5558,-2.9155,1
+-3.7747,2.5162,0.83341,-0.30993,1
+-2.4198,-0.24418,0.70146,0.41809,1
+-0.83535,0.80494,-1.6411,-0.19225,1
+-0.30432,2.6528,-2.7756,-0.65647,1
+-0.60254,1.7237,-2.1501,-0.77027,1
+-2.1059,1.1815,-0.53324,-0.82716,1
+-2.0441,1.2271,0.18564,-1.091,1
+-1.5621,-2.2121,4.2591,0.27972,1
+-3.2305,-7.2135,11.6433,-0.94613,1
+-4.8426,-4.9932,10.4052,-0.53104,1
+-2.3147,3.6668,-0.6969,-1.2474,1
+-0.11716,0.60422,-0.38587,-0.059065,1
+-2.0066,-6.719,9.0162,0.099985,1
+-3.6961,-13.6779,17.5795,-2.6181,1
+-3.6012,-6.5389,10.5234,-0.48967,1
+-2.6286,0.18002,1.7956,0.97282,1
+-0.82601,2.9611,-1.2864,-1.4647,1
+0.31803,-0.99326,1.0947,0.88619,1
+-1.4454,-8.4385,8.8483,0.96894,1
+-3.1423,-13.0365,15.6773,-0.66165,1
+-2.5373,-6.959,8.8054,1.5289,1
+-1.366,0.18416,0.90539,1.5806,1
+-1.7064,3.3088,-2.2829,-2.1978,1
+-0.41965,2.9094,-1.7859,-2.2069,1
+0.37637,-0.82358,0.78543,0.74524,1
+-0.55355,-7.9233,6.7156,0.74394,1
+-1.6001,-9.5828,9.4044,0.081882,1
+-0.37013,-5.554,4.7749,1.547,1
+0.12126,0.22347,-0.47327,0.97024,1
+-0.27068,3.2674,-3.5562,-3.0888,1
+-5.119,6.6486,-0.049987,-6.5206,1
+-1.3946,2.3134,-0.44499,-1.4905,1
+-0.69879,-3.3771,4.1211,1.5043,1
+-1.48,-10.5244,9.9176,-0.5026,1
+-2.6649,-12.813,12.6689,-1.9082,1
+-0.62684,-6.301,4.7843,1.106,1
+0.518,0.25865,-0.84085,0.96118,1
+0.64376,3.764,-4.4738,-4.0483,1
+-2.9821,4.1986,-0.5898,-3.9642,1
+-1.4628,-1.5706,2.4357,0.49826,1
+-1.7101,-8.7903,7.9735,-0.45475,1
+-1.5572,-9.8808,8.1088,-1.0806,1
+0.74428,-3.7723,1.6131,1.5754,1
+2.0177,1.7982,-2.9581,0.2099,1
+1.164,3.913,-4.5544,-3.8672,1
+-4.3667,6.0692,0.57208,-5.4668,1
+-2.5919,-1.0553,3.8949,0.77757,1
+-1.8046,-6.8141,6.7019,1.1681,1
+-0.71868,-5.7154,3.8298,1.0233,1
+1.4378,0.66837,-2.0267,1.0271,1
+2.1943,4.5503,-4.976,-2.7254,1
+0.7376,4.8525,-4.7986,-5.6659,1
+-5.637,8.1261,0.13081,-5.0142,1
+-3.0193,1.7775,0.73745,-0.45346,1
+-1.6706,-2.09,1.584,0.71162,1
+-0.1269,-1.1505,-0.95138,0.57843,1
+1.2198,2.0982,-3.1954,0.12843,1
+1.4501,3.6067,-4.0557,-1.5966,1
+-0.40857,3.0977,-2.9607,-2.6892,1
+-3.8952,3.8157,-0.31304,-3.8194,1
+-6.3679,8.0102,0.4247,-3.2207,1
+-4.1429,2.7749,0.68261,-0.71984,1
+-2.6864,-0.097265,0.61663,0.061192,1
+-1.0555,0.79459,-1.6968,-0.46768,1
+-0.29858,2.4769,-2.9512,-0.66165,1
+-0.49948,1.7734,-2.2469,-0.68104,1
+-1.9881,0.99945,-0.28562,-0.70044,1
+-1.9389,1.5706,0.045979,-1.122,1
+-1.4375,-1.8624,4.026,0.55127,1
+-3.1875,-7.5756,11.8678,-0.57889,1
+-4.6765,-5.6636,10.969,-0.33449,1
+-2.0285,3.8468,-0.63435,-1.175,1
+0.26637,0.73252,-0.67891,0.03533,1
+-1.7589,-6.4624,8.4773,0.31981,1
+-3.5985,-13.6593,17.6052,-2.4927,1
+-3.3582,-7.2404,11.4419,-0.57113,1
+-2.3629,-0.10554,1.9336,1.1358,1
+-2.1802,3.3791,-1.2256,-2.6621,1
+-0.40951,-0.15521,0.060545,-0.088807,1
+-2.2918,-7.257,7.9597,0.9211,1
+-4.0214,-12.8006,15.6199,-0.95647,1
+-3.3884,-8.215,10.3315,0.98187,1
+-2.0046,-0.49457,1.333,1.6543,1
+-1.7063,2.7956,-2.378,-2.3491,1
+-1.6386,3.3584,-1.7302,-3.5646,1
+-0.41645,0.32487,-0.33617,-0.36036,1
+-1.5877,-6.6072,5.8022,0.31593,1
+-2.5961,-9.349,9.7942,-0.28018,1
+-1.5228,-6.4789,5.7568,0.87325,1
+-0.53072,-0.097265,-0.21793,1.0426,1
+-0.49081,2.8452,-3.6436,-3.1004,1
+-6.5773,6.8017,0.85483,-7.5344,1
+-2.4621,2.7645,-0.62578,-2.8573,1
+-1.3995,-1.9162,2.5154,0.59912,1
+-2.3221,-9.3304,9.233,-0.79871,1
+-3.73,-12.9723,12.9817,-2.684,1
+-1.6988,-7.1163,5.7902,0.16723,1
+-0.26654,-0.64562,-0.42014,0.89136,1
+0.33325,3.3108,-4.5081,-4.012,1
+-4.2091,4.7283,-0.49126,-5.2159,1
+-2.3142,-0.68494,1.9833,-0.44829,1
+-2.4835,-7.4494,6.8964,-0.64484,1
+-2.7611,-10.5099,9.0239,-1.9547,1
+-0.36025,-4.449,2.1067,0.94308,1
+1.0117,0.9022,-2.3506,0.42714,1
+0.96708,3.8426,-4.9314,-4.1323,1
+-5.2049,7.259,0.070827,-7.3004,1
+-3.3203,-0.02691,2.9618,-0.44958,1
+-2.565,-5.7899,6.0122,0.046968,1
+-1.5951,-6.572,4.7689,-0.94354,1
+0.7049,0.17174,-1.7859,0.36119,1
+1.7331,3.9544,-4.7412,-2.5017,1
+0.6818,4.8504,-5.2133,-6.1043,1
+-6.3364,9.2848,0.014275,-6.7844,1
+-3.8053,2.4273,0.6809,-1.0871,1
+-2.1979,-2.1252,1.7151,0.45171,1
+-0.87874,-2.2121,-0.051701,0.099985,1
+0.74067,1.7299,-3.1963,-0.1457,1
+0.98296,3.4226,-3.9692,-1.7116,1
+-0.3489,3.1929,-3.4054,-3.1832,1
+-3.8552,3.5219,-0.38415,-3.8608,1
+-6.9599,8.9931,0.2182,-4.572,1
+-4.7462,3.1205,1.075,-1.2966,1
+-3.2051,-0.14279,0.97565,0.045675,1
+-1.7549,-0.080711,-0.75774,-0.3707,1
+-0.59587,2.4811,-2.8673,-0.89828,1
+-0.89542,2.0279,-2.3652,-1.2746,1
+-2.0754,1.2767,-0.64206,-1.2642,1
+-3.2778,1.8023,0.1805,-2.3931,1
+-2.2183,-1.254,2.9986,0.36378,1
+-3.5895,-6.572,10.5251,-0.16381,1
+-5.0477,-5.8023,11.244,-0.3901,1
+-3.5741,3.944,-0.07912,-2.1203,1
+-0.7351,1.7361,-1.4938,-1.1582,1
+-2.2617,-4.7428,6.3489,0.11162,1
+-4.244,-13.0634,17.1116,-2.8017,1
+-4.0218,-8.304,12.555,-1.5099,1
+-3.0201,-0.67253,2.7056,0.85774,1
+-2.4941,3.5447,-1.3721,-2.8483,1
+-0.83121,0.039307,0.05369,-0.23105,1
+-2.5665,-6.8824,7.5416,0.70774,1
+-4.4018,-12.9371,15.6559,-1.6806,1
+-3.7573,-8.2916,10.3032,0.38059,1
+-2.4725,-0.40145,1.4855,1.1189,1
+-1.9725,2.8825,-2.3086,-2.3724,1
+-2.0149,3.6874,-1.9385,-3.8918,1
+-0.82053,0.65181,-0.48869,-0.52716,1
+-1.7886,-6.3486,5.6154,0.42584,1
+-2.9138,-9.4711,9.7668,-0.60216,1
+-1.8343,-6.5907,5.6429,0.54998,1
+-0.8734,-0.033118,-0.20165,0.55774,1
+-0.70346,2.957,-3.5947,-3.1457,1
+-6.7387,6.9879,0.67833,-7.5887,1
+-2.7723,3.2777,-0.9351,-3.1457,1
+-1.6641,-1.3678,1.997,0.52283,1
+-2.4349,-9.2497,8.9922,-0.50001,1
+-3.793,-12.7095,12.7957,-2.825,1
+-1.9551,-6.9756,5.5383,-0.12889,1
+-0.69078,-0.50077,-0.35417,0.47498,1
+0.025013,3.3998,-4.4327,-4.2655,1
+-4.3967,4.9601,-0.64892,-5.4719,1
+-2.456,-0.24418,1.4041,-0.45863,1
+-2.62,-6.8555,6.2169,-0.62285,1
+-2.9662,-10.3257,8.784,-2.1138,1
+-0.71494,-4.4448,2.2241,0.49826,1
+0.6005,0.99945,-2.2126,0.097399,1
+0.61652,3.8944,-4.7275,-4.3948,1
+-5.4414,7.2363,0.10938,-7.5642,1
+-3.5798,0.45937,2.3457,-0.45734,1
+-2.7769,-5.6967,5.9179,0.37671,1
+-1.8356,-6.7562,5.0585,-0.55044,1
+0.30081,0.17381,-1.7542,0.48921,1
+1.3403,4.1323,-4.7018,-2.5987,1
+0.26877,4.987,-5.1508,-6.3913,1
+-6.5235,9.6014,-0.25392,-6.9642,1
+-4.0679,2.4955,0.79571,-1.1039,1
+-2.564,-1.7051,1.5026,0.32757,1
+-1.3414,-1.9162,-0.15538,-0.11984,1
+0.23874,2.0879,-3.3522,-0.66553,1
+0.6212,3.6771,-4.0771,-2.0711,1
+-0.77848,3.4019,-3.4859,-3.5569,1
+-4.1244,3.7909,-0.6532,-4.1802,1
+-7.0421,9.2,0.25933,-4.6832,1
+-4.9462,3.5716,0.82742,-1.4957,1
+-3.5359,0.30417,0.6569,-0.2957,1
+-2.0662,0.16967,-1.0054,-0.82975,1
+-0.88728,2.808,-3.1432,-1.2035,1
+-1.0941,2.3072,-2.5237,-1.4453,1
+-2.4458,1.6285,-0.88541,-1.4802,1
+-3.551,1.8955,0.1865,-2.4409,1
+-2.2811,-0.85669,2.7185,0.044382,1
+-3.6053,-5.974,10.0916,-0.82846,1
+-5.0676,-5.1877,10.4266,-0.86725,1
+-3.9204,4.0723,-0.23678,-2.1151,1
+-1.1306,1.8458,-1.3575,-1.3806,1
+-2.4561,-4.5566,6.4534,-0.056479,1
+-4.4775,-13.0303,17.0834,-3.0345,1
+-4.1958,-8.1819,12.1291,-1.6017,1
+-3.38,-0.7077,2.5325,0.71808,1
+-2.4365,3.6026,-1.4166,-2.8948,1
+-0.77688,0.13036,-0.031137,-0.35389,1
+-2.7083,-6.8266,7.5339,0.59007,1
+-4.5531,-12.5854,15.4417,-1.4983,1
+-3.8894,-7.8322,9.8208,0.47498,1
+-2.5084,-0.22763,1.488,1.2069,1
+-2.1652,3.0211,-2.4132,-2.4241,1
+-1.8974,3.5074,-1.7842,-3.8491,1
+-0.62043,0.5587,-0.38587,-0.66423,1
+-1.8387,-6.301,5.6506,0.19567,1
+-3,-9.1566,9.5766,-0.73018,1
+-1.9116,-6.1603,5.606,0.48533,1
+-1.005,0.084831,-0.2462,0.45688,1
+-0.87834,3.257,-3.6778,-3.2944,1
+-6.651,6.7934,0.68604,-7.5887,1
+-2.5463,3.1101,-0.83228,-3.0358,1
+-1.4377,-1.432,2.1144,0.42067,1
+-2.4554,-9.0407,8.862,-0.86983,1
+-3.9411,-12.8792,13.0597,-3.3125,1
+-2.1241,-6.8969,5.5992,-0.47156,1
+-0.74324,-0.32902,-0.42785,0.23317,1
+-0.071503,3.7412,-4.5415,-4.2526,1
+-4.2333,4.9166,-0.49212,-5.3207,1
+-2.3675,-0.43663,1.692,-0.43018,1
+-2.5526,-7.3625,6.9255,-0.66811,1
+-3.0986,-10.4602,8.9717,-2.3427,1
+-0.89809,-4.4862,2.2009,0.50731,1
+0.56232,1.0015,-2.2726,-0.0060486,1
+0.53936,3.8944,-4.8166,-4.3418,1
+-5.3012,7.3915,0.029699,-7.3987,1
+-3.3553,0.35591,2.6473,-0.37846,1
+-2.7908,-5.7133,5.953,0.45946,1
+-1.9983,-6.6072,4.8254,-0.41984,1
+0.15423,0.11794,-1.6823,0.59524,1
+1.208,4.0744,-4.7635,-2.6129,1
+0.2952,4.8856,-5.149,-6.2323,1
+-6.4247,9.5311,0.022844,-6.8517,1
+-3.9933,2.6218,0.62863,-1.1595,1
+-2.659,-1.6058,1.3647,0.16464,1
+-1.4094,-2.1252,-0.10397,-0.19225,1
+0.11032,1.9741,-3.3668,-0.65259,1
+0.52374,3.644,-4.0746,-1.9909,1
+-0.76794,3.4598,-3.4405,-3.4276,1
+-3.9698,3.6812,-0.60008,-4.0133,1
+-7.0364,9.2931,0.16594,-4.5396,1
+-4.9447,3.3005,1.063,-1.444,1
+-3.5933,0.22968,0.7126,-0.3332,1
+-2.1674,0.12415,-1.0465,-0.86208,1
+-0.9607,2.6963,-3.1226,-1.3121,1
+-1.0802,2.1996,-2.5862,-1.2759,1
+-2.3277,1.4381,-0.82114,-1.2862,1
+-3.7244,1.9037,-0.035421,-2.5095,1
+-2.5724,-0.95602,2.7073,-0.16639,1
+-3.9297,-6.0816,10.0958,-1.0147,1
+-5.2943,-5.1463,10.3332,-1.1181,1
+-3.8953,4.0392,-0.3019,-2.1836,1
+-1.2244,1.7485,-1.4801,-1.4181,1
+-2.6406,-4.4159,5.983,-0.13924,1
+-4.6338,-12.7509,16.7166,-3.2168,1
+-4.2887,-7.8633,11.8387,-1.8978,1
+-3.3458,-0.50491,2.6328,0.53705,1
+-1.1188,3.3357,-1.3455,-1.9573,1
+0.55939,-0.3104,0.18307,0.44653,1
+-1.5078,-7.3191,7.8981,1.2289,1
+-3.506,-12.5667,15.1606,-0.75216,1
+-2.9498,-8.273,10.2646,1.1629,1
+-1.6029,-0.38903,1.62,1.9103,1
+-1.2667,2.8183,-2.426,-1.8862,1
+-0.49281,3.0605,-1.8356,-2.834,1
+0.66365,-0.045533,-0.18794,0.23447,1
+-0.72068,-6.7583,5.8408,0.62369,1
+-1.9966,-9.5001,9.682,-0.12889,1
+-0.97325,-6.4168,5.6026,1.0323,1
+-0.025314,-0.17383,-0.11339,1.2198,1
+0.062525,2.9301,-3.5467,-2.6737,1
+-5.525,6.3258,0.89768,-6.6241,1
+-1.2943,2.6735,-0.84085,-2.0323,1
+-0.24037,-1.7837,2.135,1.2418,1
+-1.3968,-9.6698,9.4652,-0.34872,1
+-2.9672,-13.2869,13.4727,-2.6271,1
+-1.1005,-7.2508,6.0139,0.36895,1
+0.22432,-0.52147,-0.40386,1.2017,1
+0.90407,3.3708,-4.4987,-3.6965,1
+-2.8619,4.5193,-0.58123,-4.2629,1
+-1.0833,-0.31247,1.2815,0.41291,1
+-1.5681,-7.2446,6.5537,-0.1276,1
+-2.0545,-10.8679,9.4926,-1.4116,1
+0.2346,-4.5152,2.1195,1.4448,1
+1.581,0.86909,-2.3138,0.82412,1
+1.5514,3.8013,-4.9143,-3.7483,1
+-4.1479,7.1225,-0.083404,-6.4172,1
+-2.2625,-0.099335,2.8127,0.48662,1
+-1.7479,-5.823,5.8699,1.212,1
+-0.95923,-6.7128,4.9857,0.32886,1
+1.3451,0.23589,-1.8785,1.3258,1
+2.2279,4.0951,-4.8037,-2.1112,1
+1.2572,4.8731,-5.2861,-5.8741,1
+-5.3857,9.1214,-0.41929,-5.9181,1
+-2.9786,2.3445,0.52667,-0.40173,1
+-1.5851,-2.1562,1.7082,0.9017,1
+-0.21888,-2.2038,-0.0954,0.56421,1
+1.3183,1.9017,-3.3111,0.065071,1
+1.4896,3.4288,-4.0309,-1.4259,1
+0.11592,3.2219,-3.4302,-2.8457,1
+-3.3924,3.3564,-0.72004,-3.5233,1
+-6.1632,8.7096,-0.21621,-3.6345,1
+-4.0786,2.9239,0.87026,-0.65389,1
+-2.5899,-0.3911,0.93452,0.42972,1
+-1.0116,-0.19038,-0.90597,0.003003,1
+0.066129,2.4914,-2.9401,-0.62156,1
+-0.24745,1.9368,-2.4697,-0.80518,1
+-1.5732,1.0636,-0.71232,-0.8388,1
+-2.1668,1.5933,0.045122,-1.678,1
+-1.1667,-1.4237,2.9241,0.66119,1
+-2.8391,-6.63,10.4849,-0.42113,1
+-4.5046,-5.8126,10.8867,-0.52846,1
+-2.41,3.7433,-0.40215,-1.2953,1
+0.40614,1.3492,-1.4501,-0.55949,1
+-1.3887,-4.8773,6.4774,0.34179,1
+-3.7503,-13.4586,17.5932,-2.7771,1
+-3.5637,-8.3827,12.393,-1.2823,1
+-2.5419,-0.65804,2.6842,1.1952,1
\ No newline at end of file
diff --git a/doc/src/week46/DataFiles/cancer.dot b/doc/src/week46/DataFiles/cancer.dot
new file mode 100644
index 000000000..5b4b48a9b
--- /dev/null
+++ b/doc/src/week46/DataFiles/cancer.dot
@@ -0,0 +1,57 @@
+digraph Tree {
+node [shape=box, style="filled, rounded", color="black", fontname=helvetica] ;
+edge [fontname=helvetica] ;
+0 [label="worst perimeter <= 106.05\ngini = 0.465\nsamples = 426\nvalue = [[269, 157]\n[157, 269]]", fillcolor="#e5813908"] ;
+1 [label="worst concave points <= 0.159\ngini = 0.067\nsamples = 259\nvalue = [[250, 9]\n[9, 250]]", fillcolor="#e58139db"] ;
+0 -> 1 [labeldistance=2.5, labelangle=45, headlabel="True"] ;
+2 [label="worst concave points <= 0.135\ngini = 0.031\nsamples = 253\nvalue = [[249, 4]\n[4, 249]]", fillcolor="#e58139ee"] ;
+1 -> 2 ;
+3 [label="radius error <= 0.643\ngini = 0.008\nsamples = 242\nvalue = [[241, 1]\n[1, 241]]", fillcolor="#e58139fb"] ;
+2 -> 3 ;
+4 [label="gini = 0.0\nsamples = 239\nvalue = [[239, 0]\n[0, 239]]", fillcolor="#e58139ff"] ;
+3 -> 4 ;
+5 [label="worst symmetry <= 0.208\ngini = 0.444\nsamples = 3\nvalue = [[2, 1]\n[1, 2]]", fillcolor="#e5813913"] ;
+3 -> 5 ;
+6 [label="gini = 0.0\nsamples = 1\nvalue = [[0, 1]\n[1, 0]]", fillcolor="#e58139ff"] ;
+5 -> 6 ;
+7 [label="gini = 0.0\nsamples = 2\nvalue = [[2, 0]\n[0, 2]]", fillcolor="#e58139ff"] ;
+5 -> 7 ;
+8 [label="worst texture <= 29.455\ngini = 0.397\nsamples = 11\nvalue = [[8, 3]\n[3, 8]]", fillcolor="#e581392c"] ;
+2 -> 8 ;
+9 [label="gini = 0.0\nsamples = 8\nvalue = [[8, 0]\n[0, 8]]", fillcolor="#e58139ff"] ;
+8 -> 9 ;
+10 [label="gini = 0.0\nsamples = 3\nvalue = [[0, 3]\n[3, 0]]", fillcolor="#e58139ff"] ;
+8 -> 10 ;
+11 [label="mean texture <= 16.22\ngini = 0.278\nsamples = 6\nvalue = [[1, 5]\n[5, 1]]", fillcolor="#e581396b"] ;
+1 -> 11 ;
+12 [label="gini = 0.0\nsamples = 1\nvalue = [[1, 0]\n[0, 1]]", fillcolor="#e58139ff"] ;
+11 -> 12 ;
+13 [label="gini = 0.0\nsamples = 5\nvalue = [[0, 5]\n[5, 0]]", fillcolor="#e58139ff"] ;
+11 -> 13 ;
+14 [label="worst texture <= 20.645\ngini = 0.202\nsamples = 167\nvalue = [[19, 148]\n[148, 19]]", fillcolor="#e5813994"] ;
+0 -> 14 [labeldistance=2.5, labelangle=-45, headlabel="False"] ;
+15 [label="worst radius <= 17.74\ngini = 0.375\nsamples = 16\nvalue = [[12, 4]\n[4, 12]]", fillcolor="#e5813938"] ;
+14 -> 15 ;
+16 [label="gini = 0.0\nsamples = 11\nvalue = [[11, 0]\n[0, 11]]", fillcolor="#e58139ff"] ;
+15 -> 16 ;
+17 [label="mean texture <= 13.745\ngini = 0.32\nsamples = 5\nvalue = [[1, 4]\n[4, 1]]", fillcolor="#e5813955"] ;
+15 -> 17 ;
+18 [label="gini = 0.0\nsamples = 1\nvalue = [[1, 0]\n[0, 1]]", fillcolor="#e58139ff"] ;
+17 -> 18 ;
+19 [label="gini = 0.0\nsamples = 4\nvalue = [[0, 4]\n[4, 0]]", fillcolor="#e58139ff"] ;
+17 -> 19 ;
+20 [label="mean concave points <= 0.049\ngini = 0.088\nsamples = 151\nvalue = [[7, 144]\n[144, 7]]", fillcolor="#e58139d0"] ;
+14 -> 20 ;
+21 [label="concave points error <= 0.01\ngini = 0.48\nsamples = 15\nvalue = [[6, 9]\n[9, 6]]", fillcolor="#e5813900"] ;
+20 -> 21 ;
+22 [label="gini = 0.0\nsamples = 9\nvalue = [[0, 9]\n[9, 0]]", fillcolor="#e58139ff"] ;
+21 -> 22 ;
+23 [label="gini = 0.0\nsamples = 6\nvalue = [[6, 0]\n[0, 6]]", fillcolor="#e58139ff"] ;
+21 -> 23 ;
+24 [label="worst smoothness <= 0.096\ngini = 0.015\nsamples = 136\nvalue = [[1, 135]\n[135, 1]]", fillcolor="#e58139f7"] ;
+20 -> 24 ;
+25 [label="gini = 0.0\nsamples = 1\nvalue = [[1, 0]\n[0, 1]]", fillcolor="#e58139ff"] ;
+24 -> 25 ;
+26 [label="gini = 0.0\nsamples = 135\nvalue = [[0, 135]\n[135, 0]]", fillcolor="#e58139ff"] ;
+24 -> 26 ;
+}
\ No newline at end of file
diff --git a/doc/src/week46/DataFiles/cancer.png b/doc/src/week46/DataFiles/cancer.png
new file mode 100644
index 0000000000000000000000000000000000000000..2ceb5e1f87d923a1cbee266c6cdd1b116e2ed85a
GIT binary patch
literal 221561
zcmZ_01z43`w*^WgjYvp?AShCTgmj01bho6`29a(ll?EwkL0Y=IK^p1YgmibqU7UN4
z?El<zeI6hA_Q&4)eb<_6&N0Ur^ZlamRuT)H1RV|z4og}}ObHGS`3MdU;nM>Y@Jei=
zhBWvG-d;&k1g@Z;Y#scEn2Cn8shk`f1Niv?9DLw=IHbE@0e?xrUpP3#RCqW<@Gtz`
zYpDqT^M8?#QW5{>XM|68zZl%)DG3KB1Sc&ftl|v6HH++wFHM3nkK=?Blj?)}srJzf
zIiXV}?t`73AUE4UWI`uhvQM>0PMUa;$e&g(SSnWOC}O&d4@UW;)(gNNSGOIT%9@M9
zXLDKzkC2}U!J{j9BjBgP{U1NlG2xD8Xm@rf{{Q{Lf4wagi92<G65;=N|G!VALU<q+
z24`*j=sC>S;T9l(%i*?P!K4uM45pV)Zfa?18BE@VHquy(P#zi|S)YgEJw>OjF%g0)
z`Rp2XxRfv*ztL?tRPAW=3hqIo*Y8I9_|}{}%JsORyJ8oAY#S|`3BJ6ws?Lsh$+|xH
zhRI{3(!*<WmDo$h<@%In`e54S@`On#>ut!Z_kU7_d=QGt%cDhQ|6QdwLWBkty3v~C
zV(mbI^8Gy{J)2)%_CG&dHJY2LcD`Le{7S&|_2c@V#o;V@sqeI37Wg~{_1M7Wj-gr}
zH|szzmGlON8$%9_wFSqW8n!lnR|O~2L*bVLJ6~33+EJQ7x&0lq5iMkdwpp1WIvvj1
zEa_e{w@23XiMt;B`VCLl8!0`eTZbRLeAhNe@r1wqOBDS>UNy##TO(-XypCcsRd%DV
zuSD<n4YwNX0A=+KmzxaCi<my!)0OE7zK92_1A%&P-#$~yux0Ik_YmE$plN-WIgZDg
z=WUWeNPMYjnrYwnG&kX~xA*Th_<%q*$W^V~f&#PmepyXzUmm@bA!XDp{)o1<EfZj-
z>}-vG_QLk+4DFTmkW$|WCAV#Aqe+}J%qJz^ebT^Xw;s$VU7`BeEp-T^Na^!?ZP2{E
z@yh<1r9h>EJTo(+Fw^jgK_%Nep4}{9qg><leOP&~Kd@EQ8LlR>Lc4!ah#OrtnU7c|
zmbrbJt1yDqu*KF2jYZBU6*-bl?kjuNTPnjyV~e2~&+7|h4?G!|P#1HC$6k{>)9%8C
zdGRW@X|CSwr9(Hq4!0BCAM6!Dw-d}l{o9b4YWr&j`T-mopR@g?rQ;y8F6~s@4=cT2
za#iyJhN4}F?@y7>9^PjXJ)#u{*kCt6N6>F@7anKUEpi@0h(a&5Ki$#AeEjM$WUdbF
zQD$}jq2f>15QXZo6((&~xBeN09Nvr<+oM@5T4fg7SyQuaC#e#@c#32bC<qu-n+s-Z
zJ-%r+ynEUmQ~0?G4KHUmsn+qF+otaDY~QGCcLI%=-E0VD3FBQ`m*UyYb#0o|Pli#2
zDh)fzlx<$e$DyPUTA3Kko>=Ap(eZ+_qxGnLY8#k<k4u9?J?h)bERlgZ8_h%-#qW^W
zO#1!MmwQleGgzOw+%_Z1#+~ubHF%Jss3Y=}Sq5!ua|U*oJ8sYUip#|Ky_uy}%;6h3
zH=1?jKHdmfXE$x>>zkD<Ct}tSn}w24iGTg8QR7TqVp=ZVH|4NDzR;SRh&fkj4TV=z
zi<zi0B>(eMCP>R+V_|(bOJE~{ocLp*b|d`T+A4#Nvgd1=SZ(s(*;IA4LvIfU1ScxJ
z1R2zG{VINGlL~Ogv-cQ!DX%O24Cj}0Bw*`?BJ^h5XE;_;u=NRjzl<ee(%(&x_hFrj
z?Z(dI`JD32n}CHWObh<eJSc7>hk3^CyQw4MY-myORIJZ}Cn8GAR0pASRF+t@1&8h7
zik!pN?<o3bN_9v#GgS`>UZ{(wcwUdr?@ZV&MHCuPXOz4N;yK<B6@e!5I#FPU@{hP2
zle}sxHtNJ0ALcTysOoivOjLd0sWkr*$?bCR%~Qi;Af6ZJ9D_ijCOBTDS~83x$NhGy
z<hDsTm|RepT))WnVMYqOS)W8dpNlHF<=72;ynpfwh=6gIv9FJBAkwDRBd-)cxrEC3
zj*x_BsuhVmox{VA$ax$jG8r`^Xq+f8DR?mT8$Gd#tOc0yoBCG=QmCzU>WtsRJw@oh
zIonU?%oafI45gGz<hHxQPkimhX-s$Z1Nl%<kuWVuK(4;;rBiY#!?f2t)kvN;>WdgA
zRmA<n)xP$<G||0Ns3)|5+j%X#rm^3CB~{{h!+f}gK&#T?i+E7%d}|<%twVkK#@6gG
z8u?TuyStUO^>tYfce+{Fj;-U(5XF-Fe*|LeT_A3HmT<uYq9FC&!Lm%%T($29lO7aj
zHINTSgo^Q<uP@3ux$B8Zxvc-7Sl+6w=Id5sqv5rwPlxg+j=*_oTZcQgR24gCW2MG2
zXzTTroE{u}3hE)YHmHAYeB$AAbzW>c^e%V^<w*OLWk5j^@4{BK{gjXsQO28SJd>Vp
zY7T0{i$AorGS#L#yh15>18ip9nx;wDpp`fvWdx!}=gqeV5#>%8TaQ)x8aZVYze%Ui
z(5Z13cS}{kV~g9-*WQ_J40}W`pKDL~p4u=YWvQGCDOWA$$LKXhpH`X5@nkIHm#gyw
z!z;fp+_rlL;o~3;Qn8tos39&MZ;lR6ZHBy2%`=?6z{*yxjXB8)fzr#VQJZJ5-A-jP
z<29vRYSo$u@ylHVn@`mtXKa56@fNsIjiM4pu=bgZ+K76g{#H=jrnaV~!5EG5oRc%&
z?s&?<A*t4B_kE{~2u1&^_hUM<%EZB^5Pk&WqA+51I&>mt<7}3okPwpIMz33+3;vF;
z1Tv$F<*steANCiYicWfS{=JYd50tp6C%_$i-J5lPz&?8;ObM>B!DM^|Dkb!-sq&s$
z5&wfZXrnv&Ua9erh;<j)980pm$&LLr<s3K<C!vH)-zy2V+NX!FL#FgfEq?MapL<*$
zagI1_UZ^_kls%8CKi+nA*mU)}{`<3AB}ZqHSGVTloZc?q=#7tuTU{-$qw=(y)2_;D
z&0L*L_CQq<+9SFnW{o26)5-zL(S5gq%DQy1FN^%9BT~CzySMvqzdehn)R7UCY`j_Y
z$GEcYN%cXjQ~vqpXLnD<(fQPJkcC%Irr@Y&V}r*?L*O)YZYW*yH2IaCNcec<W+h~N
zj8{H+W@E-$n=_|otU&MQ$@1A+1dWOZM6Z6Y(%O~sXqSaj&>1pU)vS#nlYU`kXt*#t
zoD+C@{32E*Y099wCg7;~)U0b>IAC+*nAnSmNvloU?ejvq?7`pP_2jj$&$@<9N8NfH
zQuyi$3#c10tRu8`w<b$-u19i|-7GQ2F1RHNhRbEld`mrNlcp<k?XkC^tNiaK6v}_~
zb)Alx?zvg|UT$!mjQd=BZu}V-Z_HEswOH@4IVN=uHF5m@5BP}@`m9X}&j0T%lX50~
zA=4vtsaM<*HgVA@pExK|k|=ckVR`Q~q0#Q|+DmiaDvOuHpdu`EN9UcLY?Ho;E#Vq5
zOYf_TQDCaHfl5SoN2_>7bznW@*a#ysar?N7q6x=Q?_jD@sHU2LrLWny#ifGVCh^?O
z&D{4aS0w<GIQ*rN1V`&7I8=Lkdrf)cg;Wr%HEFl`)G4Jp4aI1a8A5>Ve(~daCnY6K
z$=O8cc&Ha>D0=h+pH$f`wI~OjQjC-zT^#KmuCiqJU87=W<8qZ8&Nqt^e2_HF-h}WX
zjc~f~{*WA)3ojBrF!euSMJ~4-RVkf2w!|c5n}FJKec*E{&~Ys`@RE_DK+7^tEJGz>
z*y=xA6`cEAG~y{qG?3)OnO)ezApJP=9E0$k(5+`jRt^{L(VoZ!e2z@yi8q|f;fl|R
z0!qExu~_iqmnskFH2Ivm_7=F^_q;v7HbP@^l$QXUyrD3nyGJtKE)Q&lpnN+tZo>EI
zX^#?VuF(tf7#C;lqje?+032y3hfjOuS}%?lR(3+`FHdrWjjdz&0N6o4trR3wDwo|p
zgl|6Gg?gwYy^>aHZD4GxH5<pyzcwOw+?~u@Y?S#-ZcSL5CTl?YyBywNiPr9~$&jkk
z=`Bmfi!|q>kvdJ~zr7KnP=7mZ%=lXya|qm7|0ge(=K2`D_{+wit8Tbzfg)?e2+Utv
z6|9&mGLzm@jvHk`B)ze2ixa={(9vjd7=oi1pOHpBH|hSik>+cc1cIMh@?=XYp3T48
zW+Jr#{BBZ>!$vmL&u?QtDT5hpoVKyvd4DmdRVfy8!|eAvMC*XY6Wgrs3$I)h6r|n4
z7eA~7Q?sOExm3{{!v<(vA)L`#!fw6E08APGgd480f>+3wO^cZ8cCyp%k3poGf5i4v
zCTH7v+9|tP?X{rj1!gXZriQA#Q`y<o&eXiuifsH5BPBGwGg!5>$fS6*6K5x9@|@Rk
zXM|AVm=||zsm;sDX0FDhD<=Mc&63}(!m6mEKl?xip9Pz#E%c%{fn7VOF9m&l_$Pbf
zBRo<;Hzvz%CIso(S_k&BU)tjZgDq3pC&;6j#p#m57#@Bq&y~~El3XpCq13GbmWVqz
zm{~D}8pK~BwazXhxK9NWJ}mcsu^cbV<6`}*<MK#!$G)Fzk{ZX|n6AEIzKF~#kH0LO
zi?(grDYAwZWWPz<wY><MqVm2U8xir>C3N0)HMjEu`T6bw&)5VCAU&TNF08!x!=NxF
zbmeIiyvJcp$`;mr{58oxFt_9MRjpBKN3Zg{Mlo)BrQcBZOs(vhvD#S(DOYM>S_JWu
zW~pg=qRSx_;yD1W&4UXxBZ$5oAtV=W<eK5*IqDFy(U#M$-VJ*BlqYtJZBtLGVm?3O
z4dq#F`>h-v&T{iJI#120XY3cljC|$JTZV68pj%DgNX9dH6WdX;{9>=O=2&aGFx_F=
zG4AY=R_a+&xNLKEOclS!1!iUWplG?pm_pju3#`Wv&>CrQSihHB`2*nlp8e<%(&Q`R
zq8R^=%Y6a7l{P|iJhsV=aATYm4DLe_G?GIVa+a8sJZ(^U0}LwEkNbpu3AylY#O94|
zr_2{8l&L#~b!b$LzK+(@=0Oi2&0DYSsWB!T?bb3;u1y{qNN@9eY|9=p8gUHS7#-nv
z+?ds#t?bjT$iAdKm6QLFqE@Jx=PAkWcG}Egy+4~7Lb@4<c!j;Q`??9;`M!%(X!?u%
z_(SJk$U{=8`vZ<Uhrl+hi}Evw=tU-q?Nd+pG!f)cp*(k)NS|2(nb=f4-+-=nvAWCB
zQqig}VshfAT-QkXSL*TC7S=vX$R(tyOn%69pDKGl%INW$#p}j>kg!+0XUw`xfeWq7
zXYYG%@O}ztUD6DfGQ$;kpDSfF<r5F9I#WF_ZE`8FkTLq@DXp}6l%pIjnxM*zLq<zn
zDPD2xaXz+mmG|l9Dq=~o{6i!{Z46oLF{1Gxzge>T+0g=d43lO)N;L*P^OL}c*Y>7X
zXGXuKIBk5pAN&m~FX!k8BlnGH4FuoV7sur#pjN1h;X+DHtK3Jzk;;UR89yuN(yekV
zB%;0c7|%vJQs;JO)@^O9K5#0Q__EZjUjfI9)(Coo7@#0|Fg;n}@+fyS-+$)(AnyRr
zRC*))mC+Qw``Ml_zxUA6iDeLaYnmmIL!uW<0(ep>gs*wc_2E&`DNMvis5VVb5)-i`
zb|z}Fiv|m!7(!;Ma!k$5dQ6For6|uk_lLD~oHY%&KsgtfqIRgRuFkQ2)qpzjh+Za?
z+85=A%<9`JmqWr2%iTE?43E{R_AgD;uBIy~!Yj^*5TR99g;uBzOlZP1FAq&345Zo2
zJ&rdxmT4edXFE_Y>!6VrzAI@F8948laFqUfp1;L&lw4epVRYvgXxL@DXwZ<FNP6`?
ze}t64wc%!aeLZKESAAl_82dz<U^dD1NszS~H#hg%=w8CVdacmZi^MGo?X`JU3(Ew!
z!xQ^qub(rq@oJ-t`z9DmxJEqjzh+`~kd{hlimZDb&IO_H3hJV&_Jc<c!XGJj54;$H
zet586;Xh{3P^NsLP-UA2m%s>7>7wVB7D4^3{TTkK%b8MYQyB3TeDr=2OQQ)ii1m`W
z)dQyA{mj(#_~PN^P4O|zf;Up(CJwSSDtKfkTNoOqwZQprh-<j`;e`h>xOV+`Flx&g
zbAK~b69|cElvWIcuuU45&u6}LO+vrxXj+V4bZ@>XV4O$N4O$or>!F1hnZ<H+v9kH7
zwfUwo<B~uU{an*s?oXAFPYT{V=5@;AJHz|ygtx~FpX`lrkSJxGyS)7STzKxS`Z(`1
zm9odoI#rL2C+G3Q1D};t<uiUxcxg`v@G#jxGCxLt&f3f<yfmZI*)zBE0|Hga1C^Mf
zisE+KU-AdIB7dswR~12DElByL6cYNJ?KZh@-?K#i-#Ro2!uJoJ%KN+gdj*h8B0?y!
z%a4eMr&M1=aNGT$s!e`vPe&&B;c<zrwN%ECQOCQjGqkv?OR)_36l@yFFySvsCbrkz
zPrZn~j}}&Ywx)pG)+8C1#s!&6mDuK)`q4}n$8V34Ea36ASu`-mm8{0G`Dah;{@RdH
z@MBJuLJ5z~VB+D<hUzV`6=-SojOdB)zokb4mEbzn&pGkmD+*0k$i@<-ZRA{?N6S4#
zm}GoN+LhLlBrjf+*^J`ljpw)f*%^(_UzHKFD`Xo8e_WqI3;F&^{5RfH+np<yt!Xc~
z-;qz-n{KX$Qg0xjto!GhtsK4YilDh<QI<=5MV!RzpfVoajrUX*>5|Q)*rOBsO06)t
z@fh0ZMS0hv{`^V+fTf~neDz`C!p=^pWh~1R&#N;AL5GczfWTJ;PCHccJ2Pv84M_`R
z_Hzc99`+5jTcrQ$*cr7!^6g1WEesVV<YVxlE7<f4uT5x<C6}aZ%F9R5FI2^HRh|!(
zIFN45m7+ps=CXH^dA2`kS3>f(gJPM}wNELk>`$3_oVHT^n_NOO`?Tk;FHfCMclfa>
zUw^8EKrebS2GCD;r%%xaQUui<reYZlxs8pYG_CbV-(6hB{%saBVq%m(G?h~<up*`u
zH*RY9R?qLaGZn(QJ=NM=%*&P`9|j6%?uSZ9ZA<GgA^;~s<s{{-#uK*A&dz7qt%;70
zMWKm1beAIk;`)kc-n-$Jx%Vebr-g*D!jpc!+!L!&@{t$?hs=9jIm=-5>+)kBXN6R!
zS8gX(XPZr)m}G;Lmg<FV;KFUMUFR`&-zKvW487|(IgR31rw|;Ea=wa;)TZSklZj{h
z`Ca$+;@DC&dA}o-TC_Y6jbft@Vaa{6!EKiUn~X1ejx~Ib2QM;9L)y)1u0))`p?>MB
za@O0tB3_#vo5$>qi4_z=cE<|MPP!{akbX#FT<f(V<-Y_iB=vyx?^~&7CD`{T)91l+
zg97&NklC7FD1UomUzk;~5Ol|OepXD!x@_R0{ToW*DP>wM6FGUM7Ngb>L~@d+mIL~P
zQ<kT1z75Ksbjk}7=3QSL=Y<)8?reN<{?;p8U+@_OUp7Giu@n1tnZ~ifQ)iAg9N89!
zlFvJGD*v`Kj55#p3NN<N@t2nQOcAwJpsDaIkHg6G!^G9}AW<$x-3l_@+K*#-FB2g4
zo=a_3Hh8@E`{1eQjd;=cplcY0;d@fxUEOaU!^$3sM=NDe^}58?%HX-pOjSVZMN<Pc
zjW-rAFK-up)pPL!{M@x`R2+reZ}n3#T=2As^4Qy0@X+^DNAcm!H*?n)Iig{df<Cda
zv0tm9S?coJh9c+=Iso<mluIB~Wl(SFR?Qo|@xBr4SLU5cF_lZm9u(n%%>F6u-I&(#
z8ZdQ;nz$z;_ks{Gs}F;``I7F>H`O9~+-C5~(Yj}Ek)Y#M)xt_u;%mzZ?3@tI(p4%&
zB@8b?|F49H_QYReDR~{Oq@2n4*MnwOx^|qJX9KV{T8k6z2=H8i`21g^+DIiS3?>W&
zEcz`N9v4T_h1fWh0!k1poqAU+lOFXpKvZIN8lyDw35AP)>U$Z#xjRX=OK>UnFQjks
z;LO%;)n6w|{(A|8$|HiVm_F=^ii$Ws2NI<WsZc;YrqwDB;6i?3=HuJ}IoGJCGAs0g
zdUej(>Xq2EM<I;12`@z@j&`Qy%SSspB%&eXX*7SbWCE$RdT+;mX<cL+b@Xa-yKLYC
ze3E{^>u4S(IQC23@FN*HneE;n_3`@XFd8viMqp1o``Ta*#>wua|0;`~!NlhX3yE9E
z@y11VEH6wF$N&5c@Y-R|oW5?u)T{?~@MwxpDD7qNGEKL0qs?av$TzsG@<n#LglgJC
z!vGTAD6ne`sa#HNDD)y-&563bj$+Z*>smGYvlY*7ySk%00y;H?**6JHjgZr~VnKv*
zs-=)yc+is3XsE_8tD>h%hAJQ)mpmw<{SqrEr9GieLD99?P?BAs8>+Jr=SQX0@DYv8
zq&wrJZ#!&;Wy#Id=Gq7r9hG^hjqCLs(INdym_Ye?lHLSBqTTu;a&HoUPks$cQ2(1M
z^4fgO(WVMa8xRC!Kf7am0sA5FZZlB5b9c5j5QnlZoF)*ez$-XShl7JNQ|E%Q{^zIo
z#z@X}qUY_62Z5XZ9Kz>^Gyy3~E0GHmt!SK@Rtsxu=FGb91pnm%IAnJ;+kgfZ+#Mwt
zoJ{otqYLvI-WMfLfj3<;SfR9{_H--9beNRo8)@=&+1TPjpvs$X&mIDP3g5tinB5d}
z?=cPxhp#~O1kF})=JUtO3AOti1CNfRY|t^16CP{Vqnhsr=sGI9A8ivg?~PaoOkcwO
zo<b8oGJsb2yXe#Rarz9G4nZ+pf@3;v2^oX1X~7p&C7T$Hou>3elxMcePSk%_I*LwC
zO1<ym#@D-TAQt)xMqBefN8d|1={=n*F@p&q;Ro)wH@bcAaDHg`36RmhO;jN6CHh9!
zr)_Z629@hJyY)f*Ou!`o$aHJ-*u1<K79+X7Dr)Z|o|F^hj%NE$S3*cbfu?{g_nqT^
zKuN}L<-PqS$=3bPCDah(F39&U_s71;9a+Tld+51VG3-U)r+TAd)VZJIVUch&$Fk^S
zB2Jg+<{Lk63!|h&yg2r6o2KM<B-1Lh@m1~&Tj)#R%2uq!j-;<_0`a<cw61)|Bm?3O
z=m^YoGuoUX_uogwhHe|+X<WCt#eN@aLVko3HP)mSV+BFR&y~^)B7RJiIu)h~vQ5{i
zCR9izG?M_<J3zT`E?{!SX?IRMV)1Fw$H@;Wxk^~XY?G<A%JMpmp5)6f)YRA~^3)2y
zs2A$xT+qqKqjyrgeavRXT{2-3C^u#mUg!NMvHEa?W7wqEFCZjg9@K?^A8a%YAhl;m
zrznm-b@04A>9f63=Q4J5J&gIKm|<&3shY&=q=z*H(P_k<C9h3W?IB!S9zkSgHnRnE
zXK&IfrjwJ?m4ksG$;d6%TmBlOp-<;hF^utHu2nD)TnJtn7gXuzGsqP__lx0;(2Va4
z#77dY(Uw+o_U_JLk}rU*%5C-lhsXW;g5XXN$Qfo{1H8C{<#-3^Kwo+_i>O`(;^<jW
z)#?U@7M00{@KxIzt>jO3MRN(h>j?hB`Ad6YK26jDP9L-;n}nu)$-Z(4T<J1dp<S_D
zI2T8oVSsW!d;bgbh?m<@k|ggH5H~){O!@jhX_sjk9#)Rt9<0W=Q&ZZC(w^ydQ!LTT
zw&jjmUX9QYvvE@nc0wPO^`dF5rMR8$8vSY87x+s1<{Pduy}U>$*(BCE0AktOaW9R8
zDr6?7jaCb0&4&x>H$(%D)`zMTMMYg~V2@;z3KHmjRd$f3VaTmJT(&h?E6?<Cy2eE2
z+cQNSEIu3nHvym;M>&t8i7Mxi2qW`BKt9Nu0)%@3Ru{y_hznJ8TND+iWsur;G}Wcf
z?@M{Jhs@SYgDQEW>+yHa$6k{PYJ^&|SF0$-{fU}=wq8ePFHHxMf-0=FNW0x{t~C9N
zbJQ!D$8PtH!W&%6bBp?75P@VMQD5U=D{Cnk+n+Sz@}a;whViE3Bo-C%V6dt28;7Nu
zPUDB)18Gt8@`h(!iriI@LiAroI!*ok*at&Vd^6#i=RhG4uCSW4J)PrE6!d}z<dK-(
zwvT_^0u_YBYNiMqaEyQCm$>ea0vxie2!IiEcBB&zVABthJP=JRq>^|zx<Q`$O2Cl%
zX8FxyHghT^z)CY{l)od6VbX_RMaHnz*2X*m7{92JWj+1iUHjKu590jFq9UQ?Hw0Ln
z%9*lhDAmWoK&Q@WcAtzr02GT_>BaHVfrmUPek>P~UcKi}k)yTFAL2m)+FIF3NN9N3
zlecWsw-ftI*;+r*H_UyP0ZBO=MA87@R`9py$|NTtK&`6ZrCs<yVu>;JCYx+M)-H~6
zspL;bzizD|^yb=ptPIVxFRlO^$fi2}v^?&6aBpH5eLANQgYbTK8vwo-Xu&9dIqz4&
zTl$Xmq_sgL{g1$yN0a)TWE6;Fm_~Ra@orTnE}`KXb-r0wonuMYv|T1wBaM&YJ>m`M
z8BUe$T95=#J4~ARQ;iK5*ki7;<Ws09m;UbDj1^t(%<{H60uUa--Whfxo5F*9ez@FP
zvCYfsF!eE*oQpz+h(#o5r4Mu2r1dn|l@j%OGJ*@v825Hi->bPnR#S8ua7~{PQE?2r
zGoW@0E#fM<T5^4J&>9z-T=hKDY-tJuu`mi<tQ8s4I}S!P63|<u-M07b6*vQ#PXO1Z
zT>Z!j_5QBmkUvpVVJr$b?<uqo#6G^@vY8Q@L+~#$Sa{MFM1*5wW5d1rLCZfZV#9HJ
z@=>rBBJ+pUe%)gQDqxK0fm}PBv6h$(tSli}ZU$R4itR7v>Xy(LI@4mlb+rz-^+m;D
z9d{A9I9?A4C7)UXZA-RVDIs8E7R-e7P4{s~*hVC;FG7D=6ILbirS0bK>b$o9Q3!gl
zoAXu5#Mky}80$tsWRs<N$euml3RBx}2ZERqZd4!`YbMLui_1Va6y<;LnczhX&(YpO
zIvf;_RUmd6rmo>XumtjIU^#aoET;3(#|`&6p7Xe*2E1Ulb`~dXW<6$d4^^$ZOpV(v
zbACSK=wf+iSc*cLwQ=*zZ6xitw*zw?gRRA+YWX_;K}0OV@AgRP(IRNloM}hGFpE}4
z2vf<^`<iZnxcut<h?GBKBq}x@h1sC@O}GMIp<ca2M2mlLp5k=-SHe4C^6IaQM2WEg
zlrX&R6Ea42n3YUqITEdeXTlhh;x7x;eC<YIre1B(s~tNwg((T%bZcWCcneQWwx^K-
zh&b}2sf5**#ahvj4)(0ZeFyra+;(UY|9{-S7&x-s*($|Pw94P-3E+LovP%zO2<`>V
zmrC6gO33$Y5<nw;5k^Ecm^nk+pca}QyVdU+JHx{x)t4`^F3U5`g-q69v77db^v1os
zYoi|zHl{GxxX&T1R%ef=Qj>69KNOLiBMfKdeO=b<A<ZTw=)A_ts8gn~O*Dpb==S=Y
z{MA4#3(Ma_{iqu{pBs;jqxY(#;fAf(DlUHqg}*VYzF~U8XVoUfs97T$oui!<>bf_-
ziew6#v8fV~e>ht{Pn>S*fpG>XQ0X8gCMDVJx487ibFk|a+`b@+7T2Q`^h6+j97fJt
z5tjQmuFXFd+4^a{>+zdt>9|3&Eq6-#^U~o4BYK@`R|5Jl^7eAe@d&z*Xjk^8lphQ-
zi^tG0!e)=lF~rG|$>!5zTCz|8!n;h(vItMGvXl2fF?(n;SNAjiCr5P!@_0G`0mj9B
z=&0Lrh_IV+h(C|%aadQlMR0%4gjkV`;dp-ZLy9XZ0^F};8V;2+q!#px4OV~4x_e}L
z7|{RDKg-uFMbQu=gfgjDSb2B09t!;s`!doRv1ip<sHg4ObHnuMD)0Ef9cJH$AD$JL
zM6Z_xGNs*j;s8g>iV302N6|@JJo#nF4uf*RkNtzI5EzW>jqp%Kh_Jcd{roHEi#fvm
zrB0lEJ)lCk6>ai4ZgJNMxKgY2YA}R0?p_<H<NKh0y<MP@R0)dGrj<<QtEN%doK(5L
z9-q(8nusqR?iUpI!Gt08lRYE&iCdNjq)(y)44;C&Of2n3yY0@*Icr_+ZLGyyF2{FH
z$d?;mlpKVHQ$ddlKMH=gW@brdTB$^Q7YzKinqyMTXcrfk9DlcE=N8)EgT?klGxcuj
z<gp5D!<hp@z<dzF7ya}Oi+Wt1lJ;Zmbc`43M=+V$P>x62w-?$LFM0}43b;3oydS>)
zyJ>Q(QAsgd!<)7P!J-P<2yptk*0{QNvUL2fi53#W1*F|)+t>*Sn443081A$9nM71j
zMaZ-GkEe%gy1G2QC{xEJcIa%e#WBepD+SnsRv;e?RlR4&XV8EnW-}HVw+Wr7@IXzh
z;cXe?#E=Lf4&vPAo!@Ilv|w#iPia_NA3=O7gWnk-$n^tgF?SpZK7-1jZM`=q1;48w
z5Z3a1t`fK)i%Xqh0$=rN+=;^|_&YuuihBG7WZI+igB5?wZ=_kkO|v+%K5StJO&Bjo
zb?H&tJG$zulB%*gJSAh^s(bb?6dx$x<;3S)EE2GNZVG~G-nd_quA1zP$BzYXOuak1
zk)_%%6>5h}*@OU4c&LF#rieuoZ>b|#Vr!xpYuQ8;C0D!LI<OsMg+{x=YGWdQ(%$xH
z;^kGbY5A0(*A2p5#RB3{E2j+M(b^mqkW5rMMJ@++_}$7kU%%U0Q`}H$4<cxd`K4WK
zF!Zj>d<<VEmZ|eXa4c1Cwq~PB&Xm+@wgL-p%=ElF`dK#CmHc;ZVe&kM20l*6<X27m
z6$6*Fr^1}qGv|l4dPQ)c0okw2-%R4|isI$sDTA!8I_D^zgoh;;=v03NUK9hLL%zQQ
zDe?m={{ekk#Js7UTKwXPU=9$gIDEVdIfupB4NdV+6ASJEeS<g-W<$M|*8`Fr!Nk<m
z?>0Ehv-?n-gdx@|ZnIP6FWvfedi9G2aaMtgVPWwD%E7O!z^UFNz~AM~g7zrA%HuiT
zC<Xq})(|=9&Q#OhRqRfr5uNV;8aO%GotA7i*pd>=xRxX3;{4WNzbeU{6!-F-5X#8Y
zpT%zwUE+pBFrfA)Gq3NsGd7vI*=weJ4jlg9t-*lQTnIS9rQyyAU2$9)k=~lA7=<qW
z506*Qg+QA)0Fvzj_>;8-q}ZVbWqN#Ol?QSOSB)&5=mE+B&##hOb2jT3|AoirR|pH2
z5SEC5cybtW53C4revI(co-7+Hqwkrk>={tIgcpBv4Ih{&yaw=CjamEaPUtMF`8?b=
zCS{ZaPRRK>T^P9_{Ll|^-=5z~YqNRuocY>8pu0Eh*D(`x;CDGd0<<6Yp1p3u!Y%uA
z*Q2%E>&envXZ!iclW^fVx^UrhH70HI;)C6s4F;8hAEtA=leIWN022eMl8R?R;_G5y
zL*nzeP}$y{sD1?M?7{xtqS@3^ZT=>?dLoa5mjA4sy;Eyk@lshi+E;)+ZUxCA>^W_w
zGAO>wKIR&8BA_dVpf{RDc^lCDOOf8h7_5IoiJ>}<unnXwe0<U_uYl{kncvlr(#c6t
zS(ysrBuB2g`u@V<N}th#gi?!kHm12NbON@Oio6%06#OV|myv?AxM;jltcD^4%o2i=
zJ)1N(@7Yg1vS_6v4bJ~~i%v?p-O5Dz+erqKg#1>{d}6=S%Y89U4Y+oS0sT6KQEB(<
z_sVpl-M@cSy-+J?0nlM23#e>TkR;-GHj__zE4cJ;$&5$d?FNcj+%tzN4hI2ZM1eL8
zm<r>}Nm#LQ8RfMW(G1OBry<<WF}b)!1u=_Vp8rd0xNiY_ghTJF2P=YHAsah3^jcsK
z(Bcq~a9E|HY@0}&g(lUqTC~}YW&rfi!l{j2u3IePqLQnykTv(r`EVs9%@VJIE`V@u
zWwzESpzen_I*sBFu>cPbQPX}d%PBD^Rs>{Sd%LDFC+-IE@>CS~STHx?GNh7qbkpVX
z)HdLQ{6^jn)y0zx#s)qH>|)2yL$lEz6+7Rh-!o-in_Cms=kw#@z4soYSfcgV9jizE
zFCARK^UCjg5{uo99EV*%L(Uv=MbN)q*Dn!ZBGLXx4jA{37%RMT{19XDa+YjTz@0{6
z%Ns(<p%U4np6{2CK-yoO7|=TvTV^>xalCyQ=2sqZv~zunx@EGj&9A9#)&`s$Ma~Nd
z{4)Gv;kJSeQGOM_W7sQqty*u^A9aRvYz|;$a@kCgF`M<9Q+vIoQZGmhs4=Dit*FKA
zLV6$rX80DL`<c}8PuU1O<t#b%(v!>6&AbCWV1Uz{SEqoxvZ!%ACLwd<PLtnF-e~q2
zIR!3C_sbKxpmm5z_D$IFCXi{~b!p>gewSScd8Ax!SbHD2)pz}i&JOJ*O#iYsjF8Ee
zu8Fc~;_ZGB9G|U|gQt`!n{)95W97|%gWY2DeY`H&{j2lC9Ax{^X0Kc0-|x~1&&b&w
z`4-zl0(}Fq7kcAMi;9aCpMHO3k<rqHN%|7+m)5J!G3)6eNRv>Pl?zl_<WMR*!CUFT
z%WMSo<BU^--{2%*Q@-AM9Zzv|Ny=AfmZWNp=4##MQ&|TAyiOu}>v4cTF*TLjeA_A+
znyn!G^ed6%!Cyk#jZxGbrJo|LD95TC;|02%-om~^>26#OSoJ&e?mewYE`x(<%0yex
zQW^i-2*)nONa)5WhYEG_T~8TwM8YO#)rvpI{qiLqXW+j$h)cTj?(M{sh{Ax9W`c0y
ziv@B~eTe07N64Mn@d`4gG9dEnJBR8WH1_msjELVZVcp@j_L&CFLh}OjZ0Aql9aX~0
z`5_+|B*T%-O*pG5Xc|3gE2^B=@gzbh(gi+5@RVDMrX1T;WmiN1%NzIcnoZ)d%rslV
zO;0R~g@QiHNk|l+4hXos`nn0sN6LtERr1oX&3fY=dKFxXV&w<@2K}W%g^2hu6^->&
znMPXs86e>s7%ne%$0r($P&Y=e9|D!v^R^;nw|b=5@%RlU2^+rS>km1Lf4nqBg%*!x
zE-rT)?XEAdwuPlI$|1R4vg@Z-xe``H_c?wIL9t9zdV%!Ss~;F8*D$N{P|aXxrn>6m
zKEg_WvdJHhDyQ9Kl8Lvm%t;ojdFq8KS1$3=ySy&0t|ISNd%UZ>)`8qJ^`pcooLN_E
z`!HSLw)@&j5765qhyi;1Nh3K8r@qQifW?5vT8?SV>*KMhaGCj#dm`PnD{yKb{M{$T
z&IVXzZ6poxD-p}eQew(vwbyL}x(h9YqstYsS?+_i{D4=H$#6<D|5KH2={2eZ#tPey
zA2FjZG@t*wpH!hU48ZW6K@lrpkS$dr(*LU-9vfR*Z>N6&o`bz&NF9yu)eY9srU`9g
z7{wpS0gRkoBQ0(!3=E925s?e2gL=vl2ijh_{zO4S_>)4b664=gl+g8~_e3+pz{X6-
zIDa`Pzm0l?OPl(puCDHm66|WJ&r!-m&(?@rW)JX;Q#=&;x1dxrNV2z3y_S#wqRe}I
z&gulO{3Q*7-7k(#^eLg1l+$|dN6nI(;HyiGR3T3?-YfOE&dS@feVVZX?X0<&2V25o
z<|ngNHXVapWApcn1*<&3^8ee<jwDJ|(rXSd1on=gbE+6R0pn+}FL=nJtE*r0HOv2W
zqBWd)kYO&6^S>IHEM1+y8PgrrYtVTO?7(xK4R%{s;pQTdw9>X`^FHC(3}PFfp8-Kz
z%1ra#=l(h4P9T?t@YTxQ<3~6^lZR9M<q#h~#~?~`?VDN~n?(aK<ih-2p@7HG4^>M_
z7dBkIJ9#<<!2fyR(DOVvF2N3M)|PTQ{$6geFy|*7*$k}Ae^Ai$)3bVF3sa>_SEQ1=
zmB>?Gxu6?!*3o`Rht^p0X7t2KlSf%=5()hQvhQH79Pb??n<D6SQd)?<k8>XqKuahB
zf{x^TLT4sS+jv#ujkzHi@C~>R5_vpj9ge}NoB&|`zXuS~Km2?n@PJvjAsASljQ|@d
zaA4FCoYP(;|Fr{)^rFk_<m$EYV7*o}gTTS{G{|u*-^<Ovl$dlI019>rF<|<R`psu#
z8UoHrlA^-@4?`{MU%o!|ligzUS0E5B(jdEfQi<Fz>Ez-+b&uvQwGGi|RJ#*bSVKdB
z*E~n&a_+Q{e<g-U({o_O5jsJ6U$k`qhDTNZy8JKu!y)&+rDE&tKi$pzZRJG#AQm5B
z13__}a$J6w533S^C)-zKr6VYFZXcaq{q+JgZ~L36bkh)$$2`VG)PkuYpv@2^JfCfJ
zs!ZzwV3`$igTRq0=#<PpGdpXyeIO24c>QSUB%X5DgvFiNT5h{VTDnlmKpSk`(PEGj
zQ$1TdCcNh1EK~|@vf0^e#KxfZ#&zTOgnNWKU>GA#=So-a562%2DUBHU6ajIbKtTD;
z_vS?9Lff}x&et3LmiPCLgr-l(_ZHh_fDTJk@5XaZ>gENk?+lm&DVsNY%3)r_KeuFl
z7=Kv|5QR2)w9dMnV)~vQu%`XDxDDnMsDS8cK9+&@75{lKyIFtkA+RuWCPE^5MXaG!
zcI<y})?_B0xHPrEXe$N{2*GKPhSM1S^!xMrKt#yY+w1m*|FcrWZqfe+mkQbDc6DC2
zbEpAXS@fX0%QNlCl1mutEcSLZrt{j?Dj!+!x;aM$F0B^O!7ZHa2D=f@2-@yZk#M|6
zN4s9i@0*F`f&(oLg7!3Hc#)vxtwz?njh=e#607*!&7{H#<c}Wq{js*Z`-|=DtIQ=3
z*afr|68v+E^x%G$=}Dot--n2^U0DHST8Wzg2g_o}0COUrBfh<9pC6EJVwkme(&t;m
zgGLuiuP;ybBPbJ-n6xXOxE1gLIfQqmFaD0I%NtJScPl{Iyc4PInX~K$jtT!7Hg^pf
zg#$2)N&E_Tlj{ntr)=FYn%S2}|7~v^wJ(n#+ie82<SH{BHYV?OeZ>tj$^I{Ng?y-x
z4p#btO3em@TN+re^$mb)6|iuj)ajbTbe{>CNd61gMGZ0!oAuvS<M3DSl|(&lIWqxZ
z1G5V>z!6*JzK|HEas6C5Q)s0(Ua`a&O*I;GMLkDHTAOcNbsL9-8_vyfTLA~I#$o$Y
z2<cePm6fAob9LEdOhhF?zXkP|gXJEx3J>E>m`jDjUOQC3U&m*-?~wsgLL;@_(-5g+
zy~eEq!0c|uMYs00(fpTRTSOqm`}}-NSuic8l=eaE6~wd)qrc;8blzYsL=dzWgwqvP
zN*I4hQevq+T!F0hr*jGgq`prxFTOm40tTvUu@H399l-Nqw>#(IPy<@;C5tDsClTqA
zhS79b`N2Di8KMS(_ml)J5OX|ftA1}ixleLJO$g+?4lOFyhU%G=2r!~CXfzIPdE6f-
z)$mY(&g$%!o-X1rfV;Rf5}4nf>rRDdLCVZWNz%)3U*9}MD~_U-#?#Q!dIwT*Kv-K>
zENz+-ZG|UjBC0;Fvt_<bz^+yUi0mOqDLS>D<hr#E>L#>S1A7tagE>a11fiXco8mEb
z&%Y~ohLY3BI8Dm`+ikFlKP0D;OPBryEhx}~NZ+|2p>{@{?*LJst<d0}!4={EX}m!1
zi?C=w3*dd0?~oItehm_6bJ)(v1U}#cbCZ(qRzR;87w`^aaB(?7CoOIP-C^tGq#m?D
zLKIGXR-$18rfgVU|30__s4_7(8hdIjx0ruSmt;sdjf=i7q&&;<pQ@~6`2wX6R-cYI
z4&(b+omsD0EUc~V4XaMp-O-J|>uS5^VEP6HDqEj=x6>bTn`2eBJD7}&o|yQko)T?5
zQ<b-HQc^KMle|`G1rvjXEi)4n6WNnz&HOH#71mSv7mcje-GAqk7qb!r13}YL=MRLC
zi`v0W2kT7F#kMb_vR+G3VAAL=<u`98(F6Ge8G6}a(r0)HM9#b6E<kfCmRqRWui;(7
z2Iyxv(H*2ut+1drLNpb;-F9R0(Q_s3=HcOCM&9d#<rhFP3?^dHL#p)x8nqbEhCD4;
z^%*EZsnU9NSv_&lKpOo_@d-Y&P8YP<{rQXS$Rhp@X_39HiRzy4WjCm8lPu@mU`gc9
zo|fijAK!D+5Au^(_j~*ez-wQ;MZ15m2>D1+59X<}eB7PW#R~7vH?x0pXVH*7E#Dk_
zkI~<Neq4}b|2o9|{2<`M!Cim8sadqj?vRcaQyzfBvi4Aqol9oQh|}4bnQ&HVwpSJ-
z4}hm8s7u#y`rmPYL-rH_#EmCHkT}e1z4``VMdLUu3DwMBE}0m_eL={i`=6OWrcb=K
z^Fqp&6w@#wDel+1_E%g-4F=8NzC^CDw411XZzu~hvDmitK-*dk0f?}yE*mO|0^6VB
zr$n>9b69>?y@m0)2!Vk=Faufk@INN}MBrIiSb9?gDGyir+b?k_70RR@f<FH(6J7%2
zdnMm4CkR9=1W$Z>bNYZ>yyp=6?<HW3z|eu%pA2bClb?9ke3LP*KbfUF&KZ`Vf*mHG
zqF782c<9SQrR(9#1<MMEH7M`KI9DIXaxoj#{fZrG47to-U58${bU>B#Zm+v4W*jsh
zbDQ?YIRo7NHkrp?W6E)rS$8l^6tyjcRQ>Bqm{=EM1L#7sgtQ;l3~^Q01g1Fcc4l7l
zxgI?Lene3KjWMIwd`}|NAU#ea*^2wumrkuh75g_pn=EoUG#&Y+t<H3jOBuxzTzL#{
z06i}$q+eolO?F59MoH26_j|{6q{k9Ii8Yu3AuP~u3?<>PxXV=?u^2`FH9WiBVFKRB
zc2TLL03GW&pBq11){eZn9I6E6K1d};iCXUat7i(SaA0st_4N%5efPQU&Ja^~n#>O~
z2=u@RKK@wn<C;RgMlr&0rkpay_W-*`Pezx)4h8^Ik0BtcB66Wy_e@wSFiVJWKnwoU
zrNBz=14hSa%=#0ZFOHrU=+$>OU$Itk^!N93dtPzJu;^2us4qQ*DJ;)$FYXZYMMoXX
zV6qS`;n`}6*I=Q(pw(;*0pO5Ck2gkXiX|qTbX#^Nrl&<IvW>waZ;>{=_zy=45-@}a
zBE8gvkrv>PgOMG#r~IuZON0rVqJ2!TS4cPg`YO(3oM|)*=K&jIXs@W}|F0be+DaNN
zpy|_)dMseh&L=|Wo2zPc;YSL;0U8tQ@)v^Ex%dJKI@^{7ud!Y>o(<k1;@psSumMar
zdEHWgb`bmEYl-<jP{gj7vRIThcbmP~muG-=<B8MGj1R%me|TB0%7l(m1bpfLL)q~|
z0t3JPz_o#*p`9k;Z+mgH4i}cN=ySX|-a0h&5P1F7iVv1LLM*@b1%I^u4$P^*YLnkW
z0SuVW2XEw)cnE;Fyk8mY>x-CC(DY}qJ!n)L9x#FE`H#b30$O1{3!KT@VPifRBLr0j
z5*YFM)ZlTMy8#ewAjn3BVARQcy5d)N5FwM6lrF=_W`yRXZmnZGxRASu=5Qw4AX-NL
z#DojnXjW5YXyQSHs;6{d{=g@hl-*PooQy7cVu93s(H2aYWXOtdfO*Z@w(=7uD%{b#
zs4VZ~9eTqSzYT*r=Y32UpuI@;#4wFonGrL_ipznI{qjWoYj)g`t2NtjqhV{n1#^Fj
zU>G_9L(`u>O8&KzB_=@&8s=7vnk7iD?N?M%z)pN%6?_#gn)=q#(y}+1pA-a@Z(yKg
zDu)DU0s&q+I*eddh*bNf<=-9sB5x_0LG{xW%9TutfQJ$!t*ey*Z5;R~$@$z)dKF&f
zd*uaS*uyBuR3hjAZ;H%ua)N0sB3d)~|9KX`e~;{oQTTvvD&1@#S);-Ib$1L?9wBEm
zD_;jjqJaAuE?YSB`pNe6GrhX1uYD}5PvyR=FnzA^xIDp?jN0t>MRZBEoo|B2q!b`y
zD^2+Q5!v%{n+s$WQT}J>`Ozie`@gg+-Q9QQ>hs=4K=r`};l5B~3?}8mmX4$q2Dsxo
zX{gb5K_r-%T_QR*PzhXwcpx^p3cumGHdcNiAB@<6YX=-4^ouhuD!%<Y6JlEM@dVfN
z&4?Vf42VB`o__hwlIE(TqeI1hRbO8pA?W4V8_$k9nx~GlHDW+{2?m+JuQMT@fwCkD
z!fk}ZR~H^5Ag?JyItt{Mms0ue^z`&z(+9_!>c)NFM>~tbj9lL2F<WZ`IH(gvhKR;=
zjlF03l{Vym(-dQwb=?z7?N{H@$|sWn<Bu8+959vnfcwp5Gu+oc9(ue3b7=mrVevzv
zY_(tCNqTYK$`2&*qM||BL7@}110{!Jcq)7!gOts(RxW|_GYJR7+aw-jQqlzd?2xR0
zKp`*^Heyi6Si`uGKwHqTOKSoyZ<&VoZ6ddYe#=UROf0r&AhtgsL<40H=A{H0l!5f5
zQ2b6DlrBAfg<P$220Gce)UW<0kWnY?D*1at0C)ec2=4DvyWJy}UOl0f6{MKCIu|4E
zt(_fg;$RN!-!zC7czL(#k-ZZ)*QZeG(pt$7lAoqJJzG|ppPy49Zh?(fvSiVzvV{j;
z?!%#nE(Xd4I<F~RF8WivZoXU`Z?*#G%gM>f+H7rQO8mL2Cfe^Zr(L~#-C9B*#i)&c
z;}Wm1o(>qi__Z+(EBtt$qXD+`YUjDzefQU^1?msw?E_0b1F!8V4$%~og<&v|Yy=vw
zfI5?&m<5pd1FlnM!_Th3=BQz$b~^&Eg;xc743{uGv4$YX{{YxxqlG-0d1XEI%DDR*
zT1kVipHn0>p>6P#C4_>X1kY3w*d#GMU8qJXA^KG{HBVl7=G~P0z`F8RXqv?B`d0Ny
z=>LIcDy0X;zaz!5esqhZmvHAWzG2uuI3CH6egxzc)Ybs3R$#4o>2dLqW8wn%rR$}N
z8t+upW?$62@aK4zx~3ZTtb!Bn4}YX$R=EN0ZgCQoOH)g05u8@%gTH8?_!fw;{;pOE
zW6ZQnG@jS<<OS?U`{n8G=<AZt1tSk58&#JqKfsQBxP$HeUv2V3r6#Axq|i%1^{HYE
zQSf|}Y8TY6cjd4c%g-UCkH!Y7fgaX;7};y3KJS=jl`qY}GIN`9!0mEC3oIoe0Pie;
zoMb-qqg_5}pyi>ckm*FvrwY)wx9NLb|0LPI<XWwz=XX0%Dcxo_?(){D{?Odrjjo!j
zqPWS~RfGPa!EY;q_T|y?HZEEpO%N~0o|^50VF31izE<<747>cdnBE90FL1@(wug7S
z<yNNkPlEqt0fi=6BxBC9w*G;*m+6iCuIC3o(Jh|@#dgzINt-J6e5zsV;srK#0BpFb
zvPyK22)S+N;Q)R>LCg6DtgcjU2LJ+r;m$jkZ9T0EjhLt9L3k7QZ{XOPrUg|_BAL%6
z7$;}@_Cotu`5~+>->U~2WmB)9Nnx0}F^|3lZPf4k&a9e?pZPWk9WA5<-vvn`_%Xc8
z;nIAwZvgR0MAqpREDs8FYj01nDj6#pm?kkBz0Tv35V#;dVXgs38zsp4=%bs(ABOS6
zHBSj%sEGmS^U`AE*{~I@wm8fg&Ov@GDyRhYbWFZqAp=mLx{~>M0nGNb951kA6~w!x
zrQNv(t5D>;{|*cU)8ok$ZNe;eCIi44nXYCN%rLj6Kmdm1ImpP!Y!_OPIlQ$&B~+-V
zcoD{p0Q~TskjCo(zz^giLulUiOe)>K58j>#SS#!sTem3h@6VgkM-}+OG|x*?cK{vT
z-0UtW7QvGav{Bm+AC!DAfK4c5gm8-w<{j_v%)akuaVi4t?~Do`8EnrdkRTU;M^mHb
z<8#0g5PelQ8hxnQG(AnqVL2ui!>GyXa`5-_3_%5qjlC)OZU8FShOYMBnE!0P(Hpa$
zhnJT|KAF#=dH8U3z@8QA#jtYo0{7_~d`A5SGEz<}@BGrH`}_mo0dQ<$dMH<6SRvkD
ztxz@^Fgp-gPgk^q#VI!ajW;nl=|5GjQGAXDOj0P|=KX_$Flfad!5DqvkXw|!*^gS4
zQ^sMK=LZ}<5NJb=!gGIg0+H>ha$tZ$!(IveUU9CbQ)D2FNzS9vK70$F!m+V-rxrJQ
zc{#hd#6vn^`5XACJ9|r19g`LQ{WUffKG3Ok!~$ti;_7U_*nH?ofBp@DY;qpb#&}`a
z-``P2f4d?!;D-Rj87h0PoUc<IDFPEC_z&3cObNssEHJBq8;JmLm&91U=1Z#yR?%`G
z@HbBs8@;q&kqxF4qyU%CPPT-C>?wW+al4{l5b4#s`h)lu1*fj5tF;wTG5w%@Rlj(-
z8s?jtiRJ}eP@rF*!XRH+F$2(MS)}~|scQk0lw4K3O+Ju6xSe*$nDrZao82E(tFQbC
z1{0?Xz=!C3ds7c4hdqC{R7-}Ey^Lipp@jK@heBZ4X`1BT!9HQ=hQ;>auk<%2Z8fG3
zFA#G9a1&kbj#gcZdOw_*2F%>IEWJ$Ve%?Y6PsDAHH&i*|AUa9Cy|I--p!E+at%m7M
zg?tnqfg_$hRPyfrSonk!-W?~<2!Syrn`|CXzZt+jdlVHFvHk@Jj`}GsK&Ne=A0h%1
z%$~9dIJ^rhE19O7m#@2EfGRw?UlHIG{8*pm!Lpy^K7iIp*DALF<IEkudwMV`ZDv1Z
zWo6lMRzAmlhX5Q`VxUub2q3W_;^`d~3!af8x}}x>`TlHqBlv)FE67St0rSDE#6I9z
zI15wdmH}tW9avG~Z@`H#Pi(hF_lxuP8zbhhzy}B{9L#ls4s~3ivUf_q*728N3MtHL
z<3Est512ud(;b6ZBittlypCH!0Gv^m?t9Rq`9CpqG7e~947W5hTdjR?dUw?zDYOI2
zFstX4v-`yn+T7fnRKycxy#{yUcjE@)|M3k{AT`7PW=*-johDkO9cW`1ID908w9(Eh
zco$+VsUXND{TCv{KbS21SsO}E5KEYRa7Yv$C<!olYr13@)j}%BE&iZlCwO&Eio)!k
zEdkiks)NJD|Nh!T+$x~o^e6KN0gYf092pD4KhG3@WXI?I5shzVpngI7yD4`6rN{JX
z<+oArY^VNMd&Fr@W|rmP(~yBv4_Tlv?z?oHUpOjsdg#jt4xI63vk8;O)_9haH&;t2
zn0&GjHR$x9WN(ORVA}s}!daF9s&4b2KX{rY#=@Wl8?C^czdWsFEzpLZTr)Xrj7LDE
zFLsqHe%57xLBA$YbwSvQH-ADlG%`v*IH&)J44%6a1a5H=;QK1){U>*0AI#5A+s_1~
zGp``Qo?aolIePtPReEpxhbxaEzlR56t@(*5#}AJn@0Zd<1Me{NH!yu-f~qptSaa-(
zz(c5@`$12r^F~Hj(Cw|z@^ufFO;*D%FVa&{@I0&I8uwk9h8bn)i+WYwDa!MDVC&Jp
zJ}m<fNh*@|i|6P20rMtC@IWdEVP!<pfA=VU_1=EpHL48YP32P%!YX&3#XW!hhb|bt
zP(CGs(UY-Dp(kuX!8-g1+RXX>r@+|{oDQ+*=GVk~r^iet$9$7xs2>r)Qd4TYe!Wtg
z|9Aa)k&zNz_ELxS=;qNixk`L$XrLJHC;O1pZX}b36L!$f;6jtS>$YwFcVb}y_-OBP
zpg#M>-W4I6W8isUL1~^1?lOP8Qc6yZs&JhzZ`j|p2m8<`>)W2(KtHFjzO4DU>EpOP
zou?qUA$EiI(gzV0+n{^*ku4a?#Qz;nT%q=Rud<dtpDF77spL-33FWr-P8RLbHRY`~
z&8gzfb{GVDstb*^KeoQsJ~#q_BHz(@%y&N4)3EPccrEdR#xqf4<Nu-Rtpeg&ny%sC
z4hb%U1PKlSf(#zqg1d%5aCaZvErWY-hv2~-g1ftf!QJ(r^SviI&wn!)eY5w}?pn2K
zRaLj0;_s+~@rJ9WOYbMX?@ilo!u_x%jNiSA8NVC7wa3`vy|jK>`bq9Va%lu}B<m9o
zdLfs1dBG}{oNR@FsNBI~Kd&w^<PBOQD1*?lUE5)TLzsnR&RGA3hR~n?h%rBJqM82{
zV*=AUpy~ERz9`>(erV7r%~Y=9y9hf=G*&skANV|uFBm&y?xBoht45=bK@8gpnhl+R
zN?O|+CyO<F?n=FnS%~>c4XR{4;hnE#cHAnvOBAx&y?Xno2-w`a-)Jrw2VHda?jL3T
zXcDaaPv#R5h5B%I_8{-SsBgmb{_8olT*`aIDUuxt=_Oo~Wff?{j~`mPoi|1ZC&5Ik
zE9i1>iS@LLpSY+afl))TSOn-)q>UWyENi}z;1ei}dPz-|C;T3P`f_qON|Md>LE37$
z1rMqRVC<wY{g-UYDL@}w;HofB`G5ZIJAf*e`X1l=9;+*XjWd?g&??TFmPEqu2e#YA
zF;2-Rq41`tRHMo1Jws=$nx<%ISceK8kw5{G9O*lMlNoN8zFK&fthQ(qqr}eM-{{an
zilugxk>y`u0s2PJyi1M_|6iOHbMYgZDm1(gRGOd3rN-`$B3b~>gHQ?K`CHpM3QkPq
zGB5c&uJ^5fK_tdup(8Bf<`3AKlMQT#j;5IR@QL3$SWyA3UPzqxW8L|(NwJI?6$WKd
zg@5a*em|gdpg!|y7vcX0U4F>4cDv(mKNK4vJ$NiSn)n1Zd-{+(o(*v?dmY21#iP7o
z#W#w7z`MUo*T^JSvfqEhWKiQ5VT%}d0u>j3WpMEIH53Pyh*u+BFyZHRRH9h{U5Ek_
zk9+eO=b&|<NrTE*&cC^^-+Nr>!wJ7!DF6HLP`cBjCGPFoA^(qGnU&s3TEy*jjmj#V
zb#<dz9|B8nOl=3ROv&gtY*_2B=>obwGMPQDU-6akyOZk;e*HM<R^@Z8{8s;Y{fa=-
zYa5eAz(sO)8fBtb4nKZ~?DpjRpy7nB{Hj4==CCM%wLJS%5?Ld(Bz*zxa5Bm@qVI_!
zBCWnXN-o#1Q@N0_eS@XusFUp}&exipWoy1PzJg_cG)oE%!QEr_Cbzi%w8K5O?fLIv
zI(v~JLt4X1Z?ZMzE#rTs3@FGR3%!sGxnu<!wN6T<zR%3s5{j#O>QUndZTVeuhQAD(
zzf*I0%D%8$^+4w)Zxv%7$@QoL5H9DSZuW%3*xK64-EC3SXZ(ca9=oA4OCrl%ni1YN
z4tqQbm@8x@tog!y{Zn^Nm;X$sxMf$r{R4Z!b>nVRzb9+FOL{Q4k=;<@1~Is4A3YTd
zXp(-<RpC1rR6}G+RPQ;m;{S8c@k`(ELNNb-BQ<`Z6df}KlwEpXNd9%M4Z0sDBgE7Z
zXES+-7eS(k3L55cBXUBlu@jVn^*2lr0CQ|4_w#P<6E?@SPi{W9*v1RvEOz1~J+`f<
z@J!e}UBe~_-?Titzu!G>@f_a|uE`47Klmol9%m0M8`*i^VF{Ysz9THOR9F)k{6L>N
z(N9}YPjTRV^s&#D1B=RInl5>sj=TgS9YEh~`?)K+v`_7Jo+XVLZ<0DR9O}yVzaY^D
ze$c&cIMEeExY&}j7O;`dVS4fAr!HvSmXWN@uDkXs<Pjb4pu-b7{A6nTr8jfxV?6sx
zH`3^e$itcjmHuIJ@=PvFss=;gLR$3aiD;aXiT28cm@v!hY^`4Z+g4f9WzaKXg6CLY
z|Jd`3t@H_yq{`!k=mUe!tC|er&BOALU0&x5a^GVPuA-(kuO}Ox>#>j+9;XeiLC9D*
zegY~9bfncw5dD4ncWZPTbr{CP$_}Wu7x}+TP)?s>(T+LAi*Z5t50&Wp^xg4+Ygd(r
zh&^9w-Ihe#0<OfhwsUfmi+<Xk$Zo~sxytZe?M3!dvgc|2BIxj%OfRYN^ZGrSX5L7(
z&;B!vIJ*u-Fqd!B@g<zxiYMHOkhge0P2o32>n~Df$!ANI**A-hnl6Vd&(9E?a1;4x
z;DfE<RWXv=*(@4p#Vf2&oc-Q>$<+FXY0r+Vowm2So-CqSjn+F6e;&3|@w6%dlO;U0
zle~<%T9}QwUm08|;NBR7ly9qU$4Yf8u1D^Xs~$ulHZ6*CpKu}li!eH@{1S!vg={tt
zYPxQ_!D5T4utc>V9i&M{`KK#9iMqM<x;AuJ&#oR^x8>sC*$G!HI#KfE2>U<NdzW);
zv2w<7Cb>KfX%mIuu@Kh%M!nYau9zVa4eC9c0y5_=v>3TPY;f!Ve&+Kx7<l<J72|X=
zNb9QczdtZN)+empfJHTy3%((5N1R_ns$63fy1B`+-yF4^H@~8tL!2@-e-3zV30>pl
zOY#I+-n<8H_G3OrUS41=t2Tdq^=LXdZ>@2rjAmWa=t*QOMMnO+F6B-(dNqhiPKV><
z#dP(+nXL5*^#ZGLliYge#PX;?g)<4Snna*y-eDt*fwx4TY+F95=QOMYV{_=_laCI&
zPJ3)!kVo7`FqaCbbx8Ec8bL>FYVbA*+#40Fl@^_&7b5>mYO>l6>j)YC$Zm@@&AbOf
zzi`)y{<6DCYNpE~1kJGh5tgEJPg+Lk5~*@iQGAY`9Y2sYilvF$M#0+~8EsEy>}F$J
z%4@a~NV2Yp>!;O9CGPg4-^<6%txqtzmI>HliV|3BaBR9}HU+;$YY;{J#%;NwF<Br5
zYblF1P0DYOX8nv>5o54HUa3HrJpwy7B|b-Mmt2o&^WxQU#^`F8^tO)oWHW<y_k%e`
zu=&x-mIGYaX8*^;TKZ|Zi)ih(VfC@hvwj`w$wa8@tWU^tmo2+WjEE)q?H1sn<}4KD
zyXB4;kJ+Zl>z(i95}5ILRi%bKZ%mo~Lci?MUU|yk7QHv%%3_9Xs$bpt-X;XP3;!g^
zy0LuJ|0`pFp=eVudZ^<G3gUKd1vu3O3i&G}mcJdgCrO>kv|urUnMDpt#3v-vXtEX<
z8co2;7`Kog^<uM5f-lV{_uF_)pA2H$TZu~@s|#eqL>AcBORs}9JInUoe0q36mX>wR
z=<4ex_{MH7qG`eL^gHU6G&fnBV*f0rZ9vzeySvq7QA`WxYen$G`E4B%S=+VP0QedP
z&q(kuO?nUfr<v=2jyaW+>{wYL)u_;y)Z)tuVC`d$#Ni<3%Jj~wHtx_9**F{*WdFrr
ze$m(;TG?xUbN`6J`P{$i-XZR6W%w9@pco!5RYe(!9Nv1f>KVZ~t7-J1ewU-Kr${Lm
z`cpMm#MZa!dD0q@GcW$;O^-j#NxIfabu*KiYKP;h?Sdy{bU5Q~ZN(efU2*c*1g9}z
zC|LW^r<QpuxN-Z#QTuIjiEBvu#ew{)I_7rSc6>c8#6zgbW%=bTB3M<1C<fXcPTBiH
z6XO0>!_(or4i%hqN3!F#LrB02Fgm5V;=4GC1{ApU>JCIlz3boBe(#9Zj~?EsTFway
z4sJi{F>;6%$`=pZ+RlTfYqZi$)0DLD`GP|p9hK|8(w-#MU{RKOYhRZB#)$rGs}Mn4
zmhBVSH;Cyzq)N}0*A^o~Hhi%~cgkOz+!29#csggbbl1!yQ_a$nyzVN5>r89aueUkg
z>R^-CP_H<#n>MyOQ9{y6eu<U`%<KTEE|zMO)xDm4jM1|IuvKBAf${;D>BF2&cf5;Q
z*;Y0yBA+-NGWqgrl$$Bv*);j?^g4bQZPewFuI7A9|H7`{qWR;5p(Jv_Uwf;8-V`Y7
zAa=z*m$eq481FD<>@+Yk5qu64NSI(w>nufSdeKp$x!jtrZhSd9EE@JPt36r~_k~QG
zpBK%ft1p5^y;!9Lfv{W&(c>LImj3Io!9h>n4R9qZEF4^J3FU9G&sX6frFPZnb8#(>
zuN}`d8up_G1OlxdYvzDvCpgbvTEl(vb}miLyr!zr&-9EpMaXC5ywlK@%GEhC9)wN>
zr?Ie%wU@NERKUHYQrl-1wN>+B5uVuYmj{=+*x5-~W}o!(oa2L}0W7rp=jy3O9_b#!
zR3k(k@6FZy-qv*y`QMBUWZxy)h3rD|96g{!0lnUQ!lZ}air-0JJfgh-q18`CgW~iI
z=TRKfOJB9fU{-QqY};u-s#?H&;zyZ{(pK%7wJxn>J2VvwDFxPWya1VH&K~(9^<S;h
zbfDo+tmqt2mAJlmVx7S;3%k`lp=9(w2kG=Mm7Gu}rPZ)gOZQ;&rnVLYDMxoHWU$j5
zBLt0qG5qOpZaKGN-Whn*wPae7aM8SGy>denA<B!MQbpYI+QA=R@!ne>0l!$KxCHX#
zJZ5jvj{ig2i0Q>SVNpl9^-Q)td+&zh?8;WQli#QsG5=!H)ZSfgX1nS2YgiFsWMr7P
z6!6%1Fe~lR*fX)l@Bb_X3V(TP2UXK>?Hol)8YaH&^r?lVKQaC?f1$Vpidzb)$}2|u
z3W;GtP~FqxXWca>fd(G)2EbBotfCUi>_*n;manw?KZ1EfW=R|XPYZ4E#CQYIv||P(
zg1qsLk_VH(nFv=DyS#TNh^M3JWUO?k?J7aK--3HbuX=Gi*P;Mh0sH%uJV$iwFB%SF
zYuS1*Whi6v7q%9=fbbSg#r?vl4}Ax87qfmVl?B|0vje5~JDP__xcOs!=NvY;q=L<k
z=5;l`q$>%Mww-UnRSQvy5quXiWQrTB<c1g)%YOu=pOZ=*wEJY2-QLV&oF2nOsM7xk
zrtA^OI9eCQeA2Id-J!IW7K%k%@|9M5rb<T*-s8$EWv2Dh3+|v{3=yeJiRW3D=o$Ik
zR7+)X5WGHNXmp*q!OqSBSP)~Lesp5L+>8MJL}cFMxMHz&yseB6IIdjs>*|x~Fe~q=
z0MFgAZg|d12tK=8c6P9W$zx0=3dm-E9JY@we3siZcx$4<p!o$Jp@xX|$F^w|aJ@pE
zM*d`Z4T0D|dnkTTX)FuNM_;GCJh(;6iBIgdg@{@I^vf9UB0^^$8l@7EP`5uY!|rgN
z4O(ubDzV7&x(k~!Rjf4#<e0}r=37-Mkn#(FR49FDz=_svsYw6%@M~bLO5Xf1pYAa{
zh@Vbd5dA`Ws*M0R_0UjM(qw3TMQISNg!yQL;S>R^-+m>?;fg#N*PTJtqE*@Ie6>lJ
zz)hQ~$|UyYndJ3r-0<r<navR3fg(7k1qi!3&K_6O=tD$4tPgr+U0lux+-{CB$mwhJ
z{$k51_`IsVd&Ov?e93Fspm2NvWbk@ogpT5ngs|S`A~M6*mT1d4r?Hxv*adnKgqy^a
z&oKTD00t~kAXJ6ztrmD!cY9BLL$i2fN-|Pl(pG0K**15R{5WIlLz_afvRP)XGn?~b
zptChrV@PR~*s{lAPGlp1+*fJ9`4;I3ZU%uh)eLH`!%!?QKmp#~n)_Sd^0!AK>s$(%
zL0czB{nN0cmoViW5hSEO`ce}QG@WQD1uj9yH~yLxd~WJTh_{{(_S(Y0@EIXjP8|bF
zgB}hN`K+{3AKSZ^n4Uo2zwQ-_iySH{M9)t{l>6_RUyRr~R5Gxuo*5Vx;|vVnt~E6>
zeKvI1OuBIr8;`RCTzY;_e>luD?SD3#Bx|72!Y79th}*OwQd?rgr5{NQ2J|?i8msj)
zZA_NLmx70YVx#9Jt1Jxk^s>1<#!N7m_ah*>Ps;C8tK?AIuj5)U4W_95Z<|P%apzSo
zIb%c}caz@Hwwo-lYfnW7*f|fPC<9R#hv2JzQ-_?_oxPD3aChxYc808z--?M>I+8NC
z)k#+Yk|mY)J?S^I@&o39t4a8n^O`mxkYQ;}PNRXhLaDN<ecopBMTZ|GeDr8&>+p_J
zL!&uQtlrlZ?O1A8qA`zp3A%HKqw7ez<FQXU+N(DM<PNIRm2ccbrgP9);V-j|tFxwT
zP@?bW%0T5U_Za}~lxrKhsN4hr&=oN&{pP~B@;9TnEIqF>r3ov-_1!4m+S9H+K8d!c
z2cBoa?Mj!};dythqo1G6-^zESZ<cqn@*Rzaj^6ThuaUA{z>A2ACcw=^=Q-rCxAna_
zM;bFwnWQgET5WL0qbWTP7K8(|3bm-{>E_URNBeWf_u13B6GVHp%S~LQ+KUPBT`bH`
z9Lr5yT_L?l%Rej^seiyitS>7u)~&s-@-`hHS*lHxIhVA&i2Sfw85@XBB71Zy(!JZx
zMU_#9_wL%FL@nYh!-tnjYf2_4o+FF3$H}(4uch1Yv{vmW=`A51n#bqshH|AsRYC7d
zzB$tArY7(sn^hs{#ipog{c(p@9omM{E(fQOKJJ3gBe-M~Y!uuV(H$)BbEN%B{IaI%
za(uS}Bi=ES0NG8ld!1`W%|WkCFy2Kqx#p3Ypr@JHL#cXmy0+Ut4g+%XJ{){#({*1Z
zFwCj+byhvG#K+F=J-cz}-HT4np%lyi3kUzVdDr2vUxz#+1u{2a{~IV`r;!7Lf;KoD
z^=Sgac=Ba)YxfpX#^fIngKFp-!5IXw@l(x;E}r1ND#bUWW$mN)+o`P{<rkXXh<aMu
z;c}2)&La{Z=*6_}iGHM9Zfiw=D9sNY*{65538P2Jls<5#wG#Dn$`=2q@`}v<7PIzi
zQd{higZ@hoFMVPW;8&T^n)xiuc0tjD@-e0bj2~%F>KQB|-y?T=Ta@W`8AQVf4EF`C
zAg`uHssqfDisw9nysGs&0T|2lExT$eF89SCc8l-bZ1B-7Psv^C>=r_LyJsRTeLe{T
z+ZpBJQAaZ_z6(b*`;H!Vi);XaId7sXZ5Wxx74ZEbYX$LridDH?9Cb80Gu>ejn1h`u
z;>g}s+xUI0N$q;q9Vw3=i%<mTnrhctt>u=CaH#g>2A%h+^=p+j-!U_kkaoT=J1EeW
z7}!c_5`lN6$~`!abi|n%bqEe$m7;Hz(6|jpe83Vpp>}OP`{DBHy};ZIC0ARywsEGQ
zTrF*x%_e(Yh2g@$+NGE~aP%|k<g7s<<t^umce#z+NvA?4L&*D~Av5*mA-v!SJSK$h
zzUBxfW%0)fv#VSWEJ^`y6^?XCtK;%IA&;|QhaCvb%1Z^tV2+0pYb#5udLaX|RnN?`
zW*geth;FV6S>)cL;*s{cAYgK#s6vyh=x{02WA$wZyOe&#H=1R@<iUPJb_Q09l=;_h
z#LOkQYM4(`myrZkf{|q@lcoHy%V)g@FftxRv%UkG*D9u03+q9|Bf3%s?WjjTcy=s(
z`u)^n*4%X;Q*d$q(vLu4<+raNKuYfthz59q{vf9xJXI{Ovjt5lwn_+|wBT0~@ifQ4
zhuxi^1kAfU=A`)Urq!igh}HG1OlpYpHR-RaF_-gTd7GfCQ($9B_6kKmv!vWjNh;Ej
z)M^mvN;&<T!w=!9(F5680hBWf(FUN<uDph^zWXB%SG_*l@`z$bec#VocwDpfrjnr1
zhA?Y2LtIexB)_JS0mp3ri`SqHx~-}&J}C>MR%%8-DdS!GtMGlYAJkv+eM!|0J4}HT
zvlb9icx~(@Ko7<Gnd74b8h)yM9?#IU*w>_ii87)S<aK`dvWKxUO+<Qz*-l&M&ANJf
z_pBdhE#QR|DS(!~@gSE0x`A`53t}KcVZ$uJ5R5gsYSgwzMF}Yyl^Nut-lF$Jft)<2
zt;vMbBul#61JU@R`N`w*Lra3#<RDM+8U5rTgcP62y8Gm@(UAG&Co%2M-#r}y=5vTA
zKR<8jAuOAkfQfmfq<yvg_dLM?G78c$3)*rEx%igT(jP#Yo<eE>K*NMn0u$?5O@nmW
zyS~OMQJ0>_s9V#1dm7M|Xw!8%uFGK#Gs7MeZMC-;ip7Fvs$>LZy9cczFnWdkz~xoS
zYADX8gYx3Z(u{hgEH|v;aVbi2!x_ar*u+yelJilE)#sc$>mZGJ(L+r7-p-Hph+=)$
zYZaw~{Ue-|0_gfX*L_nk(%~scB+QMoS!yeM^h;BN;nK;{oO07s%^a!l5ILlX6!nnb
zRwtLofoIQ$iex-GXe%BtmwsigPV}TrShJB1e>nV7Wkh;uqdicR;)##sFj&~NQ2NF}
zmno@dCqwkco5L!0MVf$g><;S|SaZgf#vx65#kQA4fXZ`choa3deWrz4O9CWMk6T9~
zjxb4&@x_7k8Q0IUf$)v?MJv$$Va)$Rr072o8Is3*^A{p}ki`<vC!B~ATGD@2c}knN
zJBkTNhSqe?+~^}Se4|Dqkvw)(UNYcH?~e5l9CRNXT}6!-`@kVyo<=NN!$eKXFDYJQ
z#7zT)fs-z5M;0%yshYYZ1O5P3(OVvn5cB!8>P~Kq9ktm>XC-c%(rHlwa~zcfW=n7E
zs?`=RTV8)-v9@sF$;ub2r;o1@IJIUs=zd0zNy8t)lt&$6XJVRRPh33Tgu6MAQ+g?z
z0MXvkXSTUT88XTj7i7m{2|J3^h-LxHsI9NV$|LrhcO0U3$YmV_m_F5nIo*ILB^%9g
z8{1$=nfM$;%_kw>gF@e^Ceav{#7PK9(n|L8-Yyt*+vEuBAbZwR#kb6*09Mkc-I!@u
z$!2Q5r*e#Ok{(g6XAl?kp}jK_Vl8u^ev}Q1gG|AG+FElh2EZTDbHmcoRNT*|*eur1
z>j2+%JLvz~2pk#?;(UZC)_N&w^3+O!o|ztnhnrQ0qa3K8a5Jj-Cs4GL_Zqed6TQ`7
zVBU`5KB!(c{uZZ8MAQ<yL$=zj>7EpBA*{XF6nLDf5ph~6v3w)T!>0aZ%-FNj5R`ka
z#S3Agsry=tr;ogDc2Y-uHu+S}B`BF*^ZYPe_mD7^MfyxdV5N$poTg4^oX84Nn17QI
z4QRZQ!pU-8>)zvLRKQ%Z#gm-Q%7rh(@)!L^-+E??CK+8r%6ozp4QnfL|KLr_+D`{|
zF;^?6;{x!<7}m@&K)7Y$t{$&!!P`}oVQE~+^(L%7`ag-H1_^{)Ck21xkV?Cf1s84C
zH&&Y{H^RDAWh-Yd%k-I|GwX_(POL1|>M_Wu4&7`g3GK~{t=K6Z`4Bc3mo^!I9A=?3
z0+#8M@#wt0jRn8P&IrJ5LcHcfyzx()J+F9MiuTEaN?~jvVEPeTe6r^_f8F-a)^{uj
ztb^!QYEhJrtqB)Tri*@^k6IFYH(&WwpC}t^==)n|h1$s}=-+V@AW<Pbi<OLFON#T?
zl#X==9o$5MuZioq`J)YSQVrFB`2PdxITRF_Phw|;#;^aRlwdADY#U4&?R=ir8mg?{
zIy@o|*faoUxf829DZZD&cL--xKwYi+z!5%K-bhO=G19o0D8S^t;*(*>lkcMY83)aX
z-p`j*21b#I>y^@N^?tQ;UZZD0AyvXUqK0=|VP;a2Oa~l6;Dh?CHY`%pV<bNFA?0eF
zRe_GJv`29iHGPL$#F^-RzB)aW{1cne?6GJuGIE`<lFayBPWUaZbd5@W+&3?L;xiYp
z@U$CG4%Z)XN$d52b2{-(YNI<!mZM;Jg?G5LL}aYmR_h2zE|HEsO(xc|%^S<*pX*LL
zBCxFL=-(dA&EzC@{*rxc8U_{_yJ&uaKz4#u*Os|>XT35;UNuy#9~pGt?G&aXFyUO{
z(Hwn+xXgu<?Pp^6EEQW^U<PDkoR{_ja^=zGc?M1hzAjHkWbCFJyD~7>r#}`~d@N7K
zvYh;aK+A5jM`JUV_3h1N%H|j`$w~ER_k#PN>Xf5MNIp@FIWZ0Hb4uo?ls6=5Y-1-e
ztaJ<Xzr{Zv=#9umCAjM6<+s467jYK^!Okc?QzOI3KMsx_{ccvsrlzU<%BVq+<gjab
z{vI$R*a~8-B?V#y@x`b+mq{y|_I0Rxe0mHmtYBdr@k%_{6EqC-$frSaQgIlb^Ew!|
z5kZ4d(Ku3C+aQ9|IUsI*(ix{Jbr5mQHU4>6WUEm_Ew6%mKi{h}9Z^~}^+*zKmR0YO
zC9^Xmt*6zVHKabSd;fgKIzONnWo)}sCKReH(6`aK&nENobYWOCZZqxmqZX4@YB*mH
zj!C+}+XVxHz0w+mXzkE7-~|L{#4T4WfDi$7FJq%P_LDad7N8QF?5@+@*KZS~Ja^1|
zm;F~(=&Q^VoBa#>ypltefEhK`tD3Dq))w=R!BH8vHD(Zz{t2{k<6U&M9hrz%7a_H|
zy__pKISi4TW68%D=5IIw0%kD3X_!ZU)D@ae<?E-=X-if8P4g(=!7MP~-h|}9yoX!k
z?LV7-mDO*d#?+|4#o&Egih}=2Xuiq(#EDr_F!*P|3|~o3QzuTijcFK(!el;o6fC5T
zXD<g>?sXjVlG5$Ki<#GIQzHHXc8YUHf|0DL*=hZ~Bz0NTXQ(~uzy*!-%D~{hrwm#k
zLE@Xq+Sy8^;J1u4%D;O9lOyXO?!$+45H2OH3K&NE?206CnqicyZFn&Ekx)fhZWBR`
zhuF8%vW2k-n&GQtP^?4Ig~%tbLxkU0p^T14g8ZzBB%7qK4yQpHTaa#)&=r7nyi1!Q
zsM74%)GV6@>*!WWRQ_pxz5_It<1{3=3%yOqss!ZVOeW|R%1?Iu^L%a`K8=sho^=IX
zk<akl*i{*GGnp46b*1CiNhkAWMSHt-cE^66Bi}WvkKF+it&KqT$<9GT-|2Jn3mvY{
zc8f`dYcy|6`I)z7_sXjL4c~dV1DbBGeU+D-6=31S!M2-CFdy(LU1!~!(9y)Sd8Z}j
zn2i5LmixH01s*UkhQ?*jHd?+eIS{)dbspmSxu>sXF{<lq3bK9(&{`7WI?9j!pARSa
z0VZtxWxVIN&!4XbehuTY+$OzV2XZMR$d!7YC+6!NieriaayU<>&2AK_bK?NX$}wpL
zW<FfjRsU>pBzwh7;A_a|;+DqS_$lK#`07&(a5BrFiqnO?W;q5J!D#8|ZAVnY4#06_
zPMvBIbXWCGPVApc8GAZy`LIVb<+J5TwsK*(M{12EZ1teNMxt$o&7mONPiq(yrd%jf
zJU7k_n){UxC>~{X#LwVJKIv834eo;!d<~w<<>hL{;yV+lmS83Lx~SoGVm&|h;IMNc
zkn_dL4Rn+Y8LI}Kamq{RhEjWMerZ7-%93XL%V30&sqtX#(cGmNYQXlWf0+%6h%aeF
z88>s{E%N-5CS6Ky4>vo}8<6cSjY+2Oy<S+a4c<k0zA7>`6O4Xf!ii-Ud&25p$;0nP
zR+#c~jn!aJ@Bie*Kw<E^Q$fnhxDT;}h@fwIfrn<2evQ=a!Jo8NgAP1+7Q{;hZ0Gwf
zZ%=<L7Kl4!N+F-#NKpFZ`Vj8B2Ib_$<=RuK^I0+)+lT6h@4Mg1>YlcTEH986T7OLe
zC{d>Xr2~Dy?yBsHfB)4wGmQQ2kmSdIAMg7*6zm)Rp&&5wOBL0){SJ)kxkJ~lf}kAh
z4itOK$7+ny2UQUinfQl_KFrBAE`y#Xl?Lg9=OA$RUA7VoPa#L;TknSnmB_-pwq{m@
zSg$Aqi{E2q*liAZ3`>nmoyoPoDP8o?72RL`BW?DdH>y&uwM9gylG=Qf$sCdn>7`rv
zA+{^|z&3ssRxbl?op<Tsx9eJATOpR<Luz00E1HIITixc<%CvsEC9w`5WI4C(2yAaC
zrEd86){t%|0gj4yE&z=thtt%jfqWgrp}zI=<WquV-ur>+$K`w&hp!`fCAI<{Z#ZU;
zy9b6E))Tsk1V|Dj>%QpkFGeI)zPGy8w&HU^afK!Cj;cL=;Hr$2_>Gp@td_$XcVd|X
z5@Q$sYq*X;U#a8dSKn0ru2A^D;tRV4K@Io5RaDxkJf5#_5J(EWrNj2Ol=cG~S7c6C
zU=+c7!!7q(hQ08tG`wIT8u#*>1I&conOYRu+@tqKLeVP9I{I$Ax^D>ztS)-Tf77bX
zP-wq4<Fh7zJQh;IyqDkpMc#Z2cwPiWfw6SHD}A$)s8nub-+-rK-s$Q7!VA0aDf<6p
z94D_~1KIuj8l<|9SPnk@#)t50{LhTHKfgv7Ta;D2RJ$?uv8?)Y%|vV@kwvL~SxJ4O
z4RD1`CZ@eApS1-qz8#0^J?JN@1WEMYtme)l)MiwC8lVcluCa`1K=G4y2UQxC_9G9H
zr@VpE9Fm+?bR!97ju#A;+`$^^LI(Tyt>kt)0g7Lj>e8FKZI#lSYy%X3vsD;4SnO=u
zSE88ztr)=q8tI7d-^l5=`2>6`9s4|$!Y89%2rU*d3Aw<Dl5B^#?l>jacM}9T&ga4b
zJDZ1rp<rBZ(0(7)VUE--`ZoHB$_g3=s%iIAEQC<}Dbhk6S%3Oi?<l4`X*@G$_pZ$<
z{X_k859z$aHOoHUpvCl){@MeN^Tju9MY?OQs-p#+PEDtE;^66<&Se1~WFO*a0a1(H
z-@fUSTHgZEReRbu{CgQUl}x)_rY{E~pX*wpjxXr5icUVEhNXq=b9~>5`{4TggTw1~
z`AQSCn+~2a*DEJ{UT}vMaw*CMH_GcZB0-5@w&M!>?t2dy8HjWD0As_Ct)HxiN0T{g
z5H(e5&yL*J`=)SxF1E<FbGNj9GlVboR!F=v^ySAxUtMO+6)iiAt@ypjFoy8QONw2|
z6cm4{lo2lAAuQJvvAaCpq_qGF2U3K*uk`1aw6|hx1vTav2(EidS_|1MS773Yve0y5
z^hIuhf3uaf%3qfn>uZ$pox<%{^M{@+(FoALpBU%9cl#Cg7wco;k$-{Asr^EzMREV-
zs;5=RcH5gIVIO9<$ZIcoOb%Nj2Ukb!@<K25(Xfmj?mT67f*DD3i*6WiMAf`qz*x@^
z_zDEUj}O(y$Zn5-6$q)we1rxv8Yt*VgN$NZYkODArc69C)l(yb<PpW~5nZ9U9em7N
zjNTT{5f@d`$@^5$=OBb%to$9Bncr;Z*i6>Dr7tSP5V|uyh_BcRxDvDykim=r(^@%+
z2}UZ^=Gql#@d=Lk)zNV7MSzN9C(t2U5lU3peOH<=i`|>T-OJ+A>+crd`0a*2N7!s7
z#Rfc3`ipc=Tg5=N{hXD%?Z`D!^yDU3jT3kV(}$Iix4Zl<Z@Mzuh39MR4h6PC`VGfM
zl-X}q<_6_RmeEkN1-q4av(lcyd`GW78)f*Y)W}-qZ#%b0%Sdu0zWP!lm4U4x6<t`X
zAZDOD`ebiVsoyS9`3`TQMLDnqe=KH%Qf)F44+q2h>UKNkKeGoSQ;&e0X`nKa>K|m0
zP7y{8XMdLaa5+P@U3A}z^fb{>&>%>>G6KC`Q7S@^VQA}1NP>X|yreI_{YYp5=+OV9
z^|dB=iishu`Uu-c*M0HuD;BNJ;ZsGy1gFb*4RBtJde*77MtS1OPeGq><~L47mSA}h
zah!{i&MRaOY<UL(1-l)S_w})&YqXXg>9iqFfbhkvsm2c$1moft4$X?%k%S8DLKzHh
z6tcm!+8XEb@NODtlGza{w6UsGmnTA~y<H=c?|-PNc`cPRrRr8NrM$nd9ND3j$uARX
zw~=$bZ_5{+;9r<n2Tj@SNxXVO%!{rb$pg85*)6Lxsf4_{OY5FOB-S$$@Zt(_n!R_>
zw}KYx!aDk!f)Jk&cu$9q^Ypceh9JY*4di&J+aj5Vqn{x@Z?U#A1P;e<THVQ@r?TYj
zK00Wtq&IpXt<W-M_q5@gsnr#=@zaw^1^Os^XtS&TF|UX(N~oxTAkrU36DS-H&oij`
z;nLadk?#DNQyfp@)M@&nB)OknJH!ViJarCx#AK+*(L)FoEyU5|V!)zttHfO|mL65W
zBXN?AbGFCS^DFjQdN}?Y^o&RCW{7GQ#L$d>?Ok_WLzN;Q%Foe~RXimy<rUD6<>0L5
zVs$*bJ(k)-zs)0{vh+9Gp-gnxs}qY2GYhC154H$={;hMf*L;D9hbFN-P^3!dT8rr|
z?>02*RS%GF(qm(C#Zjx)uF%K;k5!AVrY~qoj;eJ^TP*0O9HL4mum^pMTdNg%X~sGm
zA%FSOk7D4Z-xxZ%3gKmy@(F2Q3hs-3j6_63^it6Ob!&9T^0SerXuEA;ZFKC1o-$7&
z@>AYvym91T%lN*9vIUaKSfjtT>j?FQ2JH3mZTqrv%Jh9AjGRp>3kg7eU1(@ej5Uk@
z@~F$YhSeuoi6IAIzJOCT-{B9C+?z{xUCM(C)9Sh&(eP!{J#tIg6(I!oZ+--?G(AEp
zYXY(|?yLab20_%piKLAc<k*&wnoGrzM=e#GS5F8)MR;lvp?fS5<*0Ei_ma5&jbE-K
z=((G9?A<rDxQ!JHrnUs5wHHOJpQv_oO9~VC6wkHm4{rPwXZ!U?+Y3(2PUB2Z9in1q
zln7RzH+mYPLgX}*9%X}_#YV7K?}t*%*fxTt{dE#?C#gL&*^91D8|OjSC4(sAg<3S4
zT4IkZ`+PE#NZ8p~Vr22?Du0R7IL_;kHnJeVe|3R^f8NtyO5Ia0@X4@UIN?=NC7_)Q
zSLv-^w8>GEzmnxx@)0uuaG4vU$igY=w?mm!7i9V9v-cg$B*fHcfL!HbG+iGrqwv0W
z*@<_l43ZVG?WL?D#(MDEfKrPh`f)DL&jwL|bCJ{Ta#0koQRw;tnbdmf!D|qLolX8U
zrBL$begVKy6h_Jq`b5+5JI<($J2%UUzXGZ{X6;zNUt-r#mpr^JZ<ye`VZz*2R0yh}
z2IVdYI@H&=LqgN-Za&)4?R)d#L8bT3oD;r|#1B^(%m(rV=|8>DTIgcIJ~+2e9gUW^
zVC<)AXW}P<XuT%CqYu$uKn@bFPvCy2fvCyT(6#?HPzUWNVZeRX<yO+bf3jc|&APPG
z#o=XU^%FLLE<6Kg$fx1Z*eb2RNuRADl2DPca^e1Z_RU0_HbQiDrOtb_oueyJAQ)~i
zbAni?j3>WuUjprw)fy9Tcp0wwT4f@4$-_?yQt2CXJs(sYZM+~M^L<u<?HE=@kD?}n
zc<1iHE$%aBn(im?<=M^cYQMgRb0y#{N7vWGN7k)3Z-w%RhsfW?Mxt!LJ#vpJOL@Py
zrbqqNNEA)K`1B}?5yz{qw^fg==#z8J)8dlE?zSafSviwZyPVHI6<a)14~mc)f#6Zy
z+`bZ;{v%F%dbf{XK9^d|doE(NEcqVT@EfPY<vAEC)D|h@mfmD^VmZ!c8)YabH@XOE
zzALg&QqbP@46^QL&F0@bas=yk=QWen&MS_Kc^Z)9w)|%^s54Rnpl(PDklmC1lQmRR
zbxab1aW!r<wPHLx9>BNvl~lVQ4qjQ6{raX)!I&7!kxsR}fPV3vn2bc9gNZwdixUKG
z_1Dm~a;oY2DG_~FKXqMf!jd=fqA+$Y!yAGIW$C^eCK-?N+Tq)J^L0$xPmKf#KC?6z
zWfDzKYqJx~lt1Ep>{h%8o1fEVV26(m6xo*rIa_WOiB4YqN~0m2B$PrmpWDt5LIUG*
zH;ws4D>!jtLpW!@WeF59dnjYiyX5Y4C+4~CCym6k*8{Ro{N)|`b|0+I5$5FGpHsyF
z<}|{cx>_@gIn+D3_5vhZ`x(NuxAyfV=|CgiGN>XUKJN%~u<hBrpK}Eb-`V?-$?Zz5
zB+m<{vI^5IDFjbCs=c@E4Tjj}+m4qn(4wCrkE;uyZC&C1zOw3Qp5}_(YN`C=0`=Z_
z(n;xtO8kF_MnV7jHnP4&rmrIJH19v`6NL4H4gncHv>2RPUObX-9ETkwmXUyVEhzZY
z596=-54OlD?=leEj{?_NF$Owc&N%a);Ih`WZY^6c1bO}Wvm6`912SHsUHZTz%h8UA
zx|qmH-=5_6a^yEjRs;lREa4I}J{IG*PCy5#6!1RQUSOi52b!|D>gqju*Ir>S^=eca
zFzykJS;>F4_OkS-!u@pQr();05GpmvNpAF*zI(kFVb^-`tk-wSb4bLp))M1olM+`e
z;UkC6zZBgQ#sOL1#Ma8BE2ww)lWYD{J|Yi*;x`C=-$CIoq*DyMSx0_?vhdhxc^SOU
z(QkF4&jQ6cdpH3cS$<8tyI`X}M3%8a;p-w-*!gcv3UB5jtM+8}k^L0>UrFx1@R+uO
z7w;<cm={~&(J2Q=ql>s_vR^cXb@i~1`O-#tF~WBS&Z-J1Z}9$;K%r@;zi>F+V7}&G
zIEA*{bOd%lQ_qexk~F!+Ad@&(M&^42J?AjWPbGuJ&{rRb?U+Bt=v+!Wz<S2P2$2)z
zP5Yq9Una_)UM%;=G5%kLjG_uV7Wt{?-q=|8Pb9;JftkgZH~fB+jvRcc6rxBMBA|X%
zoM&vUTnKId{hc|!<H-uo6~D{HK~se%n+iXD^G>`A#07pA^(%MwjIEH2ywT<hqao?t
zm67NALPV0(r?4uZg;4g<WYnwqygD)bkV}-6x=Yhw;c#5=MHf0g{hzW2w1tT6AMf1^
zp%VR5*E_93vHNgOBHe(q$JR1t{9|Q@OvE0Fe=99^c@I9n{0UeMwhGlue{5|EEW2N#
z134DHyI%2|De|Nd_mS(cd^sv!k`QH(!4{xO#aC@6EI4V26i2Gq@cHF<Dk7C6)2PH2
z6Q}gF?UB<49pa&$+iSgOaq<GY5LbRkJYhAJfK-990~;9ouFM`s%W!hwMoAT$`rXiN
z^hlvH`?Tvo1Tj?C?<wep*xwi;r>fHhu8uHJ7Dm9OVx<LrX<YjJ=PZDZeuYbC3@r7R
zWel!(MV_kpH13z;wXz;|GNO>b1eXH-3nf(UPDUPZ{=4IZk-hxk5nmi(U&vbh@M-jv
zI;T)%u^w^{mIDTXSpvD7*ZMxbkz9#I8*y3+v}iQ*U(@Qh*n`V?f{qg$a|4^<W;0uY
zT&8Xed*|?=(+>Iu2AEKZMXCkfW)&28)bCyOWlZ{3*<l&7Q5lK21R8X(;&H+hAW;{t
zC`&mV+GO6nJ5a6*$aWcMVE|Un`V!dpy|iL1uzyu_m#TGeC;0#EQ$Fa&>(5B3{jVhq
z{MIkZ^mx7UNyGkjUv!%y<Z9jHPH|vJYvd9d%^w?A2t2(`y8Y7U^Bul+#oi14kjcP)
z#>)Ku&?0K4F-tlWv*K$YOlPq(7RfsJ)96<<&QG@+Hm~Zh9Vgb|Z{y1G3apv~bv`O|
zN}FnswM&1?{$6jhtJFKiGy`m|TxSh;0^M#SgzpT88e*VPl(x!8_n-I_Ib_=AR9#f~
z1>*_`eRt8%qecmHtvez(&XyI&;#4>;{h0>Gm98HZ)Uyk^^0G@fUMUPF&`ua6+EbUs
zGSG;AAt%ODoMTWle!f?6eIK!-JYVOQcdrcHyKsJ2NMnZ2eWE;W&%`{9hvTdk*-(}8
z=KQdZFS*5w+a;qVQf|*$kvEGZ*trOUF=G0s3-{}ceio7b#rz*u5ER|Ur&ovL$BNC=
zZ5d72kjgbt7@2gwx)aFsEPmYc3W`hOlG5Y?RAMjLwX<0&l@_ldG^tn#K}f?a4ZZoI
zUvXQkqG#+lO;g+EZ~#+Kwp<Q4FlS;OPo~@jf(5_A{_@LcaX@qA0(3wRDWQrwI4^&=
zwH$gg{IEt~rpVRabetLFLltUxJEUP|P@1QdhO#|us&g(XshOAf0?+ABMf(RKpc^QZ
z3g*pCB-Q(37WhAk9kmc_OjdC2QhETxW;*HJA;1<Y&4^^S_LR=q{v($GF0hp4Z+H=E
zzHk8x2o9S|g!yUAQ@7><PDgTb?c|dX@2U&>hxIfR&mbc*M$KOb(&uH-V2oxnRjw*y
zyR%4_4x}pzM1ii`V~J!!75twfQ06H8zCvCo!sEx%1(}<`#v_eOF-n$?T!E86=kJ*X
zgSVcs*kaRfzw)zhxPm+HZhl=!J=wuX{*t(2x9)J<K?r`=TYbQ)hc-Q6n)go9@V1ou
zr*s51QUyN+*TaPSH4Kx$c4g%dzq^tcN2w~kp>+y&61|Xb@@uI`c>gSZ(Ly$~jo4A*
zwT9B5+HHXC*_IVQrl&+ZOyvYrngcYt=>b;`63#l&9jgB+uYLNxcag;7vWMvLRFBI-
zbhN&eNHnE>qQ=fyJm9`k;9-AQ5aj#qg1h+s5Iw8Sgnaf`C}7lCmeKTOT+5OTG?gu7
zHh;GU8bmdAkqRM;8r(XZ;VZfyIYD;6*PS_fdJx)Q4Op5MZhSnxb>4GfTCwL>7Seq^
zbCWeRq6IAn&fn>fKXSA<=0RyBoSGIs@mwrX>4fw0mwzbhlVBzGfZ<pKX70;k<jVx1
z9(zH=bQ11=bRGKTd!l>;C3~WhrvEF?LA{%l0QNKzaB=8m8kXEAUTw6U$^q3g(21;>
zH}=0r%(Euz(Gobu0u~NqU%-^}oCg|1Bu{(2qil0NjO+2Es1eOcDoH?@tlUS=E&lF{
zrg){MOFf3?kfoS<qw7pYGd*6)#f*%dc4$v=Q?2j&QWg4{S`R^j_<M7z*@@Nz{9%7l
zA3qBv-=(vmm5Es^V-{BQ(;PqGg<W;;>WKxh-llOoAfs}jQK$3Pqth_2PAZ?;&4wRc
zr4i3fLsSb(5|FKjVJW}dMO1R566b$93dV#=v+Fje96w{x#}XY~&nPYtCiCmYo8n#H
z#ug!43AsqqG+BQdy-{{|%#rRDAry3lgY_3lh!cV=g+7TiIs>V%vsz^^?`n2ug>Oh9
z1JE3wxV5Bo?NyRQz=l;8=Flo{H|y<MH&OWcTqF)0>wO#CGCSXQ&wv-Z2+C_yaNJ7-
zF-<K4!=(2zxZF(fozM=(ok+qQ({U8Ed$Kdt+f*?NA0Knv#(&gVrqpGbOCwRvB+mJw
z>lG_;a?SDrz}{_9*fA-GsoCWwyJ^<$ms2lfpvLa8^@;F0y~%$MPMfHXgbf5&JsIuq
z!Wor&9mO#4pRPp#A1GV~P56aoi}G^*Mx|I?6xADhzqky%RJp}`><0BS7m4=NRl7X)
z`r;$m^%K!KHqeRMkKvJz4tX_)(b;%ZKmeD^^n#Ag_j8>_G8Us^(wC0+f~tv}qT0WQ
zAvhHV!RecT8C$bCddf^!WcXmg{ueqK1CGS0MiHYQRWhF(xGNYIU&d#IRj*t=9>HoU
zQZ&nli<XX^s9iJ_u@NSk_cJ}=B-nphwkH`iVi>@%S6+rTO^I-MTK*7i*Qnz+6LsLL
zP<>7X%hab=Bc&rJ4QX8$cJ^e@JwEZ)5jrFohfT)vzp*@tK=JF5`fh?^>(H^3{k$%u
z)M#g;%9x4-r=uQ(lTbARm?w^?CsV`zph3r*L3;mH%B6o?CKiB60xINK)XKG(Kqg}D
zN9Up*$#wL1%&<zCWHfD4Z`bwHO_o=B3f=*+hPoj0_ruAs@Y(I7KIO;LEBod?T2E<b
zGcya@)xLB(&MB5B6Y!L@eV0`C);s#ntLHfUxio{gbxG^M2NH>D#Dju7c~t+8Tq@ye
z?MpyiGi5r-&a~iy9<ZdU5oX^QaRBpC^Aq2j=iRAv7pDBkY>h^yzi<q7-q$3so8vJU
zG{sbZSE(&w0q$)(X@hxcAUr_%7&~{jMn>NR`2htUa9kE^t!T$!Vn}*#KkkJWyZ{bZ
z8K2)EOMre#s4JO0+TmpEuQycl>arb*i?I)y^g47f*2vd?5sMR{UWkK%4h_&SdmG3t
zd%?s>mA1R1XyiUS@-6up*t4~nJ`iow<@H9a&Roz*3P8|58}&cB#)}Gb=o_1<^CP(^
zsS2{Xoqp+$?wt8?f6EN7vTZ}en<d<g;lS5xjjsK2&H}&6?N_UgKhIN92yR{Q8P*tB
z_7bur9UUgkacQ+v<tO2LD8;?*fnEtgi4jJ67;F;Pck5~6Hj<g=jUMF;S)jKugEqPC
zD|Bau9K`6T()+hmu<Uy|%>~m*-9M6GW@u>Nr@Og8$7#*9pjwz9xtocrGHY&^E06tm
zepo9e+8#8*blLCX2|0t=W7vE^J-{@S6rW@TS8y`v2xmZ(W2pE1Z8~$Mh4MI8Q#yhF
zAXGa2m2CfVUXi2x^&qRkq(>#YNrcR+q0wi_Jd}3l&uR(FiG@u^kFBn$YzGtV*=l~c
zwsCXE&0gf0EWEU>vsDEj$FPO@-*LG-zEnR0s+9tTpmPH&2O;`u7e#?BhrYH?5Pi8^
zI(L<jXQuFH^1AED*}4f&7VzG}hwmjyL|IO0o{GN_=jYo|&rPIloiZMbh%(G;9Y4lR
zG|F!o8ikR{Qw$qDrQ*ZR!%EvghDRjL#K96@L@3y>*8_iY)1^K_=O1Pcpd3E_!2%Yl
zjU&5U3!!?4>&3Nwgt=j|xYdc)cPlj4ZN@3WM^IL*6<AuG@;M#cu$nF`t+AO?7-*@m
zkn{RR`XUrZCP3T!d-rDH;5t{KW-#8j0rwvv`QpUXFJrd=(>wb4#e5Yp_qi@kSBtc2
zt6_D~yll0GZ_fqR-oap$sH3mnDh1sImTAp9JkRf(F?c`c@eS<%P!l&(=&lEu22T3@
z2Vv7FKMbBi@Csw2w@rEWj*cW15=2uo4p|4-M*6_O(d{yQXoUkauICd>cqC-I>WnRM
z^3Lau-l=%o4}@vZuQfSm8b&T^)R<M>NSZJm<GORnzyP8YZ>;3^!HJ46(Qpt(qi18K
z{eGm<jYdx|lla>`{O5kUZ?HF><4`C}-8nJ-UW^vmuW3z-5LbrNC)QOACDw9G)>k}=
z`p^)1ya5`L|BTGcWtEQ*-53w|J+xFn`f!)h6^I1GDq&n()dQ_|&A5@^Pwd55dJi<e
zmSTce9>0M!r&z{9lSb$tDo5(Hn3E%Os19kL6dyHcCJ7ymzjWy==;m=-HIQDi>sr1g
zl-+xb0S~y?YLbEC<VMV8T!x4rgZr?7zoo-X*`%Mg9Y@el(np~^KqQGLf?+0kXD4EF
z5Nm#T`Q4!bGZp1yv_uGfB5jBv_x{hH4UCB2aHD}W4V<gYpDUuxwJ;{n$K+6KK>9yX
z660@gIFB@My%#Xuj7I{!fX>ZGZc5wFxr$VKZ>M~=b^kYyr8tJ~$tHI-nH2fEDrEv+
z68r2$Gud)dcN*%EUnf<Jy`@fjbT=2jok(NMJ0|$uWCX>a{VmeHZ5rz&Wruk!yOgnI
zXwWS3r9_(uVP;Lib00?oo6+1?%{&=D#GvYLWl>2bW6qo;Yt_iSqx`-VxXfdH%?J`)
z%tcAdhUNEJbD8-W@N?77E=8`DU>#DDXG1MApWaz{6^EsyTFp**PamSiBs~82{(W|-
zl*dVc%l>LA?&+`5n=iT6_ReJkCm)`@Y4s=M1DR@bKc5ry8d}r8Tz)X7?+*)kWx1n4
zzZrInlUSsoEsOSnrqF$4Q=dT+{@`~5d7UsuGU@C;dQ4#c(ln*;(*Bz;D-6>f<qxT6
zRh`9?H3pz{z_V^$k`J76j#4A+?ovy23E2ed?bCsY8J`0gZ?slXh1Qgf9~*Ve)|0w}
zsp7#tAsZ;y;m|QCCv)LUZ40%eVuL*K_r_dv^2Y5#n2@C+i4&mp)u-T5#g!)BC8Ot9
zm-u)M+L5}O$Sel#hK^H~t=Rhiai-g86t8RYg+k1aJFI{4Et&Q`^t`9T<Qud;?uc<&
zZUYYY`1?5A(G1Bk#bm%%s(uR>^c0QoZMN#gLN#$LI^xxEfPMW0b;8Udaodfa6E3t6
z)pzfcptkL45`I5tY>9pVjp-xZO_wqBC2Mg`r=L!4w(n$VqPpkOC-Wy#y<=U;NKRPn
zzSBJpZ`X3yeg8oaZrWDX2No!}0x#YDPy2{NnzsKi1bQObudR(}Mi9oQ*k|l&9FUBz
zC*ylfV4oJj>i8|OWM|DX#!|q#qNXw_AI_0iH`h*d`?EFu=-pcPEJ~~0#=QWPW5--_
zDS=$IDzk9T1Tj4e{&^{a`%7T--KrSb<N&0njeOEAq${7~My*TdW*9SkETjb_vOH9}
zLI)7?#vC2(152VzG+e&@Yc*uQn~%a>d8<fjdw-O1VfMEj2hX;G*>#}AhzV*VIr40C
zpyWb(gnds%;xbp0kEZDwRcc^;x>rL>p^B4UOM@kU{VwGT@UGGvcKwDcS!}xMel=M-
zH8nK&Rw*QBD~`t~Tvw0X-(Y9BmS4p~$8qM1USB(mcAmFWl%}bWD_1{VsU7=jID$Y<
z%d;S>-3Oo525i2Q3&E2bTm9sC!iF1Z&HMidK6a`v3z~|^4Z{-(%>Tr<WSY-L38;G`
zyR&KKOhM4yLd^%uWnLEB$#ou`Xw0pY68jf<p+T|3^DV`j`xAsxCx*Yapw@uB5zX9!
zX+7^#kuY)F@<$2Yi8Y|eRD#u!o!q#|J_B4~DSMmJ+@<NtpWq9me-8LCH8DW)5!vTn
zS+96~Pik>9xMahGSnAvB+faBY1Gf{pMe$l@r~O4I;ap69;)9vq91Oi4xQo(RCjQJm
zWY4{Y@A(&oLagy?+MBPXn!iMSa0JgUVwe8Yj>FP?c`k8oF1}6t`LDM{@T!!OxlR@2
z&OBk8I}w<=-Ip0j_Ro-8zTemulmZa)uBp?xU((7SsnYOzAaSl#o)VU|bQ|G5buIP@
zwKwesKB$Mnv}!20!~Y*uXBn2&(luZOB&9*RyBjI#PATaUC8VUgM5Mc<JEXfC0R?I4
z?(XhyIPXEv_lwKN2lk$sHLGT=xl^7BaL~XyN+egbaE~IBSa^tRYmFd`9&h5uO%Kd>
z?98ft3`AR4vn~^<0Vye8v}fI2>H?`z(ED{5f>Nx?dEeXTN6)=8lQDDlYlEAb&rt;t
z$X-&if1%)}ii-NSA40DrdMRn|meH8B!AL*j2BrLF@5_87k-*hN{arb%!9d5NxXY=Q
zq|f|`HXbom*6zVEL{dC2)R(#^NO}j|VYXw-e5U81B~r9$_;T?4m{zkwy{W>|W0P>1
z<ty3|C#{u-x6L}79y=Z<LeQ6LP!v-Cc^=DZrlS$juM~7vO2eD`LBBuFR8CgX|4F>o
zuwF^DH?$<xnm=@sc(8cKrrKOm=CY&oH}fbd3dKLu5{)*L^`82Vwh`QcEiGogSD-&c
zyU5R{a#1-{k#{9r4J@I{|K7x;*CLb*Hw}Vtcs;fAPXwmB3KMpOdVK=Hh_?k^QJZ`9
z*j^qSO^dPt&K$PAS+-}XBvc-L@z+CnlAN+1l%FT|&xuWMULOpottxhreRyTH=O^$|
z@Ad3M_Gnoqt_woBT1PmD#<jHA>b<O~8bo6MjG}$zHD#{jY18Fr8-?$ulbaXct73kS
z!=9MFfon=o`7LkY)Q@_-Q<++3L`1RcE<7*h&Y|6e?tkTTFAB$`c5|hQelhRrKjdlA
zRdEo9JzQ-h+DRSy-@MygV;IQ0{Tgco<!Za8EQ_lJ<5&3fNfR}F3dbtLbD*|o5s8A3
zklzaOlW;pTXf67MJ%I~+RS{6Y{&n}*{YB6DsrztD^Pg&w2Ni>VweORELi2a8eSl!d
z-4n8;;=P%u_3}vSppLwqTQ>eWmQLP3i`11m?Ox`W@q@seUe)drC6KX(HGn)-WN4fB
zCk=mz434`1fzgh{dn;qm=A6?bA7yxxtD0+lXbEeh_|7d(nlLX$AgKB7nup7N8TC_p
zds(Hi^ToT~G|S_p73XFh7kp!@XWIl8uku%*sm3^}K$|l$r0K1KxsI$({>ExB2bHD4
z9$^bM!PT9&0hi4w$rp}aVf~nPKY2#fvqW<>U`i87&&)2!gY5z~zpiS-ii)LyHq=F_
zfcFv}IAs+^KY~YVubH7hDNoy^nIp6FrR5HfGfw>QbebLi%jes#r>X0cuZ<)e&f#R3
zFB8{(M9pf<T3&56O(=w()cH*3A4jQKF8WPwfAr2+cva4m;o4!7r%an#f7X0YKkW?B
z?PleL-jpyEIGY=d?xB<l80+s!7*mfLLcY<R>+AOd3M}naGAmyRHy+XgCihk*t3#+G
zP7SQs@k;AhCf4a_Xhc?j%rx?Wh#ya$OasJ(lePY?cSTQnNeT7eGxdk4fdwr*tCoc{
z68>>F@asSTG7E8A@=C7NUJ@hX+Ws*lZ|!7?Qr2c;lZSNi8m3saW0LceZpp9|viWU{
zU-|EhU+gsCZu)%mAsrkDx1AZB@b(1tW~wU^kuU9Or0+mh!>%Mri=tyS@4U}=ADeTC
z$SfjbZSFMYt+Rd{WmV+(&nWKX=FU-=^erw9sRiw(dr{N<`}$1?C-2PtjLe>)64kTJ
z`s^7t)7&jkUm?yQi;bDU4d;PdENvT(5pHxg7Q1zD)Pzz*J=vT~Tt^Ulh{C4eHJK%O
zXmD9OF>TO)<+Cz4CMEldVlrWJ8~Nc#P+d4@*Wa1T=CD4{twgO0=Arm(1`&o#j(BGh
z;h|sKV8AZby0|rB+qih{jeW_Cix=+gU2C#m*ieP_vzVJ|#M7q3!AnVZYwx=K`Zq?^
zLeG7dT{*k@=dTtDC^1P#pL-ldzo1t7M(w<t(Nx5dGMHqk{M}9VO-W_g)^+M)a%LgV
zVJg8v3*Kkc<ZawhnuQdxGJVrzpUR7}chibHJoY#-2<x2r=PdFU71Z<NVM4lle|OgR
zIB-DJ9hX%3=dc0}ML+)Y3U%E`mWe~@3x1jYkChkIw&yA~lR2F>#GkWFv<V7XYIXUC
zlg7o$2pf-{Y%9!m#bMeHt(Cc-p#RiJiA)kpgjP`<ez}uZz~ex-_By&DN4ezG-tWaz
zc=r~0-jI{#Yfp{^DW0>ks)zc4IHO}zp--9Hy{ir?w}r2q5twwF+QcK9t*PRZ-%AvO
zu1e3F>4HyF5g$YF@19j`u9ciF#{Ps(gxvd(tJr)j4}~E8yWwa8Q6+sA{zau%>BRzv
zssKHFd?+=q%qQS2*{&Hg7^OU^V8IfyY<6gm<}6UbM&Kj@6^;%is-#SuL``1C(b{Ln
zJ-(=DXZYMYtZz54s$Nq&a@u#?8kfA?y#tA`H0>m4=DU51lVk<&U3AwD%8=|Ue3gfm
z-Qwx61d7IxtG84xb4=pcF*qssHx(@&pa$02E$moPNiX)NM%g+>UxUB2aIedm`4<F?
z!`;+u576cSXD06GR~2LVRAIaK6Z@evR_nrM!E4*rcdEa5E??osj_12oI^NL(1(~@R
zhxgaaA3ih`I0;5k{@G}l5I;GDTkMKu7pQL$bFJw7?r?1^uiwd)IDS>I(1Y+Kst>(%
zE%}tWgM9Ujmda5(l~-cBBB}4ksPT!MvznKl7o(D{o+QlXu;xNobths_aA@*~22OJK
z*>`~hIOscfZO~TzsdK}~S4BD1eDhF1vvNTP7Qbx3Mnuf7sC%QzxGZ0Rf{#igB(5^e
z(B#_>MN<V<K9+cfU8D_}shb%rrUdoyXo(B0cqI9$#2Go#lT6QcA~NBY!Z18-)+qx5
z?#;>cmJj=;AAi?UCbC7b&tL&{Oa8ui1__Gyq&Iv=UOVKysriWJTTp;^l~btbxjN>4
zTjvlyM~RqoYV|Ey+{LjvY(fWp^DQ4dt}?{*Lm}99ZD-G6zm3S0e87%8k4CYRT|okr
zAHP^Gi)Xq-<`MPcZoUHp_F3mJKWO*!0iniFou%oF3rg@$I)~+g0GTombL5VEq8oM6
z69F!py7WK(;6W&e_wGZVWY_)sk@t!$ah>=^1*Bh8)14gZrU8?#0Rn{^fy&Dgm*_Al
z#+;LbkgeK`ICR<whejOkEovg(?J={q$u<Q&T4U~ck&29BoHiS-oUWf{iI@Whnm6+6
zNfT$b(xScYOdFbACTdUiQJ81I#R>Q^Ce2jEFyCB_%B$M@{J?!q$ndUDMc{}sgD9!y
ze;){^<DX5)ID=-NyUrQbjKlqgU<bsn>{6V$4;3SsH`u6Gja-vi!X!3RL_!nCSQBm+
z5MSKVWuWkGR&WB{xWy3@DD?9&&aw!gx`H%@`l$VrCKJjE5zU7DBC^7aI-9f&;y8~!
z;Y*a+X5IL)5{~Cs>lv^xXn$n%($>INiN2$sdZlk><YtXCzIAIHjqW*MD)2i4_6Cc|
zzQwtNI<B=iN=@I`!gu_8G#2wr#YR@6`do0-QMs*vS<Z=dozADR7}#5=PMYLL_Z1vq
zdjXeL;K{o>-^a<$k`@#jrc`G57&{v;9bm;Bn3TDOhxxOZaAde+{Lnw?N(<)AC{@&~
z!(Gax#8;kG;}1fD4NgJqry5~;75L<{=m!_v9%jn2bV@<61MOr~G4GOZ%gO8!;I4Y0
zVeqb1z5TpDhqh5v_#q3YW}<Qr9xIdIxk*o%C0xX!|1Pf_9E83+-L(pbOeMHBpUY|*
z2(Z^1`2gJueR#cC;JxvhF|OG|1%|JMq*p#oXwm#4G!m|&>^mwR6`A~APLN=nimKwP
z^)S@*0YihyWP#ZIc7JEQHpOGPpKSwOE_-1=Xj;sMU($X^?X3<c595}RYS5TachqH*
z_b7-juoZu(OU$ZEH52=Q>b>eNyDPt77`?$WR;ljU9P-Z9MwxW}GwYmvt1ZD?8RPJD
zj_NEY4i99?Wb)kQi%XmOG=&lAD-Yp-0c+{cy;9yy+R4_Tkx0kFA6)Qe-Rm;gRyp($
z;Nh<GklX_X)xw1C{*(Fh2g?vO*b$ljSQ>vZ2&F<^^uqEk-Rdr7e?5u;k}t`e#hr@6
zh&<U*f>#HsXl?KenS9;hioFN5FLVb&C6!DXF@$4_`zbvW{BOqaPDxF_ZH^o<A1kRn
zqn&!&i^k7%oXJf-sc(|S_qTMv%;p(t0Zsi{rbdQ%Sd8|EiCLkPMRBOX<XtZ^iuC1^
z9%r5*_YnJiy0szhgS!bAxH*mCOlGm+fk|62ipE<|uRbbgi@VR$gV(xjp*N<pd3U@A
zY1GLX7YS>XI|Et46Q|UGqOv$GU`nsFG(3y)T+;<#%f#Rvve@QBc5+HlfWA9wMh?r{
z@<u~y&o;F8hF|<Me^?lW!-P<OIh$(Gnm4#Ig~Qx9a><|p9D>GgI{hzMMfB`ddH(0R
zt)*#!I_ml)g4V(qD~tEuv$@68-Br9|4HX;(>aY*JaCy%X2lF)*&AFTiJB&)iT3)Em
zouM3d4fFX-sf1B8bK6M9l^&^p8n4%c-eg(3R=!qRY8@w|u4T8iOm9MWGWLEUpbPYC
zgzX{rrVF>#t|HJR@$z0|<1l^rxZ;_()%iNJ_N98k{eY_Lkmr`oiKP-8i9)L3R?jPG
zXc`3wdTJAJsb~YR{(cFSf3uJLl;8y1Q})Wn>M>dlNO~eMd~_dKz+oWgY}55J)3Fbn
zW*43u@q#o~U}HrVWN-IP^(Co?x;1Y64!zJF-+OwEj&jx(rK4#nEk-2>b*zs|>U~qO
z4jteS3%lj8(Z!mZ_pYTlR-z!oa~<}|GRAg$VE%OB>%GY6vw?x%Z=%HS@^YThNMB;Q
zF!6ngv%;cvYE16gCOt4yD%Vdr|K5l7TOPXLHw{aMln{sB9=Lg``-At===S}ObSoE_
zG*1hwC42Zi$T#^x+0<5TrC(4Ro|4u`@>6WNM#(x>lYdtpK9urLS(=lu4ICz7@g&p4
zmYn@f6R)uOPd>~3$I#lp`K;)@216IErZG27o27FQGNIqih330cwbA*pnI5?=9~Hsm
zTQnq4SYK@s1nV@4HPx2tni9E<6qeUpUiiJYbm7d`SC7{3p4zWo><%Zma!*w2+Bt<`
ze!ps&RaKcLTBUm0e+|kiN)Hd5e=`+KtD=wSKfmTI>UAccs64-Ve`F{*jjJw+0FUXa
zpfb)S><PLW+u&>ka(<ql8V<*~{>LRCIlvM7af6lmWmx=g7V_>N2*>WRM_Wms0HQ?U
zBKqpP@+2ruP)0!diNd^4(q~Hh^Cie6M~snTvWiA9pMjbZZPx=R4c4sDb7QBH3qmY{
zY9xZ6PX_`l6y{)sL-eW_C&KKP?7{|(nHJB9qhFAk_kFnb8$9iOO7cS{;%4Xf_zs-d
z{rI&#jKlL#p7yWXHd!9^mETU4_%*NPKp92Q=JiKIVMBG?lL8%up8fw+MP>1YoO7PF
zwhB6}*>sO~&bQRKe$O=orLAp|JUbUgt|<8MGC{wMPWQRnld|8czwJ(%^~8X0;q{i5
z7T+GwZmNx4wGu$n8>92lm7~qi$yqPk+#lT@=aQ59&}36f9Flz)8qe55|4^v#n$@a*
z!Va^9$JnIJ=N>&e+k+l4>GI#8lvsb%B=g8=mK-VMPi55-n%^5G)jtyp%u-luH=^H#
z({U{C-$f`RZjA`i{QRCipQ?bi7zs7vjE_Yg4KJoFlb^?sn@Db$s@~<WuFFOc$AW%w
zGb0L7!)1c#Js-YeyAIV;H09JrK}F^U_vleBZ!1*k6|)a5BcY!iR%hgb$db*LSsP0i
zPb|#CNSVA1nzCHzQCo%txKA50ENlasw_uvKw%*=w4~RxwD&kPpal~!BDHyQfPsj>;
z6Z?7}T=F3z9=U9)sv4Kral)5#oh2PVEJ-L);CZrS55jxN@p4nK{ngoS2HPx;aRM$G
zX9PxZ+Iy5|8R_>V)V;6%o(X*>C!VZFi_AHp_%}2sjz@j1IbI&mG{mp!P$@vFbA6g+
zZzU2@P4P3JW&7<D`B}G<xo*BZdH(DoqEx)yIy<7~bK&SpRH_*{%dUC6NhrKiWz`nC
z=yw|Qdr+WsObE^`)nL~nIXZUJdB!!^MDYh19%sn)R;J{_Ls;DWa-VK96Ztlg7_^h5
zKIQ&fb`=Q?TGteC2c*OwgPVm+l#6JxlX*-onH_CU?40_=tP1?iltIoS5vfc@wFpiR
z?Omjw&n}@q=Nrw%k9_;y@?~9dmwq8{>|!d<$aiU)U8-_KEKO!cRt1@^OngIvV{Rf=
zk0Kx$6dS9Ywx>*jg9|zH#IE|e9U+V0t%MXjo#17-Uc|Clq~~Au!8>y?t)kc75@Uli
zyKl$fKB3kDrgRJQb8)!Y_l4I{u(#@&k%5mc{AWQ@l7}|oBjR`1kDsBDf!;ewFOQp)
zUpMY*EQagv-wl|A{M5scK!P!gQE?o|ev(DnjYAc1TimQrnC(<NhunGIt3XrahGfb3
z^%bN?26@$9D33<?T9l*D=^kQ{l^oN@C0)+g+W4mHz`(fykt4oY+EwwGM#c%pOK<6|
zZHe3aq3YWFUMY{NW1~*nJl0EeSBRaxXxQ6{V`)Z?bxSUP;93*=MZL;FHZdhPV|&;2
z0g4}~qw<cSPu}tWW!H4``#bT@nXNiV3g;XXR$O-m)66~W@|EF}WM48G?adY5-_QQy
zSoq--yFjzN`~}=!rpPMfs&Ks+*QsT_dIrfcS48ex;nxH1{&M^1eqH-ZdF#>j0JT{Q
zpf4+3B(*0h*7@;Mf0(D%(p``VLg=Z5n`NFRa;t9ZP^Aq&<L4{#si4RDgV@M}M&9-i
z@*#Ly47T4a_KsV<1`R*jjn~*98`pl0c(e@9z0i{2sm`$gRYhN=JlTs=B)$5(g=>fb
zZ@!@8*X_sBs_$8xo#DMdkt|I4t{LZHcz2wGUwpH>r)PhqM(`t*3l)VW{;M^`_<Zcm
zp2kPJu9}|s`=e1a`gwaapWGs)&^w8Ag}zUVQ1!c>x=)Jl*LLBzz+vy;TCVc~RJmb2
zhj1^g7_;TiH>)Nurvs!6?9PqO82kp>`qASK?MI}sXL5Kzt5FE0i19;F;zgqkDhCXL
zG(<nAp}yJ8R6)Q`g2;hy^qN$u#x!h?7X9EG3aE}$Se5GKdrYx?_i4Qjf8*cH9O;G*
z${O(Nu24E~*Cf{A2|w@yXDDN&5VGf|UI&#;q-2LwYW*RsP*zOKzezF1SDu%+^Ng=T
zD{4NS%C{1#+H`c$mRAmdqNnh-DF}3}?MRb1<lXlP=rzoaVX4w-W=>4p9*z)jOp-@?
zhSKJ*+~8Xn20HYX&R2Ocuh?Hiw8Xi2CvWR?F+i9t+LAXKfqYdytaE5Wi55>P?VQI9
zmKlxn&Vjsr)d$693CDdiN1~-cXU|cE@vHZ$L=UB5pdEULZ!<`f;z)eb%BJyv5aV+s
z8D)IbpLi|Y|KZvx)K|}O0v}_D@8ZOKeE5qP9JFVapbqLw0g?^Z#0GD9O|?pQ91XX*
zERy@!ys>jKUzT;r<R}c!?DcQ*dE3Q)Lv^LkuVhw}at-6`Ob5%{7*{e0Aj~~fYNA~!
zi@v!y+E%9f-=k!*a!X-DG)^3CjTU^~q>n#C!!_d9e8nvnmu>BEHzpzX&80=kD3|l!
ztPO>L9ka~_R+{8JxMq(9_y$~A6<_X~{3h(TST)r~7<%G&f2A?kyTS+zl?`l0D$bLV
z&HC)3nXY?&K)M}d*711q>N)q{dPSf~sE8wZ18$!FJb2XHZn6-ZoSaUO0WUqqjl+ZN
z_B4i#R?G18jrZ};+VtEG61F($`HZ@)ERA{XzHK~pn5!t`7myf_>)nM_8167`y2dJ*
z(_fC~l}sjm23j^kw_40Qz3GruZq3e9i8Dhip9*0P64l6ZI$f;zI?(RJNiGOn^{Z=b
z<7dH5Sd?$}w6{ZiQ<BmiT12?Ja9*U$Rt667&26P*GCU|$=>QiTDuS$s9bSv*t>uY{
z`W+}MuPf&n-dB{qk+Bg3UCc-Po)-KWSBS26o41r7S@mUP4Da@zYw-MM)UB@M#P^XN
zVIQ;29vxSyWG`^dHiWSK(5jR>Ba!V*HHN`_P0ma4l2-+ldz0}wyKW0vb!8m)tyI_-
z=hmYVBX7Bx>orJ5e^Y6>+?43@pX4olOS(HJ6c8CQ*dDzFCB&cCD|?qHxObm3U~=E>
zN_~-_UxYhw{PNGWMgV&LiNaAhh2KvHAH6YQpJg)!Ix5BEZ@wWg@tg=pZYg@+BBruB
z&)C7PSPQp?5X$f<aQ5QEl2RejbVr`0w!+uT);FWfZ|}s1ht8id5q@s@%|LG>Jg@N6
z>8Z~#VwOV5bl(Y0pU&EUj-qrxxqw3bfV}u)DdR#DI<-n3io6bDE1h&t!j^^Ig=>o`
zYg5}xj>d;-=(DRizmWH|7~p8Xt7RWTV!peJt88~YT68>Nig7~|dh}RqWAC_Zv5rk@
z@1aLHU|T4ETekX5?IrWC4Nf5greJcnV8$M)M!3eDSuCoOXK4VkaJaXuY@$xxf*h*b
z7~13eO2BthJYo6&X9BDUwm_`Y@t6pxgne1$DN}1a^z;4U{sFisjBedyOm=_GYi~Y-
zHgKiC_??^ih<bXhBhqiY6DrU?AILtoqWy7aSZ-xtyXeK&$YZ|umXG%Gv)~aako=$9
z&OPRjl21<U{Tzq2l`8dRwO(yp+O5Zj<W~ruA)qUn_Aq0A>pBgY`@BRsL%VuLo`v^>
z=rg|rna$lgiE}c6Pd*O!ejT!W@%2dlh+AF_-b$-+H-F+w(|tYCCN~GpihlzB1Dk-s
z6*a7t0bcY8cbX?Db+#RCZHvj*()|SNCNmfXXQ>MP%E1nn|3=@`>&u~Ny2aEh5ptvF
z#N7ByYR0Ag?tvJoBu#Rf?gIU1_49#gD9fGP?!M}Te6ELRdiT}QNfS?~W_QOk;7w1q
zU)<YxzD2qnh3U`q$m}X3ry6}6DG1iXxgIo~H6IMlpKOV!Oqy#6|JUksqYMH)C2V!v
zpy-}-ev?gWKIc0$Jo))9rtN`r9Sz4)eB|kwCXb7^Q~Zrt)Y{6^o|jLBzv*x*P?__%
zVu8Z#bWc;B3s3W=te)amPM`MUQdQur62s|{wT{t>Yfm@6*<4+ZoG#KSkWtg)$229&
zSy8O-OM11?%{Fh{hBlGkuUWKQKop#%baiy^uO-RnXTdAm-5?WeZ|Fg4zEM>+Rc^wb
zdczbtd^r<s<Z=J4*ZSy)VTa%TnX=lRw~RQiJBItXbr6fgP%Iajq+vWRRxwQ4eDo?T
z#bl;EO#fSU(@V!fovo-yK^M_-iC%{kLBG)CrC+ac|E4g67Yxx*ekO-W>FqH*darOO
zCX?k9;)q0F7*-^UV-@V$mal(Yjh}TP)5*8Fw@QzSebbI-E;R6Pcf{J~qcpxHh<te!
zW_NJfx%CjOp7!J3<WwiHA4za9_=6mB&3QZ7)idwv?AgI@Zj7o;ISNLlD+3^0Wwvi%
zD%^l3OTbV;6U@2&>3jocF5rn=BRrpGzdd%izk+VMxv}9dpUi7`@u}}(w?I7gM`31)
z5j;{H)yN#Us5Okl!c3opqM&HPHKV=@^{S%TqKJLJPJeqS$Vn9U`mju>sDLiy&RniV
zxvzH+0^FLq-IZHFK5g{9ox$W<OpA-Rx7{ex-JkX2qjF58eu(okkactZ%tkd7;35b7
z95KpJtq`A({i&8b*Hinwy6N`_YTnAKZ2eh3t&)Jw)NdDV{19ZBL;NEID2e05se`U(
zZ6zD#isv@4?q_Ho$ig5y|H8qF?7oT<TwNNuri%nCG)KP~qpa6JxxX*(^*-Y2&KCC=
zi%DebcfAhQ{+!v--eN=ge5RLzZsCkAJYp}V<XF@!U$a`}F{maCS~y(}-g<7Lfq27X
zdx~^-y)WPti62TqQ95c4j<Yc8JT-Vjj%H_?3NQ|#S{UbQK}V@6_0j!E<ouM=1|KI;
z%cm$?dxS#8_9%oungX;Be1;OWp+e=yz={0p6{IO;OJnn$z^vR`G61)Ls(O)P>aaA*
zagR`72HCloD3fY)!EEa6N|@)uy;rP~ZrxHHW8E-zpXTc<e<YjUwtw=Cm^K?O_c7C8
z{}aMY%2Pv^;;NYQJh;z;!qPrc*Mpvej(LE$^BZ$D5Y_u+bHwnrtIy>A7-EMHr^0Y5
z&zkg=^8qzR#n`Td;A_cGA%oovY)1{oEH-Pl>tyRzqCvD`^PBw6<-8I$YpLfg1#4-y
zl%?bYIUQJYoBRRa4x8CLHZHGQ)GI6@cXoHtM+V=~|7Xblm=Dbj@KGv%qo(G)S(y2=
z{UB)19T?vH&J3!BE&ULBoZ+{fWq3~d3xi?#qiH;sgDLE%0+l)xa_)qNa4%url~;V)
zy!0bfP-@WA-#~DBt{qUF+>mM2awlbRLFC@(e7x5AO*wbae1blE^==5}1PSuTW+bWx
z0o%g74wbdQ#ZTT7jE0;B%2OfG17@hkjZ}5B2bIqq-IItOx@l1s^g_$cY7VCNC@V{}
zNow+L(i-2sIXOS+@)<n&z7C1hCHK#XM_(4!djrl*eEQ{mP;;R6e_D<Np}gUiyA5tO
zA}*T5o=OjIVOtA@er}}__lm;wthyj?-aBW2_Tb$>PjqjWKL@Z}liz8C^}iWIOYqbL
zd!o@Cqf0*v?9KtCp^w0uowjHWm{Nn07Z~)|O`j)o<z$ThRuYjrJQ&{{1<5X|N40{A
zHc@GcUhg>x=gg{+A#sUMy|)(*D?uZI2C?tg8exXaVt;O*REPeBX2duHpo8|;#Z%Qw
zK=Y31%PXBar@iOU>?QtPnDF9}@t5Djf7s1O2!f6V_2z5xE~e>_q&hwp_c45c-Rj62
z2c))iHVTuy@e-mBqxrrAbClGa`Q-m?a{zS;-_uk53BNoS`!|6hs6=|O)QRvbfg9aZ
z7C&r3LWnx07gb|dX%R#fr8U*e3Gn+czum~E4nc>{qWEiUZ31)Z#mv3>MPA2Ni&}dA
z?x8A{rW{NCgYJW7=+3JS$EJP)iZA?caYclK+3H=t>ay{@<l`jpTdb<ZH?yyD?C_`y
zmb-qw)o<s?ZS^@@?86Mav#;Segn?lkpwONQtuwWVCa{}6FX*v`K$)nvz@xljg?2sP
zl?ZUFTZpG1+k<)LLt$|@FsJGz5y3v>Sj!6qNAiUi>Qh4l{@%yh1|q)(+gbvTjYXGQ
zGh!!6V($f%PusVis!ZsVy?d12P?5mc?#e5;wt!6dw_=aZc6hbB+rhQq<x1yeLyQdH
zMZ4X+zM8|K*J=yp__;C5rIeZ)yV|@7J-wzReFN`>HOFfz4U6^Whnwr}5(39TvqV*<
z2fk*K!|yFt{&gpd%sS3<ZDa&x=xo^p>~S}u<rNAhjE{X`y6)FnCst){mo~CVT&N{_
zE#l{UvqOa%Y~KShdtax!AwYccA|{4~efWXWkAmXs!Ig-z8T#iVRMdl$lkWh!wYa;B
zSY0j`LF%hiZ4w(Q>zL*Xn3S-Nj*j>|cCZi3rXtNM$;&e$nu-i49#$kxe80NJoK_aB
zvGp32FC)RnHc6DR!(0llH?QtY?Qw~GdeBws6B85L$pFu{OBE&wMpfeVrncg1XHfs3
zRBxTo%-}vzOLWC<Yv4YsIdPF!61Q!zXcpXY<&p|WdVtrGw>{d~`3$_8n)521IjYmy
z4h|9#fAIEr$#T0d!bb;(H^&{e--%KbUfkB5X=V0h#xmv2WA9EaT%FhstkYlFi3BfB
ztIS#be3;(2?D;3Q_VGdYv1fgw?`d_zQeTXPx-?dnYbMm$hOQ$Aan@{oZsz0T!^lRO
z&lL;A5CWXDsx0C#7$p1$GV8~a4aG{Uc|rnShtDA)a^b7wK96(wFuMY1PZc-rB`dQT
zgalDgprD=#5u=f%(O+TFw*pXDIEe`Kb9^oo3Ypm8@BVJ{g8<DH3K)4lVWbbsa)}Z?
z?^T*;tBdiK&zbd2v339eB4z-{$d5=0nF$VRj{moa5K!NzRbX2q3;(mQHc$`vIyiJn
z>bWKGhWY?+H&J06Mtv-TlpX!g>j<J;JcTW_zXv4RwuS*GW65~XoYw5-yfam`4O7V|
z_m6=5D<}FmU@%%;5=zgWK8|Jq4^_WD7?n(%T8w|>aA&RwFOkzi`0nPirA(ZL@_*t*
z412=`6-h4j=_w?1Yj-ycA|9to>=kHv?Sz=J=xXQ}1+((+<bb!qf%O-@H%?=;fPa+G
zH^HrdkvasZ)?7S1OJhY^Tn;<f542&xh15FCdvEgZfX4y_0qH$DQ~PX{8IV#7^;&cp
z-XP;~J`W=js2(%ut^NBP;v`aH_}h;>EW>}-C4*#Pdpn3xr=HlbE7<Ash+@KE`R4Mt
z)7wNeQF627-&NIsl8$2x0iCpO0hA69)j~5?z#c#Z$nFo-!Q}s)EtCW>5&`z1==#U!
zdQ190v3(9eceoF%K7b65RsxU0DXpjo`fPGjOLBkI%jAzHM!4#Mzcd_4jMDtD{;9k~
z@odIjfba%Df`}l{%K!|@Jf&>-v%T4CTVBed{|Lb-8L^f38-Y$%!N&<DHb<$o+e8Bz
z4nj>$ElWOGlN%ZTr6s<(xp{kz&^Hbgi2LjL`-6GU`)7es=-{{f`Cxxo30{2N2Hdm7
z^LbA-zU!(Fqr$oJ$>D6q@()mGK>LFwN{-h3&sbGSh~eD}66A4$A7vlq4EBK{`GB|l
zrj(7rbO$a>VF8@<(qlo?Kkvp$`alq@G@nERcyDMXOD3&q*<BSpfatrqxgjMZXEl2K
zsvrt18I))`k$9_g)gSS$B6+|<0mK_2Kq#yAxB*BJaBE-y+ODQLy~pX#tXqkD-P1V&
zmN_6g(n%A-z`)pT3_c|!B%IgI#lU#jlaJCTB8YK?i)Qnm&!S*=X6n!ZflCa~)D>$i
zbxOrN*87t&^_qV|pwT?aH!MU?%`c#PtO4McnL`0U$A2T$_`o2;^Za>PFfndy5E+B_
zMjKgakF%uT1k3#L_zNOl_)9C@lbuPn*78yS-t;GYRey@4E&gZ!d}f8%(J}&v=UVq`
zPFmIcEc~Cx_uGM(bO&ItuPx>B(f^LZOacsIjK71G|8ZEvP_TeTD0O%Tu;`e-u6sl8
zW@cxH(u5#I14zRVNxWEpRonr|W2G_(EoUc&{j3MI@{;(bKcp_DGZ0gT<`}T3^D%A!
zYt#==tCwcZ&(B|~$u9lX0bp@Hdo{+xd=im=oJ;Q?0j+@a%i_4F&Be>>gZUf)z#sfF
z3_X6j5Ycubh{m9$mis#dPl6~yIEYWJ$U=f1kRGtV42EIvOMs%oaXsJbjAhXJc(^3?
z0F2hDw^OWiFc{8Lq5V7OZ+yI{Pl}9a-xEL1vfvL?r|YvKr8!=pC;-a24Zv2*JnuaK
zBa?F_Mhfn6M!*}RLe$x;wgWx-akU5Ydk_{fBO_y;axPLR;3;RgkXJk|5izVN;e+Hk
zuW5?^c{UXB!Of*@FySjri&NuuH=y#`0S`L5quEOq?O&}5es~jo63kcmhA~QTyc0*S
zfd*jNV%`wXN?lGrFcYNs|9t{11TK(_@$m4>$InpUP#ze{(8z?(fw*cnyJ=`FmuS@>
z0ZeQFRAwB>AKkc!=gnokL@jYLUZN+_;&IDt*!BE0^_A%`Ik@SX)Xh=Xh3fBw2|zs<
zBb2Y%f41PlI_by#xmQ5+l&x0;08c$vA+-zKBNzdw?y1u(raJ$#Sr7Z88QexUG@hpj
z*%)LA;Luy1(}0ov%H_l>ENnpSl0x|JH%XDfXu|kx8OSjICzk%7VQ1>?5de%$0<dnQ
zAxCB67!CT9xO=9U(Ar5vUOWmfDL%x5{|@AOVQ()K;3K`A&$mYl{GzB7#B0kW{r-yF
z!Zj+2@7=i}HOC(<w20*W;E)2dqE8_Rtj)sWqW8}zDl+@6-;2d1!zBM%d)%j5T3Uee
zF0ZkHMXLe}!eFt*cHg$&&iA+HdV`pfNPoXu3J(_F3Uh?o=g+!6hpnxvO9!qqwNe%=
z>$@LM01-VHpBoLJ#L@j9H^5V&R(KSDEGEKwhu!?U<@is2m#4`|Ne_0Z(d}{=UYuIw
z&mV+(Dg!LVJQ`2rpA9dH0u@TY`vst@GeKCW4{qkOvU+e7fF%zmCJVRxv;Ln1B~kQK
zZUOFe`9}nqdesNa2Uo~$b2wcgl^>WVdCC09#>edeTnf<C<Ld#HK);#)C~Eq_t#eJT
ziskR5*i8nXjOMFC+O9e7&k+EMFH1k|`rqHx&xQqP@wWgx%H?`S&+l>LkI!wB1~{`@
zGtz$a9cX`6k>O$aE0vibvHo5DAF$?=<zgC36&=%HPq&YZpabl%7{J7n9+hV8|9!#%
z0c@!+zw*HWgeQO|zpgNni>6kD0H+jPU)y5&AMXOVzdLW)&(jF}&m1@j0iCz_o-&K*
z@sEKWUfRK1-E3IUAmjs`Zak1u>vBp9%-|;gD}q+kjQlH0#O1`{#6l3KNgt2b<SNa^
zk$`k7H`^QaCj~rgsu$Srk{<u6mn!K2LeX$9Ym=&c(R1nliF~Uj%0kgG5a~_;XpOPe
zdb7r+)-1Uq;>8__jL($@y|=jdRCnbkfPeECOZ<>+`kvJD;N9sb14HX&-@hG(i)p)4
zae4>9>Z8TYK(0steq12m*;L~1mJ)^oJJ$~_Tq)#n*-)ln8(q(lfL88)a3pDzDOXow
zYx;sjb)h^iSFwV8uy_%J4kR;(Q5I2w7Xd5ayZH8ruCDHbCkYlI=Vt#PRABCwiuA8~
zK!tqDvKWYcOy2ab0W2kjBII+@w`OP3YyN(5xV*Tq;HA^xU;r4n6HTr)ORZJvuhEJK
zy+OkWnbEJWdLdEnRL=?dE()9i9)x>a7%25ZGk(u86MO=K2a-H9fdWtfugl(wMboJv
z0XREW%55qK&2aIbQ9(TDT8L8eg2zt|nqV)ku1MeaMEB74*D?wS2mqcm#@X2!5W+&G
zC99cQ)XSq)HXx}H7~WkWcpcHcj3YbjV4p>1kaoO)0?OzJ55;-yN+OtDF++4?x~9Y?
zZMxCf_Bl4QAk77^<w1ZUpRm&oTuj9@!KY%EMyx)f|B<b)WWZBo_EX3GvtOUXma@)+
z2m}DmMS$vmFa<HKOx-uayDKYDL_|dR0v`B)X(<XA)6aN>h*7|COviXY8Q^QTO-@d7
zY4!uzYwZEP-4nXR!H=v-Kx{G%I=(!Bp@D_=Wzeejx>x)190$Tfd`Jd(iU7YHip%j~
z0leWAIow}&uN7Gsh`j);iK+PiB!$?9ERE)v=#>kUgH121!;ZS$-V6>7>$_GE8d=}n
zIMb__3Sv}5Tr0uZoHSx%P{}8Gxjg{nx5tVg`zaor{IL+P8))_Oo4cK%vXT^Fu)GO2
z%2`thN@_lg!UHF*3ux2_H}v#Zx$yv@*oa|&<h^m2$ke|(9R;co=c#p|SziU6;lC~k
zEEeoW%w;OSJNyRvaf;*41S*gokw6R@h$Zkubzv_|df-2I0%t*$bM3b=;fl&PB@J3i
z(}x{G@vuW;WZyyIbO96a&@|OqHxTWO#i$LtFd2qP_X4~`Xp`$Xscxea>O0Y4xktTC
zyhKL);tJIh={?lHTXzxLo7ZuVI)R}2*ErpAU2kLOE4c*rj9<vz$5@07T|olCvchd>
zR+~?X#uIscd5~U-2kG?;?HGZLqLd2+m|WVRTd=H9)W>yU=E(!_9>AgIbH8Hev)8)%
zvy0R@fby%^+Kg!fZp}l0eIpB{9zTF&AxCpu2kz->YWGJagP<?+G}6o7_4gH=uC|Eb
zcsEmP-DdD$l>N0IyziBS{R8GHV1l8P*38^p&i4pF{Q2M_&a`;y1;zgH5`eEp0wD}?
zISbOG|Bf>P%NyuA65F+3MlqMNH62=`lPrwKgRzV{JV&1>$dFke+-AvTVr6O6oZ0~W
zS(Kz$1Rj`B!S7;Vw4{48V&7AaPh*)4goAX#UN9tF>G6PVwu4*YA5H#ePt1o2R_>Pw
zfB0uZ5bvPs!)ODy;Y&@;OUwD@B3*2tf8w6Vq~G7}d-|hwmTqnUPtS*64grs-Zu}uT
zz!z7A0g?mJVgD~V(EQ=w<baFG2^p6KR16h0J$*J_?&kEcUr2Yp0vp!Yc#-dEH;Z8x
z^?D6{cYBZPTHGqu+qvcxF6+0LYs+BskDm55ZkP`NX>(B(Z*C?&SnI>Xf90aWy=o{U
z*cHC-+NIX8UNtkd@BN}=ATwr>EQqsZYR1(5fTh-2li;cxpA*X}UsvWbWP7wfzCECu
z%lYUX1MQE@I(}dA#&0~bA*#K(1a%{t>YgjytSK@UUQ>J@)_)!TI+uG;YqcK@N;CCQ
zx5d0luGx(HTl@OtG@y<xsqR8nw?jrQc!u|^cQuD{O`T69b~Q(fg<iR6YMOLfzsi{^
z!zi)RLhw6zNt|Wz<*c^qtjdKU`53#ZZltSs9avaZ+h6yfI0Xb@xWH>w8e2@FMB{`#
z&@LND74bv*=DODzgcS%#K#|t#cCp=IL<QVn*GsnEn4V{F&QC=gn;GmT40c<q6O|zQ
zy-Bckb+e?`BnnxsvR(EtJ)C06J4#~Wa`q7A6*?u$<Hyd`;z15!cPGec*G%%z*2E{$
zsGP@PNLkix@dzQ8ip(>c?CcbKrS-A#e=h7I+|#l_WQ7V6cU0i>3c?<|n7ce)GwAv)
z<+OK>afu(<%ISg8=zLBnwcPcjVEK92+g4AzqszKr!e7r92d;QT(`Az^-B!9PxsYs~
zi-PM)`CGDClZ=w$mf6dqyPQ8skDn_Pj9g%=q@)dtWD>QRlI8IU85q#aa>>od^8Yd%
zO@*N<EYU*P4(qnMYU8vvC-=uhv^G5GEk0Y)HwhZJz%_C{yMlb%<&QLvb-!3yL$pL8
z8;q3PuvjxzMAr272>83$yhiMt7E_<`aR?$tX&Ct3Z@e(1n$@$8d(|;gDhG~;lFgWm
z@CUfen9MbG`#UlQay!34T`iTj1&nMB_Lxf89_DPVvaTJOoyPMd6!fa-hO&)|n|x2q
z?@+m~bemF~XJQBq<IB9N`@O^^&}}=Lz;|*?m91o_eNrECtY(=8^70!aHK{gZd8-kI
z-zB_~+vF42;exOjKFJ+ref$}Z_%n*K{eDAXE^QFzl9q*j>-e-igJ$<~Zu}l;qlDwM
z;lNo@FK9e7TMgRua2~ze{_BTsrN*)1!yy%c;A4|@5{nU)*?IKbs*+KIy`<&K%szs1
zy~LJ0Mdvc;1!K`W0;j8AM7aZc?M;4RgN+)tcKbt+JGH9hCnyWgDg=ywwa#d;9c<oe
zt=LM3Mj!rQ41Y0@A%6R;WZvWZ_IawnK7af1qV4aVb(90%qT8^e5}sM#6ALaI<f_uU
z7U7^}JNHI@d`(mBu;y3%FJnvP*UvtE@Hxf~OI<!eOXB$iGJ-Cz-I?WU>@1UVtfsl-
zJPDIZ-^9hL@ftT)a`Bg&44~gM9C5>2352&n_4VAw5m`?iqh@^f_nn_|5bTyuMiKCE
zQ<ZQ2$N9dP{3n4l;8n>43>qcsM42HBPo=4^tW;*Yju75Tje<FADlqV4{;j6_#$Vlc
zp3V?8cl_c$R;;smZ-$@K$x;T9t~NlZ-e81S|FSsHXjFzw@uv1gFYu*WVZpg5ybR}l
z=kHamXff0Cl%~-=Y$2+*WUai4;*kRpU#jxcM2?|9%EocNMV53S$_IhZZjb<0QERz<
z;$X}DlasAcI+g9P!WfW2Lek->77@yQXQJ1R%DodQ4%ikm5!l`c{0hwLURa?{d|g?l
zY_EaMkPQCv{2>x+M*zkJnTd_$7N}nDCs33MNnQbl;`-8Z#u(PISbk6R$x${@xVBC)
zbjF<1%=Yg#W^=8wa0IT2s5Kg~ZFq9Ne=U|n6(}Ce6;1ew-09uya{S^pm$CDkSxuQm
zw>-!#eR;d9G#V!CRaBYCZuwjxtU#UbeKhrvC&mG4CBd7Wnhj1AhqF-i#&>XULE=hC
z;^?^H)w=^h`gj4DiiA}WUB4B97Lj2{p#>_kg|%C#-8-IuD>*HLl?@hwx|hR^z#fV^
zJa~AMB~Nf#9`ep#nT;@9&WZeweFDkE+??SMj7P|3FX3SMoKFc=*ONqwuH;w@dMZ-~
zeTb%4fT_nYHZ5ZHEq$jFMl@YpX%qu1t6PsswqDc){;$*dBUIudNS}60PHJ5CJG*Z;
zE^USCi-hmpK|wj8=2<3Ma@pUpy1BbUUd(z5tC5;VL0+AF-#Cm~;I1QF>JCLU=q4e+
zf)d8_j<XU-n~HY-k<vo_=BA0VvLNp2AK`-Kgn-`XXzORh`6qsW<%P;oNDkCc?q2Q*
zj|AQX@7L@#*e_46gC>UfktyYqk>;A-_ATx2hrJbpH&m%H%J@~Y7I4>QxILG}16@^I
zko4{P^73|JEW6H1%!_*jq`v&MHHMdq!WyWLTQGk}s)ur4jrXVKB-M<4J1$t75BtpQ
zb{u|x;m*uitouc41lOPSK2b6H6G(kFst=VF8JiYrH@)zIKB*Zq=y}L@5<?IJ_Z5~|
zYFJE&n3cHHYMwtB_l%eGCKf$lNi^)aNc9#=IU7w4?_G~nf-3R}3?ie@&PJ_Z!<()!
z!q8W)XE$kYJk8n66*bn!HS-innd;dKuMF|x#tI3&Ewg^_4}*}f>p~M9?jDv|?+$W)
z`ogvU-6}E1&H2T(>&4~>6+XW;w8zTE>F%^6m+c*?Zj;CDEvCdUu9KSIe8)95r=1D8
z81YKST&V%%i-E)dxmjsa`R6unbZ;?XXFShvv8)8D0?R)834ISFRNd9pD03owQ=&^i
zSWoW$um>dly<YE$woAp4AV4f4KK1bKOQCrMjUn}!HNFE`V*N}8*?Rt-K@+1K!7AMM
z`cAv<93ndhze8a6sBN^ZTx3~GLD%r|C$EX#FG8swxVOf=b7o;>8+pi#7yJ_9iB2Y3
z>;%%r<!(A2Hz%pMRwb<Rd9i?sit?4w_#mf%%Xu5;t!T(6rTED7W|<G~WRe9}FKzGT
z%<C?Qi3MP^chC%+s${*FD$)0*!`;KIFQ*Vi!z^_W^vr%-?*1Twv^E*~0V@$!>U=LX
z<9-5Vv90nuXDczxaIHjWkx%0_yTX`NF&Q7m&BY<QHr`q5lk(sao$8N*adY93HWx{%
z6V?~c4bCv^doYF0b2t`Eu2-Ake#m;aIO4bOj81>-+y-|FN_%~vI9fh;Kt$nvBirbD
z|I{a!sWWBS^1%H9Bh*wmL#-mgdun97Gw=qj2s(4TmBeH)&6{7UeewAN@_nT*3MSXF
zSZphrrd@7dTs}S|qu5!oP`{aZGwS{rOERL0n3Ewsgq?k_arJ}2emw<sl7Pb3O020E
zLOuuh5T27y-DNudV&~Moznx`DT}C*g<W4sya7XE)k)q#ZYQXThe9%0swH+1y9)w3l
zeBa_1n(k6@rrw~>LmN$xi&BC-<9UJY+%wi55Jk1PvXZ3`S&9l#VtPoOSpjMEvOnZT
zesR5-Z7V0q0}>+hWxc^=1YRxI-Z-($yxR#xFS2Wz$Yh};+Wz`T>YeLPEQE8t4(I*c
zAJIG{Q*l2(4Nne##fRD7NvK~Ytxh`jo-c6xXe}F>UUf|RK7*+RPqS^VH_WhOQ>&Jx
zt~qe2-1b~omFv}sr_T(M)KImGQS^^UO}eAe!-uP0sxH`zUFU>>8{Z-w2m4PH(vJ4v
z70K9TTY`VWzJPcL2!~DfJ7QQ<d7IC-3=6qd@d{Zab6R01sDga@2#%tbcsv%M(2e;O
z=JmVow4CnFG6Q}S;jPL`-@K8T38&Y6E?bm{*iW;RF{x+CwPmVV_py6FL^If=a)X|a
zu(^FY%$vc}Q_F=ActPpb9$F_Q3&DmfE7fZ|%1)>K%wuFKOv+pAzl`w?Xd>Dg$u^X5
zqS`q7J=H<T6ZVWUSZg^huMHAL|6T4ERQdRU!5to#7oHX~c(l}g>UGbk7_|{%>OI6M
zYW4-%mjEK{O-)3tL$4am?CI?Wza@XH)npiSOrYV}e)7kyNSG<)d*xIL;>Pu_c$&BT
znAV2f?Z1Qw1`9P!W%bIQ9gpoQIlBa>hWY+}V;#5iRiF}KI9ob&%_pbzdauKa=j3V5
zBKBw<ss|A}%G>M`3CbH^iD2x|vp&4fbnp2y*)8tS@OhWdM4s4|*K~M$#*!Gn89`_i
z$U<*3U-|K~;wzUKiKfGLm|s{bR7UO3WCpe^)e?*i^i&MJxDsbli>(r@qx@JVOlqsy
z4>2D{u92J~n7>$(T(@O*n@luQ@w;7^9D2ydlj87HefW&p;Rj>O9`SOzvCcNaFBo1l
z-pYMka3o{1EJOLHKH{(SLmAm=XSh31GeGC5Ze(-r9QdLFnONgS{p(yK%5bi%e;3>N
z;*Y9p%EaW#J_7U_UBM3Fer#m{0n4E!(U?QtHeSEe^3C9vcRVZJ$o1=I-1oX|3mmzM
ztTME+mP#lA%il#uG_I}B)(qW;1PvTs=(MuX81=E=;-gbnB>j{8MCpP3!C(IxTIJys
zz&o6b*=*`vC7u2pTHdAJdi2{`H4S4`=Ynj@3t!>m#1*A=TV&_u=ySGTW=ZHZQf)j7
zsdQY+%V85knQ1l-cqeqJDb<J)om*^<W$Qg;rAl_t8sZAYQ6cp0nTB%)Ry&P1{znpG
z_T|NUs$*~!$aC#L#~vKbQe}J13dXK=9i<RLNLZe?8l7MJiw9<~#&2eq)j}#!1^Obt
zD|=<Mj%eO|;Ds?t6i#I!{JbauZc|F=*NaVpH}#n^QZT==t2I!QxZm^1-f++jQ-&5L
zOu8-dQ{Hd;x>ky$l`5mLTgwG4rD8U1_s@1rRX8s0T?;QN`CMrW=;88shBo6Br%9N0
zdv!*P?+awi-LSpGOo0xO{Pye3I$5<&cNGlNVl*G)6o!-^yoQ|N{O9Q3!J{lnl0?Tx
zK01*T*qRh%+QSq|?v<ZUGqLuWKls6js}4ZT#eT4CRM2yOecXw8)EJtdf>Q@e*X0>>
z4LRc}KgQtz$5L-(lqHKIQ-zmQV#|Jy*FWFE+|EZ{?nQr$KEly94lNV#{Aa1}+Q-Ag
zX=>49efr?_;bq5zHYne21kE%q=U!xG6N5B)LLd9E`EBG6p?iF;1%>yoh+4#JltN7e
zkFm5(mlGhAjb*_h6vHPi@|ksBw+rUC;Q>McZj*gyRg+#)KO)+mra11=1nneY5Dur4
ztd&d$?E3XnZ||nt=~lnCRP8>FD`C-}Lfa8=7Sga)qe(%A)EXD5G*MKoZ|G*o#-H=t
zf{3J&_XnggeUol+e@r&$7ZwvC9qWu891`Nc^cv6y^qQW=5shbdsv#gn#oa(w4|*P6
zA@<Y|1Si`*m(Hv+a}1BX*W){LE#AVJ%!V7zkqz~T!dm_<hD=W|3;9i%+;^Z!!i3ui
zo))Xu?>2XKQQ>y2d|yhcGj`X)ufMghjUarIrD*FMST5_|yMae6rEj-A%&BC~ZP?eO
zJa2t8S{wzv@#9sTl`RsVjZ-h)848S^ON;4*zw&8P%wHX`RckEuI=DVK+<sQSJ=d&W
z{hlZ`WYZTn8$K^E=|{fDblV8~9I|(`ebU7~+)J*kvbI<OKPdK|<0}1Ic{@`T%!aUJ
zk?yOQ(MIbp5Zr8Ud%fSNi%Q=KY-lo5h?$&2@;ZUsi-2?t?R(mw@3-&%IaaSS8(%tV
zE24C5=Z&5^!3`dQbj>h%ilyDO=sRDJEV`^$kgmhl(I2GAz>_ZgEyH9ai;#17CMZJL
zXt7JqDoZ92+MVa-%M}M*LA{m(o|pC{t6~B&5+N^KO;cj2^Ef}fW*w4*LaMKij^^{-
zS)5gA4564Vi?$ahws)Go=Z(Bs5`KcW4I=V!+cle^gu7b$cjX*EA}GkZo^ScaY}e>t
zUiE!!a1x_g-+oizZ*=-0zno~Jo{0Gan!_G%@b>_;3=c$+7Oq2nnnIYc_&}Nt7-G8J
zw}sR7bkZ*v(_Zj!-5%L1RBm#^2IYqMR`?`Odf3h_{Ahv+D>WJAQK8ZLbVXQx-{11m
zvV-QOqf+7henF1YOCJx3j!wF97N;GRUxd6E&NlCEX*~<Rcr*BS_kYbjGFLyLdi$Mu
zM|N3k;f2n12Ztll4lwcsJi%l<gZ2?+l47#1hY*L$Fcv2I+Wh5719!GOlw&>m4z=cW
zvmR{*>~yh&c=8eNkA2qj!r2%SBFddw=W{=v?y<eAk@>XQ2T$S*(Q&-<ER_no`i?&P
zvV;3Q#QQ^Smbk_HU_4FNb_BDHWshm(Npti}aemkA;MK~pW26khD~wvf*vadX8RfC1
zS=x6Z_E@g+cha?<w{KEfC$$tnulG~h^=dM@WBS1I?X#%3jhZcTdDo$1enmLd^vb6k
zHguZl*;x~>who_M{zz~=Y}<Ac3fu`~FGjUL9`fXGv3^qcB&Bi}V(+R+JR+E;4c?~D
zW_kwOXs<cA=p9XSy_xCv5Yt+R^V;jT5Hv~aH*cUeBRkb%*Uh=_S5~#&%iY0fL`_p(
zP7Pl-o})k4aD)zl=QV^o9S<sHN}!<-j~eE&*7m_{%4Q*~l^&4)eb!#xX<*v8C)(S=
zRaTi~D>#~(6unuvGiSY6-CiuN>w5An0a9^umWt~r<;8M!(y5~K>6hR4;XgpD&*()C
zCqq_X47u55pK()}|Jpa^KA*B*6ZUX@LB^uJXgI29_w#CrrRWWeGeT93n<`1?%UBy_
zJJ`~SI^8C792)L32z9$rZ#PO9Y4f>g&T3LyLY`fhkcNq%GEA#=J5VGITA#_DY|XA0
zu>W&PAVf@$3OiTwQT<GC3!DUcONKn*rM7Fa!=YeMr>Mwq_epF+TR!YPpKv%NsUG&*
zlvTgu_Dm&=&Rjy&&{IjZM|&wH#gq5eh25d5Q?J`H^SoY$mWZyq2CYd7nMWatD^_Xk
zT-OoDSn6^*G5W$c3C)m)=1RjLB{8~ZGqQ_hJ9r~cItU`?U27Me{<Duwji#@89cc@5
zZC;JxPT^J_=4(=$NCr*`VX}E))|$uf-mgK<HQs-=t+^fgMBf%aoJ~se8#yYQGtPkf
z<1O42eajOqrM+`vcr2@OeXFt+e8R}co5BG$ySr(fp_=DeER?|nh)v^q&{~1U8DdX$
z>D6U!2Uql_-hZ5$CgL6@_gc?X%i}!GQi`l$_)yX{78$+KU>ea8okX1Q_2~?ii@?5i
zyLVxaK>A)ITRxGHL7%fTYU_QUM9?w_?4bj>@Gw^2`;Sq+|MXQ|=!94K+=Yh1Dd>}_
zbxQRA(R3C}adl1C4(<?aaEIUy0fHvDyAvD+_u%gC76J_J?(S|Ag1fuB!#DTy*7pZa
z)ts}dr+2TluHNso$PUEP@6dAHMjulU43+WAR^q*u?QV@LeeNLblz^g&B$r&abvS#M
zRb5i1|AXf0LhrxFeB+%$8U;wccV!J`09BK9x#CUVSzrr);rQE!&Wd=|Q49BB&xK)|
znxz?~V4!Y_qs-)^qSt;NFdR3ji%Nit^iceV$5!kPD-@5;Ijhs*4G=n^7i-Pqqg=1{
zotW1LGlcMANJwRor%VOfexp-Vwr<~ii?4d!jCcM*DAEc7>`t-JDPmd+j?FhOb!ob_
znRSC!Xf=@3+*YsFuluxLu@n$e5N_W3wpp9!Ng!z7SVygxMTs=sWix!7WxyDFZT>4X
zs1hxaIyx?+5RElcHO=-1MxB_VDDsN}T6s^_vwohq-zBB!j8&n-eI%`?f>zG*@mJTX
zyrfct+Vo?|=)>uF!A?QFdLa4vDG#qa;Iw+ZL*3&k3hxt(UU%t~51CN!kPq!&qX8xz
zCaKT@B<voQ6{wVSL^ksm#nybsZln>dM&=<_{p#fF{d2Nd9-8Qf-0mUc{^S4y8n0)3
z-#`C8XTKFQ4cUtpZ4nzyw~JR+CqL=6AWoyl94|BFkQM=N+HO8g=ZmY(w+Xo2G2oy{
zJ+#?Af(%iarDL$3YAkmoWVNZ!Xen#lF7#V=Ct?dC(4vNiemX)h$Fc>p8_t&{GT$6-
z3sk;TZ#+M}cY0yyKAeVS*;<_ndK3ThJ%KKEv1GU1B{y|)6{~0QNxA-m(r(Zix-L3w
zw6q*Ur@V*@eg)-A53ft5a<PZr9Yyjs@E2%|E65}m<Qo$dDw&J>1a$Dlw8Ir885g?d
z-*-Bk+>`9mnB9#!<NXb+m^VqJ^c3s1lM;Kd2=f))+OclI`-VaMUE@R6a;pbGka;?7
zGIHqeY0&8z6Ev*2MF83Ua2fb%p&V%@-@(g%y{mT7p1He}YGe2%Hd7?1X>Ds)2fji8
z81)dp{t$guTj`gtyTbZVUV85ilxr6fjThEL=PFSjR7(|cT<Ekj8rc2)*a;jKY3CjS
z>-ql?a$_@hxDa`o%wju-yWVfoA>-N~B1v?INcaK|52`Y@&U+AO%lwjV+_cY;qqtEb
zv#oJWU*qE?^KCmZTt$VTT<VO;*DOtS_r!m#*DWM1lCNiH<EBLiQ}`A_z;2Z)C-@@9
zqn1ZR<G(*%$W^Ln3vTttV|S)zG|OHG`F#(&k8wJrlsEF+;n^^}+e08=wc8+>1-~m!
zaz;CQA@6|o6z3imZ2{E6kYL;brUT#llqS5wH+{}Hm7)6WF51SyzOzbx4^{f*la_Df
zxYP9$w2^6d#dc-4Z)q3XJUV(c@>m$1CFi;YpY6U{3zL3kjw)m~mB~UZwtzC)*kc9n
zmB)H8FyL$K*Y|-^4ob0yzqkMU?HDnnYc%27n3F>|?;}X_oSwfx_W6ETi8lz1xy-c3
zwRy#&c{6?tcNV{staCtGvHzy4@Tw>$Bs~v=`+wEuhbJDz-UC6M%7rdXv6a{IR8+%f
zlOBnM!Ro*YUNQye`1*}+-1e(>QhvGuWI~ePb!MldL?dCXuB6C@w-DZ|=g_r$e1~yD
z;RN(;jE#cIn{F`jtIG3rJ}W1Hy0bfds{z^#Em|%=>&PNb9$p*il7{t;61UIF@mhS}
z+D2U>{-;-ySA4E{9m3YAl<EAR80Iap2{Fx$Sb~VB!wpt5+2D{E&&ayHi(5uare*sP
z>Yz>ZuZ;^(%^5BC`3yt>j}NaKMv@26BbL^ibRQu$Ps8VJ0UL}!wYmDM#R~W(WhzaE
zx~`9KY}4;+naM+Fqn=s9HjSh)6^@X-l(J)wGDpjGmkY$tHUVA7w?F6XHm;}2&L3;@
z=*4e;s;slpmSwxn7TcLJgZrZ3LWaqdeN3ag08IrKKc|#Z&;0K7dFU)Wlmk*mP%?JK
zT>3(A*UY4(eIkcmP|SAW{GS?#w(!1{W-`NlO@MnkFOU11my@OAt##q5c*hUF)$qGw
zm2+`G+*Re!?WYf>a>>NKLRG^jEO|k-v_G!IQPQJ&U!9%iXm_ghEU^t0@`=ls?G}Em
zQdH#C*6wWQ)O@WPpLJKIi7rgu_De|N*DbxFL}vZLh9=LMk(SN{zrVTRvBVVu-`qRq
z+LUwf6UAQWHSAh@1UX~BfLeP&GZzk3zlNRW`Q743Kh6L_k6$MfW^xdXGm2D)R8=yY
zU%v*@ja)Ll)4Y2D)*E}A4@5Qey#>$74bZ&hGjj(Cv#X=rK)S2ipHr&9diSFa`x68Z
zSD*eI!F2ajwonL5V<a43bcKTJBBx()-^FM>o>t=y`8(kfOe+T;W#><;hIf;1=kXlx
zc?f6A6pWAE60R7Vs&VqE4^#(LL+N&tDY*2ETVbE^yt0`>y)MI7%*-q|0+_nEm*i5n
zmNJiozJ!(C@96|qK5Ji9r5StFmkL}+8U3xY6F0LIvfI)z3z$khWWtSVAz!G@s*Om#
zh$&AM>(=OIAoV!EsLDTJbCaSikjaK+woV<q;7rXmoRXvaNsc$A-<4}WW}1#`8YSVP
zs&{L%N2<0ml=3L(S6Z=;SG9qL0cszk^z#O%$|%2I&*=`hRhsBgZ@$i!jY9DwKvOcq
zc6g(uI;iubiE?cBAB04zuV8F%9?7N)SBz`Ycq{~dVYNsTOPh9{eN{MHkT`2FJn@!y
zpDsA7A&AT_c1%gsX&5zmKRsf%8g)&w#Cp2H{Li;T?_K}<3{lq-kGA}Rs(7kQAWSh^
zx#Rf1lIphasOf>cD>?}v*^(N2-K+mnwh;qSG#F&xr0enWtoZtWVF~N&4L(f)uC2e?
z(!xFIa+T|<Ud?l=qv|GYcs(L^bC(4L<kRLGcCFIscGN7r&yg#*+r`mmzb|zm%#EiF
zm9@eTcX2bO3@%Ud*Tevbt<;#JXv#mnuN2+Q8N+C%uTN)JTo<QTuITlh!$?lt^BV<L
zGyhvUI#1*#vZU>`s`;0RNB;%wEDWBz+>qi-fljTkYqZGy`|fL({Fj+P?ACzGP4V#l
zKN|3~{O|ZYXs)HI-|ZXyGU?(%6`!%&-_a+E$<#usTI;taj%G?AUNI$YDzk4JZXV%6
z)VzH%XZq?+B?lueKQ$N%FUqD0)u$P|z{$oLibXbVni{M8FX}%aV|MlTfB8+uTLIQ|
zY5l1<xAsaAOj}D<nV_h$-P(SMx?rHS)ns2r$~KW#vTEwJ2T5Y?eX^iTZ^obOpxeIr
zxeCLLxo#}2-Z7__QAT14nny{`#clDBwdB+rXTol>==I%DVJkoO5v`R<)Vc!cXobBx
zO6O{BE8^}ptU16mxW?FBxqUnI;QG{3$#Ykuu;Spr&Dmr#3)N<=Yi&~U4AHRgB|EC?
zP>Tgcu-H!p@S$!wfwGLF`No{K27n8-QG2~W3+EV0zOBr_&Kf83B_o&-9<kbpN?s3x
zr1<CY?|x;FR_8A-^`MY;i1e%R(Axb~_@Sd@sYc!T%)3I(^|F2+kuB{yqwtBPI}Y$8
z@$Y#+8yxX;`q*`QnEZ>`GYc#-*>fiY$3O3r5Imy>zL{+g7T2GT&r8`ib<{C<vg645
zIh(E?2(aTNepB=exxJ0oC2N>a1oS=5e(|n>tN8~vY7HZGFWV#no>v;oTt_#*k=uEf
zN)E132_`Na*wq!b|4-;6h0XI(G@@tx&%xC5@INDCK{)WKINVZEODg~j$}`P<O|H0!
zOM>HN^!}hi;cU_&mBitxPy<)(*n;it-Pue}TmVH|M&(wi?zjHKGP@y?_T`?kN&WvK
zwQajp4$CGa3`o^BXg$Bv9j5b=7m7N1T^`U~6*y;P=MA!J>rqqCw9sn%TO<4Uc<SWU
z8$A=GmKWsgch&qz55XPF?nh&l@y`CDY!kh*S6Xoe-td@G9*fMzXtXHq!KzaMM){qu
zWHZK%j&0PG^myzHN0ya1@Uo&f%d9^}$54E^bo(AgeM}Q&(i-R5VBHht3N>&aXN76K
zzGTR>KV+R*%{~2T&cFelB_E?KB=n!yOMg3Y^k5o5;agRQNE8Ry5XAMhE!C-D4XgBM
zkKSBlKD(hmpc+>-k$Y}ICf4mX@a_uRa_Lu}i_FGJMYAkD!fw8`BNemUTh~)Y9<vM%
z#6z1>qp2o!-w@xsAlDxPv;c--4DyqQ*b%p#m*k+Wae7~9MLn)}mLCEG2^EFvbPm#d
zQ+K7YKC3H07fu}9;7FrAd}H)otMYWjxzBPMuDxj(4-WR26HW}_A@hJT+^VtD>iX54
z7uE{7EJI2)-7-na?=DEIUW{|+J!M-&<Ljy~hA3nAETq=J7x(%)eNvArlolcLaPy1I
zK)pk<o^Y||0bj$8-yd20I?reii{V|j%RM*moAzKmY_Y8mkzhp7d$qjRTofOF@lO$%
zU$y~3H{wx*xHqA$^LR)&`{Lp8h^kLDT%|fP(|EP2D^mppKw|iYQtNs4^HdnxY3Eny
z%WVi?!gj7NY@TI-J<?T0PPdCdZlm7o++lHAPdhQ$)W_$dt2bjqMM(OE4{`PGPTVyt
z89)RKdG?=9N{0S89CbdNDdqO19emM#80Qs1e#@(3374P0I~Ti9DUJTFJ%W4PsYh<l
zQn6CUne!f57)9Sk6kDPP$wMwtOFlg_D>rEwC9`V7!n8HP+|(dw9%I^~Gu~(~mTj=e
zGoWU%!DO!s@p`={MjoKCbnyGTWgX%njNvfj+k|xaCdo-Xuh~LOfb-hb=l1$FS^~DH
z*ZZ$^{(WlVUq{i9mFKkAGy@;6T!0i93fY~AxU#sSU8_q$Ox`C=tnN^T=fPW03Y#7j
zfo@7JZ<X}OZM*1;N1sj2k}1?9cX=J3%pK|^#su7s{CH{3GEif8ojG2IuHOa}RAtTN
zv6Yzkrm5=S_{TlAg@J{fjLO^|vh8Y1PM{Ea$8`Nr=I7SvKM&Zrie8L!3EjxN#fUkb
z(tl<h-fA1$YV`_3U^Q0vsmI9Vo9;_<0(~MvX@%~nyp~xlpM>MKhI~W~u#zZ{7+tj8
z^_rx2w&@A5YHNP+=$2c9_hL;ce4VXeQicrnhkBB^<S+)3e~^uGyl%+4_3|V6(wsQh
zZ<qQhw7ZfC`b9y3=<QlqzApd0+;TRd&!ATqNj8@TtBI!hyaf`Qh<r!ZBCHl70~_@B
zDOe5nIfBEG!^T=qyHx6&30%eaE$-;WW!0h&{r8&~kc%|8N${CE>bJpweLPWlzN)<k
zoFN6)<JQ&Sdi@w$5UBXBPr{^P7kJPnN77K}1}9EZ0PZ&`!YmVV#B;*X!g$%envF$Y
zr<rOO48;1=qFj00Do}DI(*_aEgE3&4tf%`V$qMlxt*zl8BAezp>vOXTwn-xeAz(ip
zX!eejM_Ve9c?AxS$}F|alccdWTM3R7+9wdir=<lJ5pK|E4r=hppnek>ECi>XWS~~3
zN>5LpQli-GZ-d$%ikw1q(s24ygI9pkGAltbA(rM^=|U2$9hW(2ytnZsqX*}uszrO!
zG6LsC!5(|M*xV?pO>Eq9UOFO&&08oWXsm{dSYg7tJ~u&Ux++Qw1#5^*daT{i9eHW=
z^4C~ozR)_b?6jlXaOVj2>Skr>Zl*|NIP*ilGKi`jh3;`{%GC9jQyei~M#_9>&JgKD
zQHW)Acaf4)=))oAxhSp<{#^9Q-&a4NUp^s5P#5FD-lOH<2`>l!IjS%06cQ?7Qx&+{
zMpZ@LYgVif>QGX3J$Agj8;U}M`Gtb}iuJ(gct&>XUUrzkM1MOvf%ae7s!0r`rw>eo
zbwtbT(CKn_ga%n(f%Wlgj7Z&z7L25209!V#lV-evZX`i@Z4K5e$rRS^Cg%sdwOv~7
zjvl_f7_kCWh4o)p4TzZf9y`7dZjT%%E6`ji{Kgj}>hD8-iXTz^=!rr;Wz%rRA;@gH
zv$f@S$ta0jXN2==_e?4AbA2{7JDYk#XEri`W0IH71-sXE2Zrtvr{O+#lx$nPG-9RU
z-oX^T7XpOb4?Z~`TlWFp@4R9LF|X-+8UzJ1=9PX~9qq@~nhf?N6^9POIeSn;&;gql
zAJK>WHq5R$HW7W1*`dsvzuU8!$J1Urp~7N88h~e;yQw}GZc7#4R&(C-`~<%+YSZ7<
z(wU33wf1i+gSp*N73=Njgc0vlw7olzEYz~sW>5y>_QxgF_HRu;#eg#~zi0UK5Sck#
z>=&T35KuJ-v%A?|H46SQ10aXNVQ;h^2&FCh@1nB(VTXAdhCR-27!ZRa>?b0QGm^P3
z?Q`pxizy_-&#?H-Z!c52x6JOFGUi|`E_P#iGRg?}X3|CcCBnQ^?W}0mWu?Xt?2CAh
z+4=90!UyJ>5Frm6cft9iy3#Y|Mtt7+PM}J)K({X-WydoDxX*XJbmDf*_lA2GJ{Rn9
zyp!}D2E{qIIm7ESfeMY-ca>E)>aG_dM%33co%h9E7+LK13ocsUS%CBCpXX|G23&m$
zwmaV(2(A&g^HA4Q0tK<d!AK^Mu)q3=1cw>Z`PNyVv|IH0ajeatTrmzqwb%5$qb8^D
zB+lllYSc^Tv#Ez7c(<1zUn4dyb*LOXyp^ykOJ4j_<+LCR+xcEEHwzCh1r+QFmj!Jm
zD0=(j)P!6%L(JkQLz2q-iO2m?oCpb_fmHCpWY{a&;b7c8mGg16Uz`UI6pHE_4y6ty
zV4Zyv48Agb5>88gs#paYz%seL>C?B(mx;~IM!3o+gA)KR@aoF#VlAkuM72q4sJr1j
zbh!)40)BQ{#K;WQ1^@F5f7U^SeLq01)F~wWpF$lNA#!^-jj39^_*FBy^7f<iUP`+q
zt}KDz)FGK<g(IzfNO=md3vZSEASDs$1K`;%1J){HNh;5mB!LZ@BmQKU>MRnrSq3zW
zSruLNIY+(-(S-={hnw-q-W_}?u4J0n5c&wmD4HEDoBBOPI}T)Y*uV3Wi1d9iEy9g2
z>p5W6-F{e@8DIl@!qjPf1QGwMgE4550NS}~d=e%vuJxJsqn~8`!tGqF(!dKw$`35K
zg-q>yQ(a~c5DH*H*GpTB!*4#zn$xomi^(<p`tlv493L#ogzPb>tjJn@8F(07>JaG=
zNmIqY-b81AIrG;gAk>};!7$DM6IXfF-)afm+r5Ttj5N8$o2O3L`qt(YxSC#)<n9J-
zq6|E@Y*}M}g9hJBX}o!kCNhw^)57RJuyuLwio@(*W0cnX=%!FUv#|mhXdgH~QG@XO
z-Hl8selmK6GgxZ$6Yrw*UOJtKE>O!_<x$pApIY!P14C*Gy)kK+^;dbn81!{Fe%ymn
zx=2pcC1Rt|%Q9{jMy}SYsA?!=-wXddY?pq>5^H}$=i?--f-34lLYrP@N#}I-H_<5S
zGioE7bn<?VWR8>sWlh;VKHy(%l4-s^4V~482t@fBZALEG?$|&b^jh+hbHbss64>Xg
zEY-*wC4yW<8cLMm8^>@|#$MY8ra{g6JU$5`TW?WA<oR!S;}aqgUljfoDpVQ!oC_m6
zG<p5$W1>&Qch&mudyhGRlAox)@hCS)K69b-kns`}N2DJ1Y_pGD_&c_>Iv44L<2{Vq
zN3?oleK>y(7mu>6@NT3gAU;co<-LYTKh?1|a0VSpz7?xA>W<}Uw6sl3M>eWkBhoP3
zaT;WdDOsm5tooLgr=J&IVp&7)&dP)|1q2eh`>M5SjG3SRvG~U$GZto@&TLPdqczBe
zG4YOG*+_2x1#Z39XYzVMv9zKBPQn!NapW0#52HvtyQ%ek<g4@^W&_7)MI2syyk7{#
z@A_T%{iZ3yt+O*D|Lrf7?P2mL5KrheKzGOeAyjG>itinTl5DQ{Cke;j5n2-J-kT@$
zJF%Cc-u%XuO&pU|!y-}88UEI4Lxq13rEjNUMcmZIMPxqu2xeOzE>AqB$kjGoCW-g6
zygI6mMJ!gC5}V?yXkPDc83iY9BXp0VY$GE$M#g3g)T+D^k93&Mg==h``OhS0Bc33V
zJC3thw5zWAhV)KQiR$(bSueR;`iy@M#81$MMzW_cnreSf@h(se(LLnTQM%4&1@ZNV
zX-3$0DKf}OcRY`YRFTjqikq2^gmit{|NR&exhCuB;}Y0uj>lleR&Z2@f4+9#Br{~6
za{0kTph_FO#@zqd{VTeR65BwiA<6&9`CA&e%>89)tZOT}Grj%+IU4s_03A=K<}vD;
z=Xa~Yq^M{fy0K)J&EAyS*N`wfj_F7WD`0n|lzo8!SSE_7!#N1|rh~U<7@LRmN51Jl
z%^@{`SMP}~L@vI4qny(TxhT{QUCKYUSj-W}G0+>u)d`^IiCKaA2fZPF4xM9}*fec?
zS=Zy$ns`+a0%xc$H{+q#x<z9a(7^u{t_(yiUmf(x)B=|O6)rjQ9^{oxsnKZZ6k{G1
z?c<stP2Wx9kdt33ylJmCAwQYJ(26eFBO=-br@I3==|LL|&%ce2lF09Xg-u<RyA6%*
ze0+vxdz_4}`H9ev?9=tTPZc@9;-U#`u#4bCWi{UZ0+6<$9%JoCSui{9+XS_W7=J?F
z@2BILBsf9mnKUVo9AA}L2<Hq;ZR2ZF8hjLokI%Ei*=i{+-Ulopg7dDkw#KbfXNPV8
z-?N8JwREYF%Q&}sG9E7%jZ38-JubT31;^7u1(bSyUiU^QBfjF@oU^oi$88*&@yah3
zb`>iGoSaYmh%6wQ5FJ$f*0mts#{j2+=8uT|NvwGa<wMRXrs57z=})wd$J)6!3JEoE
z!C{g+W<b^l;f4DqM~5LtkvRahDjDiz)bUzH<$U1_jnE6Zg+vq=+D-BI-xEcZw>Q)0
zuqH-~ey%MOQ7|1N{mzGy?d%QK(bXZjy4CsenEZ7eJo3^xbE7Ee-nP*jJ}@1KgYDmn
zrte1k_z})$TJEP=k^+dTjYUFkBTYanJ122INLs0zYdonZh{C14ZzPCzapgGRu<tWz
z#-FdIF-{f{<?i|pY}SUyjA9r<Nh*pcts-dtysOU&v~u#pY26|;Gfxd`+&_>C282EQ
zEKp1%9Q;{F{|!R?o}n*)xQd!LV&*%nlttI8_IhD>e{I`~`t0Tl#Iaq+TN|FDY6Dd?
zM3HwbmpUAUVvq{mg>lXe&XVW}*`213VO|ob81k?9EICzQ(6VZ1S?Pk(`d9TAhUWej
zDy#2(cI$oqc_?6o{w${x40)`IH#e%q&xIDP9=@B+Ffj(4DcM;WX*T2YjfQxlpPlK7
z+HIHw7MZ8z{#Pc}ud!%lIbvO>i=LM_ShV3`$)i||w1Yajjw4phdK69$9<KtC2jijR
z+0Ap3vIo!ZEX>jt2F|I=i4E~opED(4ipqX>d!rG~RTUA^>9cjKbkvLcrbbUOR4L3w
zrb~H05xKoaaSS2{k{sx_c%yL+zZ=!5+VrltJ;G8`u6yjabj#0gUPR2(%fgHLbd>Kx
zFtqNEk4P%i^{-?-Xa#lBkj%@)3`-!w*yyi~tsS*?utAqv$6W^-`%85SoqEG3QbAwj
z{d*o-8@_Ij1^whA$AMxvdRIWXw9?!B^;JR^nw$L!@!B$%D!StV^P4W;@!8Wb{w!@9
zc@YxfYJPr~lnC$9!OyQym=uULH)~j2lBXLYtP@)sacmc_y4cCo!vjnm@z?&os*NE8
z%)+eR7Exrx!zk3wsTDu!pShJwUH-O5>~<~0zdqvw7H7Yl+zK%&7@nH*%+gw)XB}J^
zE{4`0YV_9ln~6>VeWXXnHP)&}BmG8n`UB@tJXwH)wo(NeLSEZ#-6fer_vNYB6(egh
z3MRB%PH;D8iffJu$$JMW*Q;qb*A*@-*?hQs1!%7j#NVXBsPegK-R%Cg95z=fUErdX
z0Fm+|EH}KykA*S&xCq^~IhLhkkD1ahS^;{&5D$XSDz>&QcG92+rQ8<~%tz)ueB5AW
z^r8cL(=5;x(Eix-%d}t(O%6&e=J+j}{dmQskawKkDQ~<BJ(cu62e{~VsL$T+`Cu&G
z9EIeEV)HijczFPr{=_V}8$c6QRqmKVUa^vBvMaRf)th<PdeJB_(G6A;>bG1yMRdpH
zV>BNtRMAF5$zzONSDFe~!hpH>RPw|2W)(nXMX4{gX+5^|CEz;1utSS24nU(YQw6<0
z1KbJ;Oc28ujA9zKkQ~b4E=~vpgt=3Y^I`!WlpK~*1R&SQ&mNfJ0FN*hDoA)MAUyb!
zYARbJvg%k3r7$}Ey|$6(S+DG`-CBsN)ju=Be1J`*rEX6UMORO;#7xO5iRD{7E}L-8
zJ;P2g`#@*+YeDrGcx?7w{7gMPrAl`6xNCPe|B^smW6;vz^;)3b*Okj)J5)m!%qZC=
zSIJT4X+GC|bXT|2Txdtf^<Kl!_x6@_%^4F(A0l)_Fls(aKG6@}$!Hl?^;x9!wOX?&
zW8z=7t$&5Sj^isxKDpETN4IxrK`07UCj-pnm^Ow{#i=#!fY{e6#|_8tWZeu5|8?sh
z_amiP^yYACm;J%s|83v?DzJ`SEpXv!IOt*?X;K^9s6iGto_shwj>08}J-c@i6|VSN
zJdwL$HwZVI=_7-?M*=^QYu_QLWx#q<^wqIZ9pv0#9awqqQM-5QpJQ-&cXzoh&ZD!L
z-Uf~W#CP<fr-#jO4=xa*u&SMYY0E5Ya5t5?-Pt37UcJyiVaY!XrefeX=ohsth67`x
z@Vl5FjCDBl^|>on18Y!KFt2g_{YiCWU-#1J0-C@GrUm|$@*H~>O*<pnbpIOSWQ=`(
zO-eH}zDDg%1%7#>E#uPYc(Bwa!UotYA@4n2ZKqf2Zo~puH0k2Dx4|1Z2(PhKebE7D
z(=`|FX5Xhg(g8qkFhHi&V81yaTpYP_dM$%-xQdhB{tECmZ9ir_@seK;S=(?>R$2B2
z*WboK_!|ilCovs|8re<<*vx?=?Ezgww>Z$mz*1WU9vVHx(O16N11M%+w(9phKUc=#
zt}G%AZguKKN4@4Ra^(%~Vs2bi<IjXYB}*(`X6b8;ZR|q^IrGSud>^BiDS9GkS@ktL
zp1RR~7pdpQp0XOzq!`H*3J$N%N`i7>rV-!@|1KP)ISh!2s2<`mZsx%&MHBDUv5~}m
z>HOB9BJli6*KIEEZ;`a1lc%m|)=9NuY>84G-PrVx@8#z(b77VJCvzpiV*~wQm1Z4r
zIfI($^LDyN>t!w6VxOxi41iBo1(*F|6#@s2X>qpl-{fU88t$7nykC=1;;NGQuH{oB
z>Sb*V<6OL&_NxWoW=o&{^{uTmiNx%a;DlhzJhOcWUJ!4(1|2`7Eq4XCfC?jVlt3>S
zWZ2d3-<S%Xt<+h<sw)5Zj^YQ#e|0@8!S#{D(d@|2u_S>x*AXBCJqZub3%<Z8pVXL`
z05q@N45b@3goiYC4=8d#_Fejo<IU2TK)sGPHTrt(+)0(+ymRepYyxS`v~U;pVy#JT
zoA8aB`XHuUZDoF|d?MtcW~35HqTjBp77rN!uw+pw!9ou%OhkVc37CU>(w>GR&vEPa
z*v|=8d!`;{NxT3SXhicoAOWYAelk}@gKKm<myo&35htVzz6m?hEt;u<Sx5Rji?79}
z#}5kX^Y?iC4UkHbjzP)U%Y?Du{D_a-1}-Js=`pFCU)DQJ8~suDy+>j@yE#WdN8ld%
z^cslI%AdRU6FLh()@yJkLq{V39<A%=QENZVY!RHM(#G;yPU(rf?rLbMOToFe+3Wtv
zT!YWntfm3d2*~_$^P?}B<UGCgrLL}rOnE_^`JY3VH5aQK+yKA9Zw(ilpjSf>G-M~b
z>RdsdIF;*rdGY)7LN&%h?qJZ{E>81&!OzG#ZZy83tfji}YnWUxt<nElMo4l->?a8^
zQTpY-pGRSkiJBG>GzuQez&cz$TW*_b30h2++)uB|$ciNsa%Miqxo(Set6dsxxLPb;
z(Mb0(Xnj!Fi7~0Kvx4#y?d=!kX=*B<4wBGJk;UHARwBLiqLUB|{oc|KVe0F2rl%#3
zRFY}9qmb4v)(yY!v;2<9CI$*Oxy<>eOlf%hfr*l0?K|RQz8c9ZSVH^M{qdcL`pcM;
z5qA6VF+r7EF((-G^ZU!j6^ns_^jjmW>d}ld=!4X7ysWd{QI{;S@y1n+{X53MUIFt-
z6=_+Nql8vvK_q(s?J{2vtBvQ@Q-}LSNL#4f8TItRT;JPQuRNtBA&Z%!7yxc=+*IuS
zBTxk-@_Lxdp0hB3?Xxr^sT~CT)uf8d?bFHQLq;bw*Fc|CZqXD8H1mR1ISf;KHRz%|
zdH+bB<4=2wt3VO-KJNuRK6COdhL$K)Me*JL7S7mv8eb5@6np5171pOoe|pYtYGi0d
zccrYjy0N~>bu{gU*R)@w$iB63C<*}-Q9dPXzic9nH4q__<!N9d`5`t~tfPuMnTsWT
zgQsz`xQ+K%@QJV%&QGFBgzX<&9e-)6s{n)e5sZDhFdBUyYYQey{yz%<n^kxN@6?5d
zx9<Uq|KOKg8sKVYP+i>ac~yDJ>1o;T;k?r+etiDbQsuWk)BLj0yPx9mAM81^NY_b~
z?#6FSBy&EB55=_*>l_CrjV<mp`<c*rXyLe~STf7UUta|XIg7{bBZPvRnKjQ_wwP4N
zFP}KNE^#qkArutAk|3&CRA_`y!88#t^%R&N%zcjW5@(!&PfO%$j`q6x_B!cRm<|*d
zn7M?#^EK4t(JkWKLsdmO+VdLfF<V;B=37}>7V@=$F)QivMZakG#=9$ei>lAJf_tPf
z9QQVPT4%SqyK?4po};A#*1hVzU!Ngl(m)ny=F9iR3sX6EzABiE@=98-%q9%)Ic|Ue
z#)+^et(k6!Y>i1Yey~N$TfDBJr?Rilp0s28*H%~gmJ1Tsly}qJetyg{vH!D31oJ@$
zG$I?VXF{+q5D`tp5LV04`jqLbLu@>#)>x2C`x-%WJ0Y;Kgy0#+gr?Im>y;zt;MYQN
zePQ@eZgm`n-Qp&~<_UtAZ_QT?CnCl-=Ej6fSENhY7OX$69BgZG^hoscMaP@GLoyq*
zLn$QpzH#2=9H5n5$GzH@uKUKWdT5XvMLKr2qU=%8*p-R^qke_BJCl)-LC-7`i}uW)
z5a|oU6wX2^<HZQ0CjXG&U=Z#v9Y7O48d7%$WY{+J>k3*_3P8|RY+FgDCy7dB7M^<F
zDY*{~jLZVjd{7O?&AWWOG7<Net<^jV8pis(!)j_hO_8Z4#vi)>p!8<lo5-M7!bIR-
z@xMtI+sQ9ZK$2p^G^FvBphdu@52cfR`>HFImUI4sACQvMwu?xj2L7XQ+fA^XWYlu;
zqg_MPM~=0hOtlXzG8Cz564F}gw7<QG7?gFIkqBGB1inb_Q&XwgPWLm|*<8ST%w~7c
zHetZ`Su5&G#W~*Km#7r<Wmy0dur{&9Og@Y>Gu*VwlWOW?a|!Q1_w&PcsV_p+@zIX#
zA2&1|I63bmLm(HQ5e*Qv4zP6}`?Nf%AYTmd*>I5gMkgv*7&dVrLP3M=ng{dUJd~aU
zdgfd3uuy1$VucuTX`!CfaT){yLJC!6y=kpL?M>O!4R`JJS?J&k<%5i%{g{uvD>B+g
zzj+eP7A@81BtYM9iSRd{2*^_(e`>rp_ic1Y#lb>Bz(VWh6+WS6-hm>#lH^QGxBXA?
z-Q`{j=iO=wjzML?&BX0%SDGujL1We|Z|)m;7RrC8^y&sa#9$Y(AX)BZzxnvSKFSAR
zObESi`ac>v4_PiE;ArnzIgfqN7{gI+w7=G27)MU79?3_O`0hsY_<~(L_Ug;`yz5lD
z-l@!PsVOJ4WSj4Rc+`y$f6sLs6rXzh(87?eB?&5b=$qpW8mrVCF~KA+VPzM=?|a;7
z`gB2PRSFeRY6B-MSnr=V*Y+`={1{&;7z81#l7LNe@?`$r@6y-*9sG0(@>F+<pe9_t
zW#;cwo|0-P9X=Z^0M~F`$7iJ185S}<5>#}pn493;Ofr??v-SvEJhE;2EAO_?B$M$B
z1o0dJmgq(sKzsIx%xBV$D_A7UKXxG%^nddk^0&8kU5n4|`i4**t~Q_eah&7=BX={K
z&8vEJJ%q9RfTi@q$vfLte5#=Z?HZtkWv1K8Bexvo4J0~MHaxywTg$6&9gZ)gk=z$y
zKb^z!m8vEHH}JFx7s{^L?T#BrgUuHif>aW}Y$Mf1lhT=dgC5_OFj-j{*uw0ynH2C>
zI^q5m%f>9PNp96{I4A8jO%eZKtt2mQ^{Vj&E3bOp;;75+N=9XvrOX&4h4g7t>mEkY
z{N(+E!+5OF`E{)a-HR5jexlHOr-%nxy)>@#BsIt3U0@^T(+7NX7CXj=%|=kMFL)%^
zMj@JHmZ(%UqX?hR;gniJE+ZsLvzp2oo;H~_tjBiL$RRHnN$G!?ujg4)kZYAr_(?X;
zw8!RU^*^_2eOOm&V$#8kcVHj10gRB5a{b->QXuhi>ckHP?-K%Ir$0occ#gYiM4f1V
zJX0j0oDyF;`yQTWN$_ezGAB8}olz!EsFhiG^rKJ}^jAL5$k<K#J-y`*@$n_r42dS&
z*(px20!jAHH@$61JdV4|FgQ|*3f?#TIaZ`F%%si-tb(TjF~yybtROc-LelaV@fkmC
zp3c%PZQ0C5!8G0RjuGDfo&Wz2)*{cO>cJ>{#Y7^7t|!`4LcY#%T1mQ@gQU2qoX6b1
zo1tku@A`En{{W>tT8=)Q5BLcx+wRj7pK2d69;ls0H_c)3bWnx7O6j<}UmwH3+qOjl
zcP&!!C;_Bby|60x3sYIhd*ektQ{x!ZMuqL$)76P<@NTc`oBqStHk2X4{J6fdcM-4>
z{`LgcZT1yMZkyE!YYA*!y;o+N+9=-prhl+LrpS<7=ZOe8Tzhq*X+z!*?IsN+6?`SZ
z20<K^2wM{ys-L5H+wQIdh^<rYxIa1W4EzA^H^#Hs5l8=Q0aSjZ$G$~L<jd50yNl2F
zXI#F@X*G~G(I+Kv{<Xm8wnqab{rHGON5gIYm^J%T^Rmd7S=6*RZR?Kryii;ALff%J
zvh>_+vkwa!2HY53M@M+q4z>Bs5sWOBUo=G#@~#(6v`XaKRLbr&#*|z!2Yc{f_dHg6
zan}Bz)T&HWGqVvP`m*&Dm3GMO!o;H26}mV3QMM-w2kCG=_)Lny;yTx&J?H^^8@5hd
zuU=q<fM-&GLIVu<B?;~Pt?&&ATRO8<&#+A$b-IY*xxIKw;HsEP4xQ^`d!r(psRGHj
zuHRp5orjap&hekiQT~02DNnvrAG#RW5-#(3lT+Y#{R3W)d$?tqM8Qzh@96KDE4HB6
z?EKd4{z6|!CK6Vs&6eNF>8jTfqv4EUEV&UmvJ)@IOw2?T^4>$(iW~O^YG>UtaQf@z
zF2#MfN@|gl@HnuXJx+C{Hue`T05L@5(eETu$w3APi4c<)Al8V-I~@9bKZBv~mAtql
zr+J_%kikc}`>A3QwBmzWCN<Aix(!KQk?qR-5mDZ1_~}auH}d@6ap6o%4d-sW-EpDw
z57TDyJ)Cq=Qduy+4EUocdE*>k=41(kV3(^M&;W%aaAhJc8naYf^D;)IHi^0D`Oc=j
z=L~ZH-n&<CvI4Q*nG8xBpq9jwdQ}TbK*CCFtqNb&eSDn^J0M0zJ(_P68ixisb+U0A
zy$DN$)x95c3CsZ%5ikOjj)(Oe-XkE!nzO(31=0yP5MSKDF-aJEhqKAv;jf{x%H8k<
z0$bloe{+~Gxk~2~yl#UiUW(b0iNROXe{c>P#{ok@wBXC(3EcF*?dm03O^T2(ccpcj
z=Giq|zk<tnBSG|87&yoLC@`Rk;v9_qfw#zI!scY|qGo9y4qfun;?d%8R#extfM+=F
zm)e;0%xJbgW{aRS#*L7%K$pM0;q&M}FIu!ywm@PEHmu(6yGe6c3q4q3e{BjmAi*I!
z^du8-WoSBG4&y4d`CFkcU$I<+n-700(lkN921yz2j;D~EWpIn+qjxG%`{`%?qNQ2v
zHIYi>#eDu{7h_Pa(eox5$L>C%CPyIjNt^Sx+d9eWQU<SeKwZs&t2Wk9T%1bdALEm5
z1f_PTR*|_=es5^dp@+%Mzomv6V9oJbjV%vLLZgsud1yEM7&!HVi3so(fzO^7<c3$1
zzp+6$hm_1J04h0<-}=kLFW#5?!Oit5Cv**RUUYo_szMtZ1s{@`O!IQlcnuL02hv|I
zaCe&&LsKC)zPpZ~v9`Cs2i4PVCq+K>GYtx&vi8NDts9=Q?CAm}u-smkCS>th@ctlY
znQLp+qv<aL+dhTUHV1^cYeJ@z#;@oG?Qu6kigSTKgz9!77smHUDjv2HbGr(@twI14
zd_tcPAzvSJ@-+N7`bLi{At&Q4U#I)NnuXrLCDGma4y}ux3+zI;9i6Vu4F9LbCaM|b
zY(~l%n{h3;fXg6%hr=*XYkW2$bOIJs6Qc3@cT5=Qw~yAm<a?S34O;P6{KLBEJQe?!
z&i5RS%qw4buq1rf+J9r6o}uy<@!M||Ho)f$^zv0)`}cA;ki*0{24o9(MzBV`8s{%A
zmY%yV-wz}>;_i-M-hb$yQmrxc34sVRZv)}aDAxt@`0@M6Of2j1DSyw3?b*sCzQ|Q%
zi5Ie0)3E|-aDIL5z*ieH4<h-AZ0Iq5QB1CctGk7_!7<@x&5V`6d)prOC+>Ga1FJ3@
zk4@s2AAf|+=}vBo3&??dd%h&3>u&0geRsNte7;Z}=c}4y_b+j;f3F`(BY8e(91qC8
zc_F>C>X&v585Et%Gf(keybb%7P4%^+J>VAe7s?mbu@`sO*3|jg!~sUOZ_V8NX_Y_K
ziqWEYcV22qOzOr2*q*@seMokiltv*eu8lbKLC@?8*Di0O5KjyqNfoN0{BEG=j*PNk
z>z)x){erdbhtY;g9B_n^nyVK<StI>@tiukb#`(iYz~f#HZGU$OMNHaRHiT8LEz&MX
zvhOApW0|Wm%nO4w&+?#a?D=i3e1X^f6(?Je=hYRTDtGOoD0ax>LeO38SJ!t;%`|^%
zAVU9}&r^n^NG9Mv#}dowMg=6h@R7Sy?6|o|pUi;eeie6P@X!{5nFLJa_8tE8Op3v3
z|0wJi@wF2^-*e{v(d5<T8;>A*<s`^w*HLHY!`x!W^1s%4KK!S_Mhb26zb>_TE~)Q0
zN&ig~)@9wc9<H+p7mxnLIqvj2<94|ZW4vi*1uo>{Sttv@oC|x!xT0V%uD@zt2yGwU
zA>7K@xu^0-49*Hx4H6vqh&1r%8GO$hnjT;e>2(O>dy5G<@--HV3hZdO%!$aChrX-Y
zym_RFOsLBS4G!Bl1rc4nV(7YQbf&SVd`4lsar^h9S0nFXYcRmWtD4f|RcplRWiLyl
zoQH#Kua%Mm+!g)v^P>kMtgrQp$6I{rf8{-q@^b={bH8WgVp%rxTs!uHH<&}Aj89S}
zSdrS3&^dAX3+a6GWU)Kj%`%NxmIqL3^c}hOaBD4d;+s24e-mNFW%NyfB4(O}3o=6F
zy{C;5@0S=NL2@CF*Yt2ADu(Qi*XO-~lTmBxyftX$ym*2qpb;{8D(yhG0bSvD$L`w!
zlG`;0?JLI{1(>E;JV?6&U$w&e2w$!CIEkz5i?CScd$iwqxKM`v{GlNUb!;Q~+FEe=
zl~$X0#(4NN*Gf{n%Z5Oya7Rh6(*uEhbZjn@Eo8FCQdaH@ut711v%DBy#ze?F)ieSd
zcHqWXR3LAbeJ}(x+6#D{Hr)zcOK-NLmDwGSotQP~ME%o=Wf@^rJP5S*RybLz`0~E)
z_!;db7GvaFk=G4kzbu&e*?Te~SP{1G=Asv-uI$75P8k1hBMJO3`|3|}7lze1>-&Ev
z5~m5g=9x4WG9fB%iBdN%MB}Aqs7T+BG`!9nJmfJB9KDsBiskLLSIyySN(V}3-$o@)
ze(~TKp!!xRsrU-Lx0h7Vpz1H|3w}-bqRC!2{}H@cD`m7)cBSd@@qdP4xVq(smOc6T
zSJDRZ&qS=2M#i@dd|K@u2nwfK*oIdh>f*wDKbUTr75$;ePCTk=1eSADm&$zilobcL
zH_Bzu)q5gvwsxBJ*mQZZm5*&sl@$a(_GwP;msRgKp9TE}qDkj$e*bP~-HH1Jlq)J$
z{QXJty@Zr*@+8-2i^;NsRj5LRIy|O@J%5_>X+^~2Dfosq5?l37TGr*U<I`m_;ERm8
z0$S55sOcv0*@0BEco$0V<L5X~wH<5G!2e7CNsCzehgT)o$Oi}ItwuuY_&+wA>VJSq
z90siaV4%XN0he|W5B>f*p>_hEA&R@1b?MIk{LQ0^6`%d?t6-SQ@~yl5D(DPzBV=$t
z^V7N3K45XXnnbau;#;4x6MDte4y>E-3<10@I^AlO#XqAUy577UOn9cPnH`aja~ycE
zxSc0+a$fA6uM0Zw3px_sGV<<ebQi4bDZ}NN*!$|<ETX|Rj=XG)ah11WOa}8Q9~~k)
zZ`XSl)LP1zU-vb$#aJeLzibJ!V^p?*VqO4zc`tj8a<=+vmwOj2q?G%7C9zTFoWEn?
z0u}|DjZ5t=myY(F{X6G=7v)K`Vp+-k_WZ&UB&2#yH)Im_$`IFfz-rJtuhUpls8Ntm
zqS?w-!JOyTVu3@EW+~_U?sl;e#XvSnS4{t#)9fF|wRNiFv$7K6v%G$=D6l%uHmEF}
zz2%PvZuf95<;eMKLT5wE)e33`sxt@9z2%13sft#EC5FVrGYYenri#~<H9bcS%KqQ9
z>^68;RjyEz5fH+u2YBd0XB~Hzs_JK>aVES^(W<f^{O4xD@p?Q{sW=)f=Cp<CBQ&D&
zLpP`?)qzzebDPZO3pLVyvIulhRSs1-6nUy{bAsZ@!yD7R(@|d8c4Q5^)j<2MlNiQ<
z*6MJ3k<U#^@oVBU!_((khZiqj<%#-ibEq<SN~p?MbQ+<2B%=Ebh?;XX(_CeogTBdP
zytUYeA(NdwbhPN=yB)z?ci>&@isxC<Ea?hD-Wx)PdZzX%lN$KQ8Cl1gNVY~tqQ><K
zx`y(?&&Svf>urL5?^GV#XEK9RG(OFB&1lxaf1flJJ2e5qYq-fU?{CP`bbQ}Z&6{=n
z{)^+BtOGOnS23C4(scWwMn4J30a>W_J=39fd&y#9&1(sUgne6hk=Plw?q~hNi)fcr
z(r1DryYW^TSX6|7^i-mLf<MF&Vj`6BP@maM6+2kbKFXUJJc(y)rn3c}QTH=;IHoV3
zr7f7Cd9IHSsB(>;mL9<_d=46o+hkaB2aqX>O9&nc@8|QA2eJgJ@?+~zJOYj+<&EAx
z*susu(8?Y-i-cqhZ=iV(IcMxPY0*&-=U+~BRMETjQEoL^+U!&exuP0C^;$-KfTP*B
zzg;Dp%IkJp=LbmnCx)lErVX%Mpp5+yE~KyBg<Ux&CYCi~N-{}9vi@zZ9b>uy?0#J|
z8Qbkg77D*Utwti#*?QvI6AXR)l~%NqvY8Biv&#On#DkS2)UQ8LryMw(IW{X$iwz$I
zY}bsx(C(N7Z_)0iN`}lvF5tGR)z)jx`SS7RnV5|YGIt2}*do5p(}%u2o{L~E{Mdy2
za@$gl_j;%<tHHXw-;bz>aU#=wRqV5RZLX@Yp)9IpygEK>BE$j;`(xcc7XfV-e6kt|
z2fkk$Ec7&GhyQ|n|F+qE$YbXLGQ!(2NUipW{)OEBjeD_^!tk;6rd%%pjLDCtSB4jc
zYIY<Rdwk8!DvzYTeOaIUiuWj0qJ_1F7G<!=sOD6UY58?AYzAFxog#-(F775;;3|LJ
zhn&@SPEY?YB}{kHO>7WJ+N*Nrp|iP530(#U!UxR;n~l%|;(Ex<)_M|Rw8e3NFAj!(
zA(!%ld)ynIF6$#QRrg4;pT%&X<6u)_vl=BirW6^XRFLm;i=tC97F51?PrTq&7uO-2
zu-78!&W9d))9E7(hEl~Fe_Y|&<llOFX4`5)l)A#{Irf>yh{r@47)OXFNIevFivQ*i
z4s;8@$}H3CrEMn^E4BpY`c@3dr&0XN$^wwX+joV84py=^MYNX;3sc@;Wr3Y_VB-_n
zU<<;0`y6DBbmRp4ou0=?Fh8Zu;MpDfrM57Y)*0^DMgQ*Buy@wXd*o|!iqz1i@Y2n@
z7>PW2@Sc03W!V5&`oO>aJ=n7Nd&A@2!f(`nYs5BA&?%CuArc{vppa`hiTqbDc~cx3
zFupo{E2J?10*8(P8*a!2!*BocoVJ@n&I39~@KiCon;LqPBN30qK+#d;fY~5QS+%jW
zF<D_1nQv=0_HL|@m(#%4m7~L;{g-Bz-|Hz{0*gela9f?x3X9EZ7g5|4nF^k^b{XKy
zd;U+U$k(?EJ1%>zFaS-SFjcx0u%|>CfANHWwt9=X3|;^oz>kN#DB;D)_ziUrax!<i
zu>Xk~5P<?c)X^D%7h#c~TJ*+3&y!PW!Y&NAn;1Rk+TweCHZDVT%pfoP;vY{$y+Ojv
z!w>N~sep8NpOBO6lQhPHWttK1ca|q73KWmK?r=}V0MEvirCpBPbWLS<=_pvHjO>rp
zhClvzjP3~{GF;F@fE5N$a}FBJctm^-=zlr(BW{wh@{j{DqJ-Fj>DL*hj?Topk8&Rg
zmXNqLXx@-^MC}+E5h{uG1$PS&CiBV+pPFry1f#uw!<GuLW;Rqr!7y=M#Whh6+(BP;
z!G;^%@ijdAQKCA4+$zUmG8@aEKAnf}JzDR#GdO&xA&$$I6!Y|l6(PT=9ryGm!|-K}
zwq+A9maJ@%@#k2*SPwZS-L@$LPuOdEr#b&(cvcMb2>S0@romyITvMay$J|E!PC%;~
zig`u|-)Mc4b_}vsod;;M7ia~-tk;OFZCflwwGoa5A-s5N;7hE24709(y7*e6-&BBT
zS6*7K5u(zk1s3?oh13ZJag0gh+W3U1z~d?X8`kNk-9QxPFB5v5@=A+rZ}B)0{_Ipe
z`Zpn?IYQ3UGBr>Fy7LwIyq+qNimiER2)-nXy4CtoV16eq6zBaBibH!q<+?wi`dI&H
z6?tgvAC{rx<AUe&Gj?~JQ<U9d);BL?DnYQ(@hXkQN(JWB&$dWymq`?WT?lcA)%`4n
z$2003-D7nT?;>u`6XU|wPW(oV*Bh1ql22r9HyCc_#and9j~)xkL?nmJW3v9ckC5YP
zy{325#P7@Tp&NRkMyCNy;ayJ2?y746zI?YwdNasZTAU$ejC;x{YiB-XmWP8wk5fmA
zUGbZ<T-=+C01>5#ttG-GM6I9sy}dxcA8KKC;WW6xC?C62tBaMGvVPd&`P5!++RoA|
ze*#)KZt}u)L#TH0uw7=$AP)q)-aum@Zr#^my}qC$voh@ey)VdL45Y5;!WMiNw!KR0
zl!fIGfF6<{Nz5b~drd=GPW$oEZNC{dr`bEpZzHuK65v!1REBfdWzLGUC=y>k;D})`
zMk!G-;7emNL>I!MO3J|Pr9yUHu^cp;p7^Ty)#h+NM6SUsZEvv?xLa9$Z9ev0H}>Or
zI|6;2HGh9%Ca<HZohH;<U+hq4lvL6uJ9jeiklYY0QrkpoEnPo&m3K%WWGMPAjC~6@
ztVP}Fjc!^*cFRbMDi&AkJZ-e-vbD6HjCxe=Mhuwc@)qW@k22R$FKPp3JScEht}0@<
z0k5+^ChQAMhn%z73y4Tz;q&LfN#c{Q1JfTDp!|dg_giiBtj<tlmNN1B8;Dek%)41x
zb*jq+@vrQg-geVRj-}C6sE2euY08rZOXDD>2W{7B0>3<;5-xS3m8H^GsV^w8raMWJ
zoriu~kN(Oc`YvFYe+RM9{I%VbYB*t6a`;=1@M3^NmliZuXYp=tS;RfgB}~LM^8@Pb
zib+|Gc<k4Yse0Pz0T=_n?cNV*HP(1AeEU4Jh95jJgd{e&wkAe^aA1~^X+R6cK>)V#
zE}IOV(py!%niQ)KjbEEJ;N9{5Wies18na2RSQk4qVR_%<q|N($#HPl+`@tyMBNq}!
z5k0^*)N`>L)#j0ZVNiVemS<vkob|<23h4WE3M&1@wb92lBLK74x_?0=o~J5<uJmvH
z$NyvMor5FmzVG4Kb|y|Hwrx%(wrxyo+nU&#*qYd8ClgL=+wQmL`Fwv>?|-RVsqWMF
z?sL!DYpuOE;T-QxS!pwUgtS@470Xq9`4r(t_Um$~hx4LCSaB<H@^Uww2p9iB^=AXj
zD$omVar6T1;jg1`^EcY0yp_`GXX5;>k7$-x7b?{YB;cy}l$Z6BO-9ci9S;qYPvM=9
zaBGd6+b;=<z9-$MJE0D>(ytI0;87x{q}KuW;;}+mB{EQ;D?!boMkjwtc-maaEjZ>K
zrsT7gOXTZFth3*Ei`JYd75%uP@A~igr*U&eyBj#i`|=-c>;26Zj=*#;UrfB=&x%Iw
z+VKga-p$@y$pOUj#wh-Ve}*~u&r2LIrO^zh-PvlyjE)~+%U6uiydTf4u@ZTb(XG}<
z#CP;&>jOxg`^Hr@9g6EsMfW+zrKH_+I-Cg}yko_#chnnHmVL*whk||>_WDJbgsE)R
z<6t}n<Dlok2{1U10yD2Eog%C1`Vv_dPCp4CqNpUIcgSYjRi!<B+KN@rEMU0XK%-fL
zPu$n_b7(&$Q268=Z`UaoZ2saKw*n9#19VVd3k|_QbFgAeE^BX4D0%j2eIX%?lekfH
zE)Wrz-#iJd)jan}BDwFq_GD9pvqW7uEAix<Hx7wbJYLzBs^?4T*)qUOZ$wJvvthO%
ztFG6$R}p!v1*$pD^0tgKe~hLIJp84}OWy2%c|F^D+S!@<J+P8%5%RXq>a*5c=&2rk
zBUF16>ET?o(J-AH61*m*YDH_mIoJ7N-fL;pK}21Zv2{eY<1j=XOrZSW$f8fX3MaSj
zj-8eY>ZWGsKgX1)33{4=5~SGZGCpTYbP5WUX>k9Z=m|XdBTQEP5pm<3mz5yy&c<L0
z|GB&a+=%bUpMG-bV6uak6uR;NYA~N?mNKCWjx=&~8E&}xh|lA8+ejN~L@8w@t;dls
z?{CgK@+1VPAL)5Hs^IF0UCeWJapuI=WRm2G%4kNYA4$m;<OIXjc<gO;&v_OTM33c`
z%x>anmav^rsDn!*2)<v{bF9!sq9yw9(6>ESZgiO`=;sVk98PK9&!y206u!N6!5B|j
zP7yueZJKuC?&hwGVhiXR=GqNS8um*LH`LYTAb<GEB>nfJFu(^zMOv6A9Y-jn$w6nd
z3=|s&#9psEtV6a<%QM6(Ll4~Dd1gan%zfI#ED+227ACCkUUX}o{Z*Lb)d?x<odc(R
z+IjnGD*}H3@4)C^+%Jf4Co~K~OG<YC)}Jf*KpnIH5kEcoDc6lsV)5_3%A+aGfqZnI
zD+wA+C`!KN)xkYf@-P(qssFd2s6Z455Ke(b^P&YU$X6R`W$=MM%<6X1-9E9~YxQ6t
zEzcafMjq$Z#4;f(1t&jIX6Gh9KX^s<TRz-1#oeN1?ZQ1T<c0E`-XFv#tIQ&%l}^W;
zg3K|o&5rXHAmzbF1ZtZT0QE;aw2`bUi3&ynn9CP}`MYFKD5TD?!T@4Vuav%idHR1}
zfe<@w@K(V{SdiuchSkhzr_;DFLdN?#{8ozY=RgIE=x49~d~E8qtjTY>@5QHM@>bv-
zyhMfAolrKx@8!tj^=>KkDG=T;NNu|QHw7ear^z#4HP8zCR4D&Dx#*!(fXEy*7F2Kl
ze)=D12*Qy)Vf+b-k#4ranz{-~;o8pl#`yOBpda%pn||-8dmb;_k&1Z45xpUpb1(Nb
zJE3*oE0Azp5(Ea~3CMKB?;8PePR%hv=8uIDEAAKYV&t2JKV1wsXnw>U693kDh<gYo
z$pn7B0jx?#UA{Sx`N=hwAxBgPIO0{h<A-k`j`Vf9HgJ597+sz_AugK&qGCMX?|F3=
zYl6XZv+#Wxpf-(-<p7FVpP0YIj!q8l`a(^td}&Ya=h4*Y&Z4yUr2Q+H_Rb?95B=Zv
zNZ0whQU7uZA!c?0TTfi--?mwy1mQ)qOObC}>S|%o88o{bzRW#Y<Z#o#htc_(Sqb{h
zW}-#-+k69*j4IbAa#E=l=!&BquGVo~8X{TiA;@<VE{bP0J;Au^J!%@d3{3t-z|=`!
zoFI63f)<GJjcMOfpPS%*MEqpT-rO{e`^X!6`3A)X0a5XEk%B&(?2_`o=nDEu!D4b|
zZ`VT@``Z9oX>l5YfElXrsKa(!AHf}q=4n6VsD3OH`ypD&#oIYhKh2!_gY}~iviZ7M
zvYT&d#fKiNOR1k&E$+K7FUrkVMFXxBI6QC&$Wo^M%RSs{)wdBxH=4ll??9<KCzoLq
z7>bO%oL~@6DF^|B1ru_)K{%|_ikYpqWJS{Qla3#q`H-G{icoN*-;a!9$1Qp*FdUAD
zAyU6p0fh$(DHKG$^v+n#=!*Ls>QkP?HY4=-VoI`yxRG%d(ZhGi(P^4%FU@{f9jKLW
zBcF(<`6lB#o#omLy%wsO|9fL$1gVLMe7RSJ*Q(F}ePSZqF?pZV5$92`<K`1I;q4(>
zkb^DOZ>+^1(7gi#d;UEBKzxSrNd10ZRH+3IN%z}@C%;D>yefW8nn*^?yzTyZngx1`
zxcCD^ll|ZPZQ@vbgghP^XD0O=)8Q|pzUa@p$TwG({lBXep|(8A5ZsT|7nMC^cD-M$
z1FDOi1Y^7t*ihP8D+>t_4waR?91UX>pOgCEX+4)kRB!W0)XLC!%I%#GSiBB_+ZUvV
z;~f(!TRtO1{iiNLcnU57Xpdh$Z25}y|3a{K@VRm$Ep9N}NLTF^RTEzDXa(%BUJ68|
zz<l6YjmM^l867p;X*E26&FFO^N+SsIcUbZ-)Rsp(Iep|YA69P!Q5@Cv(<w5k-7rM6
zcK1rUBjmOFPWqn<`$2D-M@MnYyWXE&bApDYglbo{Qjni=>MY;kJa|<z!YYUn{qecI
zwa;mD#YA}-_o)lO1=aa#U8&*NuZE<qp^%?ThL*DzI!Lb^s6uzd+D>_=QnJjCGQ)vj
z-pJ{A$ZWL*^ju}4zR)?JJGdNTN-6~G_awMq9#Eef9(a?>A0eu^O#S1V@en(Yq8q1l
zfcdl5E0e>bR36`5hH8epfzB@65!5qgfu;1?jZnV`#R3p@pw;;3a3g?penTYW`yPmo
zsJXpZCDUxb1h-c%^!<*hJ8W~$l6m_%6Vb7mKG3>2qCgD)gVhzMS})?@jGN;8FJcO_
zDnha}O0XaJB1%i!NY46O11R0h>2=O^ZyUPb?|Ius|6FauKEs*m<3>J<uGPyW?A!AK
z<_TdoP5?spcbV792qAoGVT>N4-gur!u~5=V$**fnB75+kArMj42O9VljjYp%u=Q(L
z8@;S=6h4nn=<jvBA#$raOeVby*HNQLDO!p6d^$~r!tdHZWWKD@_BwAaegDKd)ZOEj
z*mjdl20HL_DW<0&-Pz_Woa78^kw4R1`qE^2{B2+|S8zMPRv8BRuJYKESji9(b#o-o
z7JOSDa(aTtL?-#?y+WR^_e_;Pp#C-kRp`{~w^sqG7zIm#D139>Ifm)_)!4JJ$CiAO
z!%xl!cf8SKEsOu*H>7xxTj39Oo<Qf<{{aFrxX|90Km*&lq){&%Bnnf`-fMRQFMRYA
zz?yEap&K;rhXF#d&)_bXmVdOrRSq&%<rb*v*d+K4UfT(k-BeSdmOncKCF(;~K#thb
zlp%b`6TA*wM&naZd4Mnn!eseJZ8q7l${Y&i*ov@w8S2u{LOsouXJL~wqLssm_)9ar
z-iK%GFVliCnhwIzl_9|Pu<xHccydzjvSGJID0AEI;sy_|p{-dX?#D|<ymt`dh{(}|
z25m;h%x_N;qk}ORmc-p6+ykX=r&TMSHw`+|Ll&3M-C?JrXg=uEf~WKoUHZ3`XJerf
zq?G8MFLVg{{x6VeX4NDcRYcBafM^Y2MJZ3S6qcy~epkL3!zVIvKUIzzWpG5`W{A1j
zm?b<M{NN*n<;~j!W4UfSM^R+j?wfKGtK%w$8r5KLnGcz=`|~=dG7t;@HX!e}nt&oC
zKezP4St~i1gUkw|a_vh5E)>kqGGB08#2wmzjZxgiUV`|@vE@kHsVu$}S35vmkNrOC
zyLvt(boHId^Az?mmp2#FmVaRXmLEmj3qrwdPA$48K5<ZdH|*daa?}x&lL$hIhCBJ?
zd^F;l=)9K%tNCtWP%0b_F4$S19kQCCDCN$z*#rdzZF-gmnZe}%|7D>%nntJjY~6$A
zH7-Xa9LPqaUxZgWqaQ3zJrns#J)<uHG^9>!xC+2qfAf*(@sESK@b62WWdW`J2vrBo
zND$iIK9u_XLDysF!;N2JmRwZNbZMdR7v|lROZHz(sonb-<m0X4rXn%F9^jbbugs%>
zv6YJF+AT-DE^o}iSJ(6rUrt#zV>nQ#`6m#6j~^UhAdup?!y37XzeAblfAP6d*mT=b
z*7L*)-!yY|IJZXSYOLeCqzL~!{E*hWMIH=Pw@hUAK*4IA6gN2Wh%B{$ZJg!jCFoLI
z_X{3htBZbp$_q7xZ`pDiCeg3L#u@qC`fQKnq`n*u42}a1F6C1QeGndNSHXE27k=(<
zcuMY>;J3YUF#Tx@#~@NL=O&n<`jR7WNisY;h#%$mn_y1hDZwh#O?WZ3@Jf)E1WF)g
z;KYZ_$gYEtFC4i{y_dJGk(L+DNmnQuXAHBrd+Hy;{0}6Gig+KFq138W{g=hhifu;%
z*b!Xc@(a>wuTo&A$LIb2<h{XpJs9A<5Jd_5g8;WD))eJxE+zA)KDYQ>O>~G3pl>zo
znw{wV=cT}H0n26GiH=T6ljW=Y?X6~LNNMfyK0Ng#;qy+wBb>Jv%A_J}fnR&DEVeO6
zzSni(wUyy|Vaiw?h&M_t%N=6ct#y!C;1obHz}j~A#cG_0n)2?t^D+*^MZS-!a#J9D
zYj4y*hsAjlHT}VDO;)%yLNF$ZNMzr2JD4B)a{jeBf~8!KhMDQLcc^z++;PtdYJ|%v
zLh!=TNLw<7SiE)JgaD2Xt@gZv8HDn`z!CDeKs#)9OOH}<DvcMz=`2<SZynYSC)cOf
zPuZiC1=A&b{K4W#KiE&zeSR9_eP8*V<juo>7$4rLDmvH0$q3`>piW}y<Iz`D&b=bJ
zz`0DjVjt&BDd~sO^R_$`L_VgIY;&O0<>gRKp73nJNniQUK>ZFf{e(sc_)Ii$l<z_(
zr1{l<Tx;KA&3rk_a8VAFm){8sGPD4eWR@u4mV$NRa?hw3y6hS!()CCt=Z{3c+>sBz
zZj3sf56Cm%uW44*b|u37l7@OD5&YI)w^ihFc37@g+q@kDbemv+iVzA*l{DbtUqMXx
zU@KAMAI^xkSi1~)uCRrDaP_Fu4JNN`CRKQUV(fSnt(&=E@qY&EDhlf(f{ow)g1=|0
zp?u+Zb4wG?c^7~Scwc_7>F;*-t6pz|77ut8tw|8vpq@66A-;AS!~8@|+NyWw#Hw3`
z)Ab+$8-0wlpK70)UhwAWM}!<4QCyx-WdfzCP7l}a2W+B85&qys1|jYnP(n$~kqz#5
zy*$pRgc?(MIiy65KP9%mlR9RUn{Otv%XG>zm2aEGIZW`G3@lB;!3R)SnG@Ei5x)_m
z;F$1OUPc%lrGmXHvGYAH6i89BmOFjBQMjD-Ai6zZ>wgt$k@#2>We#IC$VBW-t~*yJ
znprwLHkc<8m2hafKi2ov9agUpyV}oZF<s{5ptVmZ*{F(o@1R5|8aTEO?(F$y4V%*q
zOu>O(m*D?#W|I8fp}@cZZ(9=5Z#N4-NF88XSx2)Z0IZimzp~@LuqjvJAz^eE7=HZj
z-$r~*xaY=mZ>mm?2-;--WwFEPi7+2n5o&b*1Hi24M&j+rUJxxhzTk^`j7swx7?3vn
zgITv)hrb!g1ZFjJ;fKVFvWx-u(PjlpSf#&=UxZ)H6VnAguseunn%JfhdI7|jzWSsd
z?#5F`9Ltu)JU$P$(OY-ik)(?6^<z)Z=4<wD=<_G*)Rjc@c9~&M!ai^LcYaYi9EUK&
zc1o26?H#p0%Ow{?1z`M4@&q62zM#au@jLA>7eIrfM0eI`YOkPgD1yaZl%2*ur{SP`
zHtb<)Kdmu$%G?Word<6RiF?ak?V^d_^a&?2ySCI1IW0%3d>tXb`wH<*vY@aP_g4E*
zw3uYH0K~nP>yBd%_Bg;3OqgnO{T5z27YHudPYu`tHWW-6uM$Z*XC9)B`1Tw^gQf9d
z3l9HfD1!&3Y!rh%Ht7hK8npUq0-ZUa1YDZ$GT<@#NFYsGw6Ixxn=`KmZnWU!yL#q5
zHzs4o%QNiUv(7NDg&U;oLUAtS5>6$w<TOYw{{rN_8b8{AAj)1CS_s<OA;SOiCZ`kW
z)){Xi@w5a(f|q_oDoKup3lVp!7Eu$2ecEyn&8JRZY!F9>H&;h}t6f!ot1*)qtRCPn
zxXT%^w<&H2s+vaM#O_+4`x`hj;)`PdW~6-VJ!K^tC-`Y7DHZMesoEl@{lE@L;a7Ns
zF7)iCPbGeZjU<971mv0al}E{P3EpLHi>Gi?ygtRcOr64{lpD?yaB0wmlvoIrO;5|Y
zHSKo$W8A7O>_8`OjUo;5>wtYYt>9I^Sn~T5Jr|2Yjl_9iG>V0b6#N(ut|x=^jS+q<
z_Ey$ipwj?@BuK;ZLh5_3-CS`z3@Ruv&dO$g5{JHsr{Xb3tgyNm;^&>yU;0WEt(9pR
z;h@!=jI4TC&zc{0)})-%{mMC<5PR5V`(!<&8%xUUC;6P4GumfxeL;_lvH58kNmUL~
zGp+WG6IAkXiBevD57~%Agjl)9i<tRmx~ellh`ZawgsgvjB6f3pT-17ojDgP;*6$Qs
za}TEsL!0J5V&f|WZM?|2MjF<Y?Eh^Bd|dEHfvrzd?fyc*=Bu0w`;Q~R9Fj{$fz2v8
zVcI@p(*QRFv^SE@oU=%L{1wh&4&8HsR2<&sxHCE9!H7~8<V131iL+vv$v?koekvY}
zb6?te^aFaY9#t90<AK@{#ORs<PR&foh80XRKe$S_DOhi_elvP#dqG=T-xN6@TN?Kr
zWq>mfVm+n9{IGF*4=FW2dFOGz;`ODm0Q0%WE&aTmACQENro6UqhcH6&&_hjy@#C~%
zOZo{{-~Z9Kd!r?)_i$bV;NYgPtQ)AgxEmK|!W7VbK;KxTy9!#x7BF#?%RW%MpqnF1
z=Ec7?#XQ_uPI)2`Lipi5-Ib0$H_|6lbfEl8Kg5PcV@#9*V~ESbTv^n5kRKK#5Yp_O
zNx=AVDLGaLJ|b7ahX^rS@!(FSE2%f;{gX|`j7@?Evo+ySKK%CotzeLPs;`^<ZHOIU
zaKq42r7o6hlX{Z5{29-={1<xAZI3NQ_i)i*$kO56<4RqPU_C92)(y)T)l;&^ehv!6
z6<Z~dKC+&j9u32o_lwRY=QO33*f<q@UpuS>K)p_#&)(V^&-O~Y3WO0}Bo`a}PKEo0
z2f-T|;6J+U9W0*T{g8QRaenx{ji-UPJ2@aMy0|sErxJIX8KUd7IG3FQva>W}WT)#V
zciDuObOj*i&ev+<dl9(G*V7GY%BZD-T;+ZCxM`@{aAtFdpzdXehU-52sdYbb$A{L6
zavdJEFYSL1AxLN`AF)gWP2p$;az1LT^rRhN_sE>i?sg+V;>~|B=J^)RN;jg%7uYEM
z-<3o%wC%vQbGMm(gBPP>*IezghHx6wl@;Lb1MK;kuITj!J1OBkC+}@!*8{KBhGzbW
zlY*cp6Vi4TAJLfA%m4?c02L0`t0Fb?Q?OTK%uOv1CaJ(KZ#fgeHa|lrL-79Tm;3aj
zC*ZGc))x`&GQ_$2^ULBQtZiFK@7iuA6G%sGWnFtZuqBV}8PzEVLyUGIAIuj_e}?n^
zEt5HD4u&6$#~wV2o-%W_VMoeyF<3eO;E4*HqoX#+Iq`fk`ncP6M}wBihlC;O{C5Q?
zAb>%wDKA+uFjF+?SH?tLb=KJ<dit0Uxk*KHj~~smR&mGEALOM?mqtfEGGtx4lUZ*?
zac*COwwwX$_7YxIW!^usfyW+D^QyzXW1|>jO<!hZF;@$o>QMwtE~D8wq?v%AnKsGQ
zR564n-OqQ81A}#WKtvi?Df}36dfS9dsoKIIyymh*!4wF<k8kP(hD+a&o8<YRk|-|u
z@|IVV%Nb(gy+cWEG<)PUEDL=<Xe=4d;F%5wI<X%P^c2*<i4d%1a8~%VZFC*jvf7+7
zD2$-BiYf54nEJ)Ox3LkPt08{)2ldlvFX*d1p!zb*aOKB+IJKmkV+1jWfx!+T7l^d<
zq6G+n+6V8H^BTk6>~D!HIOlnxKM=!4q2`6GH4=8AfB5C3usW0KPkY=Z3z0bj>OKU=
zLHA?=6B;ccD$Lrm^rDR_c$e*B8UaCPEkfXcxBSOKC#n_jh>StNm7oL0@9{Ig?IL=%
z@NIN>gzzn#dto~MF6)7)jW{%HqbupGX%H<TVmNIV?*B2kAfW+7^320{{QZZe6Z_Gg
zqXo1^(MAOv-0UU7j|<WrZa(LMN<8@$t1fvV1FMG<QvJ3eP{Z=H2o{t5rpe@TCG6g#
z`mWS@uf5EIBgnhUBwF9sLPn#H{WGE?LZ%x<H*!bV&Su5!I$xo-JoiIVwKZ0i(cm8U
zuQccJGp$gJ?<vZf{p9LMn02u5)TDKL*)9e6I-x;d=pO$ZT8QufcnA8@3<!LrNo=`f
zxU;=CUZ6<Lnhc8SY-@?Kg0(ro7Hu1wejZNtmtXvl^nImTQ84Y<##QKg{{yMiubiif
z1MVL#4RY&%nC9KHF9+z)Ut?pBNWzkG0Su^+;zr-GKtoT2Bj+`*`kG|O&{?=?DU5|)
zrPRAa=lmD)tE`<a={ZPON#q+E#^P)YhC1cdgopKs#^|?*=hJ{8@^6ZTZhj8+umN_U
zRBiACffCVl2(`woX}Cp_=>nR2rrNL_=~h)_{U(NsR5y*m1o3l9eQo<Nkz`n(@cuL@
z6C>z86pT3%m57_DJ&LbPJA{vXmXQGPcWegHk^yz>I!`#CGp%kbR`rOv*mc~5+~|kO
z;kfTz+}T1PJN`vL=6!_cA;EHOZOO#Q1jZtL>R_e(no?0DL3EPl=RPmu&C{dcKMmI)
z@xK+SNZ?|33q&H6wf^|s_Ivg(*{{~03KoM|i%O*V$RnSh)Q#rqepFLh^@L&8PJyFz
zlH^7-aMw>S$MU!*xdQY&cy^>tt4-pF1b<UEs_jAn>1$B73Y(*UVHL;Og&h8Djh_g<
zF3aQ2g8Ekgh2GG(A@W0bpE-wD*Cmv0SQZr<g1wYNM{Y+i&sMzq6&>GolX2ha5PrSg
zKJX)1DM56(aGo#fdh%A_EOWZ0Qt%a!4PDhK-X5Rh%xX$=111>wQRWoFcbb!p&0G`-
z3!vi<upkR!lPL+Kf~BMXdxe713S^K?AfKyk(tr}ACLo22$nSbv(h;G&uR2f|i7l{A
zk6Rf;ApOaYG>CY)?W}93Qlho7js7$f$@hCM)Pk}y{2ge}43hWJlVW?Yd>?(RH2W}a
zT>)7qz-za0UkTg2ua&kOgIA2vyKm~6b1pTN4x5ZDc~<W0!m5`@mHKeWPn)=EqU#yR
zr`=sdVZ_tIl-4&v3851Y1#rFv(8ye?cZIYYbZa3M{<X@v54gm8By8(VSjCbNy!O#C
zuFJBIJhB(74DB?vFNz^1OL!`pYuU>BG?*AcX5@W2FD>-m1#cO`!=MwPhL+_=2%UK-
zYxc`6KMGL)eqiK*OkdXs)F#MG$q=4>7oh1rImCs;WLexzPAMZ@hR!}z6y&z4m%;k%
zV@wl(R^UziiCPK!2{807{Ft>qzMUi~vR2qV$a}o=7-+jv=(jT^X!?`j{OITbLjsS@
zwNis843=B5W$`f*V=;Bt>po#*M%ANF3yR=)k&iE%<=@r#wosoX3FR#3OSQ#0ib$J5
zQ#mnwUoL!5u;v+*f3bu=T?EkvFyX+sx#&F4`{q3u_a6YC0_Fe$J}m)ndw_wx%`Aac
z%QztPurR>@XYbXKB2fs}##^!YRlgYH6;1}L1rPp{{?6qo;aP^;=(ctd&9h82&7Sg1
zH3h-kkD4t0xDie`dA>(18!zepa7p$PcG1s$IfR);`BPG*Do_6fnlutWjXVrk)oVv1
zJrpte&2jyjIr|&dIbMS0zkx!w0OZ(Gk)|z;C175~3&XX~?-6?5>L@^-2Q&lK%u97<
zuh=|W{IRCtgHV;WlT-x4amO`H2eb0U<IV;I=Bj4}t;dQCAn&P4aUIvxG{n1(k2Qbi
zCuk;vw6lbZ5J-e@ng;SN<@yK8AF1(2MPo9#bgC&R;5OF$U+{j#3<6Rpv}H(4kE_vT
zfB%o`2qmw#`^!eC!-qbfefo5%P@Z?Mk^s336V@wf-KnSayqntKg734A^;H%>t$xg7
zOe&nUiN}*S^>@}$KKI6u{jKNL-%i3yjFtrZ+H;%LFVo(#S*_s01tML9u`yoPZf`MP
znE)V~(XojuEUujkwY-R;Wk!H&Q{O0)y=r@M86?RZ+}6UB{xxR)U>4+>Zty+1uMZ$8
zP#GmV-t2x$zvNSYb#ZbZTEZ9KzRwh+Qs?_dRw5UGiH8y1IIIWl&{e9#G${veeQ{eM
za441^LBWsTGd*6Ak2h{uj2W)dw{gjM)<4QDgLlh&Y;HHOzVSq3^y9gTraz=^%}~!Y
zUfxpdP8fr-k7!H8j4UKKMqda$hrQ5#zS2NxjD6vf7v@u<jC}YS?YZ{5r5dd|9+565
zXJkM8Yxbz3(oECW#(A;4Qb&A(std%G=yQ;Ga^IfFs8H)W{V2OyX$6h$*lJn!+vt?&
z(JJ98^QlmzYX`j!3blucJ3!A=<5jNa4m4ZU>1b2jZ009^tpbYE3`NUGf#TIGBY#8w
zZho~+lZK&FI3QaHG}i^CcA(s%gbfL&XFm-1JZzgnHQ08^&zpYp#0%SO?)BwHYJrua
z?sI4?P`%8F|CrJpwI+c-!H_&4F9-Bo{JcRq&dr|)%rbxD-TvhZW>D=fh?2uJ=W54z
z`cYWdJG39=;k>GE^o{sU0F`zQPHxj587EcObSM&iF7rQ&lnj&#KKKb!(gfJ|hku@e
zb-3%Sq8_k%r&ihC5;@$#2%)*@R*pq3!ed)ZO=6#An!HD+MN@#0=Gltjl>E5(*czo*
zo{-Z6i6(vmBNrgR;N8fZcvU-RCp3N64U(nQ63vt8WVkmFm&iodb9UMi@VXxfkB^G2
z@Y!}>U*=$ZJCY#3`^Cl19wQ5|`Fs**auiT-MU)Fx3ZGk7o&E}kw|W%R6hiboOWpBu
zuANWFT1do&Kpmjx{haCH5#=W>4|?3+9b{8(AR-!@NGb12Qc!MsX!GB+cGPx~8{XeD
zn5eiW@HTwui-x|%U-P$PtaC95?&b;@o?oGqV#T&rcG`fmG6_~FQ*DkXpS<LM*<*vs
zeb?pc?i&39Vc?AP@19JMqICn<&F{Z4{>N~DAkPR~J-Co!JcG3W!r$(H7)T9@!8yDC
ztIuv1Sn_JR9W^Nb`y~jG=5;sAp5G<;-kcvs7zFyif*3G@ZI^i_Rglx^{}8;4^*^Ja
zECt=exBvF889CbPDN$aJcY?`FGLDsvT=Kv;p{VO!Q-Z<vMFcSN?hdFF%yY}EvaOv4
z;ng@8*0qPMUJDgI?{2PM@TaoYE2>N@d3r%fnsL;e0dCz33yyuag~=$H!(ro0_plTW
ztFCBsdmWjia1zux%8v2QM!%DH=yEZC9h{{GS@ovA9#$CkI+YxqMFT1?G@Ois0ck2E
zAFJKZ#6u-(*y@gs4Z)IGAN$2!0y1T5MCU9Ea-lR8L?w$uT^xMcdTFL?O<KEAnDs`9
zk$!!TD_Eobk<P7rkutHTV-4G+d7JDk<RXmauodu2<UUg=c*JFAN3}5AI!JH#`czTY
zkoNCk^cH(ptT(lq2E7S={znRw=IKumgA4KqfR5-whu)kn3DOG?rBuc7+h!}m?I@)F
zIh5o_Y-OQ-B8;Z6eOo;Dm2iKORqzP6oAg=<uCX*B(7xM-Ywh9x^#ZsxC54{y!@~)i
z18xqd_Q5qQE)*LvT?a#`?pPt*-7ROkz@+Bf?LMUqOQ1W*>-!Fn*jo1}TB8pp?;^Y_
z4TCSzcsiQ=b7&pczhk6%+=?(FaCAMlkthgx2BG~q<1geICT9T|r0W#%=3Pwf@*%fA
z_%N}*cbJYO{)qa}RG52xiVWzr&UbPwFlr$pSO$ZAT<juVs!8qW-C@_<K8b;2z@OL5
z#WmaoZRTTPbEjCFPTbi3hVR@gZYRP+7eKt2(R<NOm_?I??^Q@BG<G0?TxTak=c9BK
z5Q0s>1)8<44%bnMV3Db_GE}x4@%%*Kw?-824n14p=CV+yEP3{t{#j|%Qhl^?a)%Y_
zfp(d@S%ysz*nW7#WEi3=^wiG}o3&II#v1Y#d=iL30h^F7H9yl`Z+$}eC4$?6SwF0}
zY=#Wvu_3=S#1M7GEM{(Zo8JSfJ0Im_Jkt4M%#)ig=zE+B;ShJlt<CF}y@dKO3kNM@
zEl)(12UGoKj96+y9oQVex}d{`fSuiN?$^A+;{|gNNA>A8e`S;Y@B;^sZ<$<rsNZAn
z9bXtV1Mkp2Pt}Q9qDNhkW*IHt71sSA3<%*f<_#ERj@5JG+2by{1&Fo~cwM7J!vBfS
z><brZ$*FL7#1M81sW~Cx9)=v!3!NR9-s)~&i@PWI16)v<(c7K|%Yd_Hm%RI9Y4Cow
zJpDPrPoGZLWUqE@LHfd}%W6TQ*$&EJYuq$Pc+-H6NE}B0wVjZ>O%xiCR={5yQ7}0%
zxw2l&AaG%Ngu19AhueiW>o=|L`#c2vUiqi}Le^xfM#<H{OR}{?S?^9D?xZ~_AF|@j
zp*;fWoUr`qPt98q#lb5ibltl9GXqH4>VJ>xR4yx$E`o6eDaH^qBG+S=_RXM%Gbr|4
z+jmK=&bl>h;%!DKex~03NZ|7+gdy(&GD2BzKm#&}hel0-80YDe<Rhk?sbTtzUK7R(
zI?vJvhI4*rh}<_6$+(Gw2CG*d$L)=+(rM98U#{b=Zgy_P#83fb@iznEy}^itQ%`Jq
zOU3^9kBGS<+kB`KnG@MtAl(lM_2(gd>9kTo2oSY3n&1{4+`Xy&`%_O7>Qe9k6T`I7
zd#go6&lgLo)21cY?tVIW>%_z05!V3t9c4>$+T`UyU7Q9Q_VV{LUbhRdj2kAa=k`Uv
z%IK;E%r(bDqE5kdjN|u{vuPW(GYUpfY+&I2EC9|!d*+{6<#`v*xAPnWzv{~7VRBh;
z*#Oh4HQ4Z<4niY%LvuRA7pDx|B(up@)OmO@0SW%b`%4VRuglogLRCeL+wZob=vLj&
z5VCQ?!45ciqj}kdS&OQSesNv5`MsZk=0t*_mVN|v5Xah!=;@FHPJ2$F9b4)g8Z-0A
zU2jjrMxHB~rmlMf<%UyVRC{R=P}*9~Q+hk>ec?aBCpzIKR=)EkJlFp`fS2Ar-H0gv
z+;BYSbyD7-Ii1#Bc>Tn!Uhx*$ui5fLl7EX2QSGjM&g$CM{P2;{*JN+7nkR0?=a6Ja
zya<;YzGZw{^wKMH9_^=DZ-eEF%LodjR$Mf%$n<=>gtVyHj8_Gvm|t~wdHt_op^ATH
zQ*?0E6}veW=>0>`p?#?Yt;!|e_|N$~)dbJWm+O>kv|$-^JHPGMtaU-g@!yMxCR7fb
z`5lBao_1tKLDMf69z3l|oh=F95B>pow=pf#ZWUl7q*gH^ZGDRNXVIx}cK@?|uQsG3
zUF2)_aeDJRo~OWpby<F>fJy*R2j3n=TA~l?Jr`l(beyPPd|o>J2bDLQ%+|5ZD<rlN
z$!6VWMC+H^ORwFYC@YCLxgiY;0MW@boaZD9S@oC{+<D7y3O(mz`mR|4TB^t$3_QW=
zqgNGowJ*ZrlbLVIzGGAz*$t?jbx`G#D(uSR(=dWI*vcfEmoJWaG+w0vy2K@fyIU`>
z@%|xG8y=XJQA?v=t3IzB{eOdAF;Q#^Y2IZRhV7eDY^wixPh_Be#s+d6Q{(gfRRUd*
zfrsjOJ*~ncrlc+8u5nw*upBhBgjf`YGr+W;2K{A}zYIk(UUdM<(IOK*ORjw#wANuJ
z6``&0Pf~AS1p>a0+-QQ_@`luq{<D*<<R9DCSUq_jD2gD@Q$%YqT=-h5>s4>f7U>xC
zS_6?z?$_;sSb^J7jP7_JAyu~vY2!LczVB3v%agg(bx@Rm^QB>H{`yDMP{*Pjr?G7D
z4KAO<y2KEFiFL#{m80SG=etPe_p!SGHHy~f$}j5e4a%?}C?ql`_UMU+hs3o?-}tmo
znbYl$n<ts$l;?R8@NAD_PXGx1=Pn^`$~k6tDl`!9+WJ?M_a>uEF{61YfR%UQ+Yj>)
zq?%q3*aGcuvxT?p=R0g7Ch_Z3`;UzbZi8TNPQDiFq;>EfeOY9P*Ci(X(?@Xs-Z|$F
z;+_*>a8ulBaULg6f)17g6d>Rms~psy-Ma!UV_bo2N(3G@(2|a3op#0zI?PyeI(?D1
z9!TQtw403H>p#bGCkcg72q68!J$w5`x1k~A+kHQ9AcP1mfLd)3A5sb;qPhh4ABC8;
zTOEj3b!%k1>c1>7-S{Z;y)^iI24r$J+rdA4Ts58wzjo254DTm}q<r^%y@yBrcu%7I
zKEUm!nB9t%@D9oPvfOVY7@oL&CI7NM>QAlqjE5^&$hPL&?y}9y)<GZ7UKbVNf51w@
zqm>oyaa)U+unU4vPgH=Xk-SHv!H>$l%4hay6d_oFm?1adQI>X_)N&9&eai<`jC|JW
zlo+b_*VXW<<S~fW=wJ|lcuZNQrv0#^_q#UP=u@{;%0jwgoaCqf>y*dqdG(1NH>BJw
zLfq|UdDRWyTUN(XDw+8;^~~PvYo=T~g|K1lzQu$!$c2Xv0UyBsT{`)v_<yE6R0!b5
zQ~r|s)jj)doAm0yJkC@YVExi9MrUt3x6D(#-<(%?_d>e=n{aa;VunlJLG=~lIXU3<
z#}HXQUZ&qOdQVa7*oK7D;Ctr#Ucu!#yOt9lm|L`#b%nQF4465UBmqqe!pZJfpDc|X
zrhU#b<8rrqaR-Pk%X_A`o-2|q>=Hv|Ho4bwLC&Sal91e;SxzPYZNa1P3>DEE21*Bc
zMDW5Sd!&CPhx&2q7eFVSyj_{9J~fWv$8ab8$RBa8d57ajd-)fuI7PA622~hbmxA#i
z34oLy|FV^sFNA!?oko2He=VxC^jnLw7Q1<(R-&Cw6T~O4BSPnRaj(zhD|`=Qwa+Qr
z?QZd76?Ijt6>_C#f3Z2qoOA>gXR~X=4KV$0fb{__ZWs~b_U?%G!~rg^Fe&OsHtKs4
zyAPMdAC&Nl$<Ka^ft~$I>#<Cko(pESZ9wk-KWa%41^3RAdi>86x&jmd0{EA}zwM}!
z>j?eDp6GTPeqi!vivzEc0g;wHUW0rq_gBQ{e2!3p{H@<v603dwoQFM!uh!{4M_4WY
zPYag0pQ~_VXVkZQC*kt%(8zP%x!jGdifE-zr*;0J!u`q~0eSw(bIuJV;2$hZ9VeUU
zUwVtrZOZg$uLU!oGl%#HZ<8~T`s|}B^wbbiJ#ZAzpeZv-$vDYtd@=qu^zn)JAiKKL
z_dx1R4TPFG!c&abD8<`cz$DHW{0mq|?N_+11{PW*NlbemUY?Cf#0>jHa6%D`;m_+}
zdYFJp>dJK+3({&wnTxhreVa7$4u@})<&}5>{<!B~lAh*#?=fCsZw@M*Fe~}H7O)y$
zwIWb};&)^jRY>1+i-#P`5aLI-e-x9np~g!TCBi2{YOIm$%7V@-@y}Z8N1lk2=V=I0
z5j_N-pU`#hk(`u={5pkdNZy;dY9!QZ4{z7?j8`=l10?fm$@jX4Sf^@mPPKBFYP1{}
z5}X6|TZKFk^13{ey(G+JEt7ZJuzTQ4;HKSuI_C`k(H>W9mgDdZ5eGB7!XGSd-Cw$#
zKnfA})l4Z@%&KQwpvx*LDCCk2g-W}pNFV8j0Nem$QIWZb`BdTk&78=E;(D0CPan@d
z?|xSCLwBvThdb@ZW<>FkLxJo`QLw+|$`hmHsNph=V$OXUUaO-ZF?9I$pNlv0{D3ZM
zDPribpVDu(2vP*`pzgGl*%n1aOhTWcr?!;Jb<WmWy2lC3Ns}D@@p_Vqvbnk-I|Vw*
z%UrV*bfNROu#aTo`XWC{v2;}m``2F!p4fsQcd_UI3&8rpV7AuFl)<1I9W+l#auLB`
z#qSzV$`^%vty<I`w52Ux^F2j~&(do*v9>91lSajXV4>1oV#iHKcq}$WuP|m4De#1|
z8k(oe_u+A={vAjzO~=q~8}8%Xn>|lxFz<Zm0K;(8%l*0INBb&I^`3$bA|F(#-XFP2
zbCar9mtmf%zf7P{3Q4%zM$QWj#}{E<oU@H^@lXC10Iw*8Q|c*>k^wMxFZ$gLJ5!K5
zEdYZ@%SiXqqAIubz!TbT+09qGvHZ4&#*TYN-qv~r{c*|7H<#Y3>g<}HLw)FPsa|gm
zg<Q~8=BOVjiRsGCh#ALG-WVl{?P5!(MDWp<&l9QMBfb_#bpGML{%6WzjshLf0pFU<
z4gclYNKk(UFqub4Rmeixx3g0xsAC_yTAuoHS&(X~*zDLB<x)?mCF}l~2sWn09E>Vf
z!iT#a&W$GqIx;q?qWxKDR7S{J;gqWM=#L1s*^AAsKSXO&_@V^aKNLGlaHHs`**{Nr
zD>kvOqut4_QNkeFn~E<muu)$&sFmS7KVha1^9NV2z89N;yAYllqOaA3ZBtGc|8yaq
z61Zo9S?=j!z34y>7Fl(2yzNIUyds*zYlvOD2-$neH$_qg$8W0N;b|B#6(1Qy6NaTr
zbH<f=3rNTt%=Wk%p391TXE?9f0@dKO@ncJv1B#B93oBU6sIm5rc9M2KyJI8OxTAb)
zisp8aP^fgP6GJ-WzBvZ?;zc8!-o`B{t#Xw|=UW>OjB+2EglY_P^du6$TKhuoo(R<l
zq0*&3U`f4Q_@M4iueI)ii2Fjki$ndl+1gPNEGIXq)LR3***pGwd{(BsObJ|JpBt5j
zsR(I;NAlt#>z4m%p{6RM*iMWRemKz@_s}ZrxHUtwPV=+%8l3M!78%<MiZZQTF+sGE
z@i}}ES&V!;8*fiVe?k6X$LAq@IRZfrG8>f|lmJUv)YX+8^_}E4$&-~9d=2>CWbP;D
zq40TBbjb!&ly$g)`Z~cO(m}&s>EStD$qw(icXB*Clg&DV#0Fh2>VTJ5qV8lM$-|8@
zvWT`NeMaWcy%ivn5mARS5k<O&0UZ9Y&+_UWWL=>ua3832tk)~W=5ykPk}tg)eoLs_
zA~fYT{Ac{!{*g^TbL&*og|Vc}SNo8k)SX_or%M*W^soL8R1g~8vEkeZDnbw!@r1&!
zkRIgTuj$!tBcZzLX7=RI_Swhe-gda13J{fg@$;iEfWmLlM%h5dwA>Kx5Jp<T%&oA)
zaw1juOw_z$X;N=K`UpqtO68%-g;HHp`_OMmENEU9xO%M+3ywXJ%?44cRt?9&D_-?K
zV_!-oUW(xoIz2e04^(nKtj`bk+xI0weY3Z4iJwYHpn>A~UlfXRPTXr@yT$ksY6<dL
z^b<-y%RzlhC`lohG<aKhKKAC54JKcbSgC4fns3MA&OYyeR%hOuT9TWGQ<zus2CvZe
zYMVb?mM;Qb`h9siA3f?VFOutVkNK?&4cz3m2EpC1cLG}{T1>$SlJpzY(BB_TeUdD`
z<iW>q4QwvSS{Dg1ue1@E9@eQF+EVFV-0k7r>;%vnSLXJcbzN;W(4K10F?Jv$c}eB0
z0i!&@3~dO%ok)BCp@eQ>UZw&?;&Y}0NH+Rcb?&&^R440u>;7UtJaojM9VQm<-qscn
z<9Qe&o>#=}bbN;Ce@H=FP)lyNAMahWDtTyDr;}Kv_nTt`<$}7B1sQdIVFg$1p`>~5
z?mywQ(w~m;f4XV6#e-Sd@_Um$-W_#3USz(Hn|w+ShU7YnOz||mPJNr>0Sp2^f1#G~
zvc!pO+{CV*>c~taw$7ppuB}rCvz7J?{vGQ2$S-@%>+sFu>~uFQmsOd(7v`ePJ<^-e
z^m+jOEy~Y+!+bLf*e}SYf&}fHbwmu)XkR{In)ozTxD5kd!5?t#w>1%h4ZEZcSl3b~
z@s7F&zw~kDe-e4I(%G8Xb#5a*e%6(V-xMoT&W#sA-xVlsn4G#NHPbl%eZtqXn^F#S
z=f$E;<ezbN6Q@uZdqSY-X?M240Ed(!U41p_3rP`A1`%GdxZpeJDJz0!Cr~I{OPxAL
zL%W!t<=QJ48Wwc21SlDkyWiC~f44UpU$m3vfZR?hYWa|dHV7^Yy}ozos#eMcqE5DQ
ze+rg%yr-?_jPT|*<kmb$K<+`)tC}*r-bq+L+YzMTVNtuxLn^KY5L5`E`Q4Ed5Z$%+
z7jtWE7~-wmP`*N{6Ge(^;gM^fmrwg}cEYl<z*!>G)Hq8W3edtf8C}}k6cqOIj5;?H
zPD%7VN*^xkRq(<&2vt&yhUOZf5)cF4;${L%;}&}|l@QD7^`$gb`iCzN^E?R%GVFpX
z*uNeEUyD0BYSay&!b8bu?%P>rQ1e?<Qo5tPDS~Ax=~aRSPYn~Pg<xYpI=sDN^IE0!
z?M#yh7Q3QNr)MZ=E>A@aULG@kzmuF^VWn(<olc~Hy8E(7z0c~ff6x6rW?rgLrnZZN
ze<2Of581r<QtbT~L&t{5s()u-`t8|z7vemFUx&d4)Upgb(>kbFyCz!44f#UM`w#vq
z(Hya&jz4j|XaK03_1Wj`2SXO|1*mjEoC9-+k_3;CJoZj}R31AdJ~B4;W~xd0BPJdF
z*?yIfqs9=iS0RdPY)wnM$h*W@tHqjN1<W9}>y{dzE%R}K9s|EB6IgC5aDCRd+3|OG
zm0a=OqhJ%^+t_-;70c~Dm$ZM@Kbc_hiyRFV4B%jC;|-73i|W<&Eb>e0WYr|^&VjNR
zp)3`%H^aNFAYZFJQ`bl8d-H$I8j>^xJ)D+7#VmH!9@Ccbk`b41%)YiRn5;zEdUQ6~
z(pcxI)?Q0!%by_ORXO_ftS1?lO7T8@e-01>5hrlLA{wF-k=g=!wUOBZOeT6YXMdZJ
zop*nUY|v#Y4Dh?SQmmaccoYAUzW>Xhfn~auTexRx&m+GeIvcssV&<3n!^~bfUhe^$
z_kwdotIyQ1c+CVKZeD!aws=zF^|F@#0_+8SXMqCzBDuNmfvm2n;N%m6X?K&e0q9jM
z1!C=x-#*9YzCpzjgFj~M4r(0`pz6wznfgrrnPUb##AqY-ykBCmq?MG{(7q;v`#<hP
z;5E&|YM8BSuvF~*fJfzb5tY0<N|H7)?#qnbjdrEcMB^su9f%D<9qxZY9>3ly>qu6S
z^s_*HCEkobJ=>>ZdW}HD77*6|yV}ISkkfc-VXM#n0ch|6%>C?o7yf(_76l9v0twgC
z4`z~O&O1NUQq1Q*^@~a(n60>5&?r7NRphVl4D-rvg^Sb^6_}A8Tmq5LLHfL7esC|H
zzzvd9{$t#n(=!s5{p$Dgl;hbpzEidPA&0vq1O(zkE6ra)9{4G=r{oYLGg|GbU4I34
zOh`Vwu8)|zXZ?>kTwe}Q=R!lq72aO>jn_3@X)Hm-xz{%(AQ!?<!sGOi$J1zQ6wK-z
zmOyu;YVi||4~YhC1xyCwdFgY5*%}9#W#53<lNZ*ncJpsq(e0fGr`c~k{@Z=6(f3!I
z_#kODRCl;zH}hz_<)KoupN4J8C!Lw%Gk;|;Gc07?+Y1Y~-=S<;H~N*cvTu<#OU(9j
zf%Tt}oeynG;Jbo%nnf(W!ib-yQ_ah0lZ5hh(w8Cu$cfB19-oset=Qu~4jCHhQ+oiI
zk1{~M=}{1c&We@#LYq+*CJr=iToK2&FQ-p(S%HORs@Ou_2(dIV-S^b?y}?Ug)VPra
z&Y=wbTZc!7<YJPzUjtz%S^lZ3pGV}pduC-Ga*IGV*xwJX&3z<T@9s031m%(f4>L=1
zF(*V-ZQD|PF7G{e9Xwjb6o0l_{hK{HA*fw0o~m;j0gp{%t!)H^OgNY+{wa4&nhp(D
z-lSjBe`_}W^wkhz$I^4f9@ebY-NLH{amnj=mr*$(5~LKQq{y%MJ4}<jpe`eLps18Z
z(kua2l)RdKU3P{c$r{Lc${ydI?l)tfL){GaQZo?S()=yXbz5$%=si5RnIquoj)|>+
zb~Ai5mId$SY~?fO@>a2hHEeX-byUzerCoPCf6mVNXG%36ulDsqIBot*xa3T=)_e;`
z-n_=qs06H<KyS3aOvU|oNX?uQqkgIz?Yj!_D}T5=6-{fKd&|dJ?$V=?L}Az&w3_(Z
z6voSN-L=D=5kX-d^am7zMs42^F{rE0ipBW_c^<8cIqc*x>45n2zloZo$(vDD9txkF
z*iY{vQHK<=p7+_=Qp~Fr*d@-p+kfC$y~}k<c~+O|5G*^7q2Z=wk-NJ4U;){CNY`S}
z^Kp0`EL#n%g2LhP9^M37!ynnENbNb6-=1?eBhFg=pvt=b9+579xESawf6Ag+{MUKZ
zB7G(o_MCSQK?q`4in4Dg`-?ZfJlyLC|84{f9u?NXB~B85RN!xOjhG2p-oNxvJkO>?
z9HH76jL2wKX`v7bWSm6oVs6G=Gi^)xF7{}(lLY{LS<!ZnWx^B*v11FYL<TXo#Z{kk
zeb)dF+uw;_<tkKJ92At-Yov_L%AOS2ldF#vVf=AiWUq*p&_pF#l0$M7Z}-(M+OWg!
zsWBN(<$=E;!Z~QeA`0t5nI!X7s5t@WizFQ)Z<!jeReC%mI)6|gsH;G?qD1bg%SFa-
zgQ3QDa|2CRSdX>GnbBOz+1T#tK+0KnpW`cpR6^hm@RaRUmz0_YeQX7;9#`6qn0O>U
zHYDu#+{_=fw}_y__rFRc>Gh(zg&?E9fy%^8lCmH(<5#R1OS&we-w7*p{dCN&g4fR?
zCaLT$vy~%p%|k-}&>8GK?k~r`_s#^l#ZBYB7u5NCPqK3<l45%}G;8{Jw~+`O{tU})
z4zG0FwKezeQF?xQNNuPch`$+)pIzo6^Ro3I4)e$~8Ou{|AuZR_yrsdgyu0J}y^f51
zFpnNlSR6*UpU&oeRMhW@L(LC48vZ+n%VdV?Cw}wgb3j!QcjLwfoFts;U0Ag;qjRYq
zOTo8h3V@3M?CD`!I=C$AUa6}yP`T5<EI}pMYO+-IYo%LGovhf^|Ehk~KDpc%gJ)3E
z99INC*Xnmg1{x_ZK&Gx*67RY7b-m<RIuJ;WeBCfFyrMSdw!rGD>sPuzx2zrnkP7fK
z&l!cW5Bo#?oA2B7Q9-NM$x|@D({hs|QlRXQ7p4ATK7~|V62UM-tzV#`bjx8insq-)
zFjxnel3<HLWm?{Q<RgQzB#?TD0dZSHMX_Pes~HstPTof>I>k>j%z{p1ny>5ld_@Jx
ztNN%$N+C4AuZI{d^8k)oa~3><{nKJr7*wEm872j9LlF(Ibv3ce6R{r8w08v(e!*lb
z$%RZ6jM<)TuWIu>B6F?dtUQ%ZV^V>GO^cOYVyFxQ>rOQp_Z8ts;kpnGk|n;ZpQ&OK
z7*ctNv22PE`X7J#i`wnaRSfL~VYwKV2uNrfDLLaV^vQLdu1qtR#TiUutV(>i+ch$|
zCld04@wgNY^?+tZV0b?W2#$XG$8dU2#^#C(FGZCV2eT<?VD&u_x$EC(b|)vJ1IKEN
z^zM9ct{4Oe8R|_PyYbdvZ@A=w5Js<{1|!M?={j^`a>r>4{%otsp!kJ!QaX!(oKHc4
z_p<=BR&50pa-%jlGc(kW{moQqiH1Gkdz{kVH-R@fK=Pl)f^PS|T1E-%6fG|~S5Z4X
z%N<W}5}^;pGrr%s&<MaECU1MMm{}tiUO8B~;S1s2$Fbv}10CNgl%DzKeIC>oKv!we
zj3UG<^$+c2$s|aA#_HP)+q?c8geB~CMFdwIV6ivS)KZ$A<nIaM@qDa|MDOeOdiX|#
z;FzuU(jnz_Z5&As>UwljE0HjsAq*Nx+8N62utg5)UNTTmBzTtc4V!5{G$dQYhIYRD
zaG-CSy3OUPuza-NOW60l+-=&G_w_4pK72qdQK8}i?qo9YtCM){9H?216!qtyh(h#h
zYE;j(KMHBVdDNd8QB@7)*{vEO7a^zC`lD|Bt|ANxh7$05d}T2zU0^8hlOyg2e4Oq$
za&}en?X-OCjlv_320zSNr4^*fjhaFEX3yNF4liqz$g4<TfYCSL`>%~EbxX*G{a+;K
zQ6T{eYZ%wflO-Ci=bP-VTK@2p{9XlVovD1nwL^pTG_jxmA5&i)Rps}5EhPv@Nq2*E
zcXuP*EhXLEB^`=%cY}a*$ECZZyStnBxqiOCwO;>wSa)5YJ7><!p1tSHxnrn=mVooM
zj16ppmh>f|#mFk2;@BL8ok??&m^gi2RmRwl<X^+gItMLY3A;ZYW?&B^tKokc9fbs@
z(1YaQ7oYqwiMjUaBX`+hL5T^P(ECmCygFzIsa>;Yz*#o^4i8EEn+JP~12Nt9mv=8k
zbTQ0hz_D!qhr3TSJ?532(^#mQEQM7TXuuV-@8xwOK#Ey;sTvM!ReGo@Nw1=<FR|h%
zORgVFDOBsr!95S;PyuJu3)VA7oV~(!u_jA%7M8|UkA)dtM$=Q==u#R>_C;z+<9#*L
zdlx6MLH4okueDa*>g8*{KRG2;b<t-%Rga`eLjNIPZDeeG&DTPJiG91&`VlNKdUU(S
zHaQ(R!eRz}pAon=SU~<U0ML$%CUbbJ8`%0<&uBi?M}Jd_`nWsW{}M?OEt^x4l&8M~
zG)(8cWQCn;R2`B|RN5AFS&bPE!x${T<yF|muUgIZK{(z>htX6fH@smI28$UoME@}D
zuRJ<DQpbe|x~Wg{_+OibAs2WlH<l5&(EycMObD0ahuac&+;L@AU7O+G>{USe-T098
zCG8CMv1<}gmEkWF|5KGAuU>^E?gj8{dnW`fuITRtyn-BKCS{o;74A102n6;IPKX<C
zfHr&Kq{O6}CP}U8u1w@G)Th3$E>%Cmgd}VTqT85><KhKj-#dQ`^(AP>Y`A#jH-HWf
zHu`$Uq20i41qnI&JM837kGaEZGxDz8ax`$RWQK<e<!nd3cG70XT1>3%J>L1lfDX4&
zKSVt#natPiH*V36=zd?j_b>PS@2~fBUGPLQLvwtpvMx4yKw^`^HNZ%YNDU^xKJDO8
z{%P&~6}N3@o;!jFpae0Qymshze_N{K9>@X^UW?m%JfN-EBf`O6`%2r&Uc%Yv_)TZ9
zO${w;n?$plf$<j6iufnA^7=Aa1EI;1xZtuET0jGlz=0Ye?#25ROrkuu+qY}wINPFI
zTFV+>=-D-;D(=tUJ-{5c+1ZEov1K2qk(af?KvtdvrH*w);Ft>_hUe&(;SGxAqQ;FY
z=92vQ2tM2Ql@!2ggV0|%l7PxNA<X~!y$qpmbH#({Wzom^uI0P1)6fG<kmvuYzc(vG
z*<uh*N@t^CZ%p1tHp}k#5c*t(8+s6c6uj}R;-vl!ECHjOjPLh{h_{tiVj8Z#m=--V
zg9I+Ru7c4XsF0weh9ebxk&10^Ui=MzIy;HY@49R@-aSUHwm<m;XSgFpevAh&-OT@N
zit>NXVm*EfirS_gSzB7W!Hc=fq^L=XIuDQkxa7L=oOo0RhvF1}*IA=t4}hY~8`)VJ
z<SeepzA_HDpmrWsD2i<-qNmYc3%xP&$}*K`RWCy(0{6vxe^sw9YS+9Tq<@k`WQuYU
z(f&QG1*WSZ>iXcaNA7}vGm(eS_h<xddiABixdJ`isslYTd9LgF_a?nqmpCk&X|zR@
zs7zm=+r7m|p<qOiWWf?C9`tI5U+f*xjaNnf=yzb`x*y?12$9El?6>V-#o@{ElCKMt
zAsd<Fsw*+vk6ORIk@Vg625E%{SzBBjw#gNJ<v&p25{9*M&uIN+^$P(=*pjBpSs0dh
zQn>1(dCAKG4$N@!Du;u=(93rS5@Op{*SMU>Yu8!+H3)Mge99Ppr0rfd5)k&WRq3QL
zn+!-7YU2|-_~@K0G)9eP7THr4mv-SMuP;x?mNQRDXReE_1>m_ZgB?f;&U@gwKnEFL
zv(F~EC3Mc<O_@EmiCtcc6tm7J#)-6kf0gsAjLj0*WS=CEo{aTB^wE8fFdwV1>4fL8
zX!7=5Li0{#R1<G})$5b5qJfcIBn$IlHOM!Xf@&hQb+2vZs*hUH;=Rg|+?vNn(bS(J
z?aPyA2On%0k0=lW?ug!cTHMai)V1cotxCYDpJ0I1E<%Y@js~}eu|dnp%U=*h4KHz7
zN|ATW!gM4hx4#Ko%p?gZuoO~BEL+QUrQbzdWb$8VWPT7z6!B0>9@6i43nsLx%uxvm
zEZ$G&3_bj(!L`10ZP1L3LAhm;(p~cBvi*7Pw;*Rth<VpkwL4PMCQEuSG}UgJx!iQl
z0EKgIEcy)Lv7bdYl-^0DctJ{vrjqm$^iA)((^IrB!}7xQz;h$2`_(N!-FHRZ2qV%s
zS`AzHnPg_n#}o$h<)|i24kuDMr^51@3h5QxK-R3YOImI0W0B_<h}f)bw|S&ck1Z7s
zRXOvYv<<KC)=QRMn!k*nq4(xyG4y{!4B~q`{|p>Fh!6}137Fr~F&DBq9GM1pc*GiX
zd?v`FW5)UMM-Nq%w^{!we3Im-Kem%<k=)d_a&Z*j&qho1pMIZ@2=4h};gSKj-3g(g
zR@c{yTjXE3<5Ho9e&zI%cIIxI(-(EytY$n)z9uU79oNr0Nr=Q2;d_vmSun(bnKi{}
zdWqpeF54p=Yxa9T`#HE_J7bgm0GmdUY;`qJ<NMha?Shs2$<d$qd1*=F;7lA}*C=uL
z?eSPp&Uln5k94MFnWtK{*%0wFuNxZ@V_I;Es9&egn_>B?1J@vrX-_g7Fd`Z9&qKoL
z1-44wh_p2HB^I^+dR)=xk||=BHg}vD;9z*3L4!i^e!=E%#pnRrH7__q_B3NqJ6STX
z2b=Hy1P^;NvwKEYj$(XPyH!4bg=;)658GZvQqpqafHpYE)jQB+tvc#Y+OGtncU@`1
zn(pTuyRIdOKgy-ol(vX4!wHgJn1R6p^*stvV_zjAh|hyphj;d5BZNLDf?Y9Oi1=U)
zC465@{l8&1H6-f?2jOMyii3DnOe}tduHhAxBe4VldYmFC_m}kGCN{IH@eKkSz<3Qy
zu%I(R5O(f#xMC%&c=pOvq*&q)96y%}u?$l6+`+O2qMB206!`5ihNzcxk2`5}$jlB@
zpNf}uHW{?D*PK_3nip+2eY6(^iaow$k9{M=%9s0o;^8gH^-mWy`WGKXQ&^ZAEI@0E
zOjbW{X(F#U&+^nm_Jt}#xu<3bX0sXAEL<l32&+Bw`F!^F4ywD{5x+#?a(f|^@k*4<
zMElD#i$bl}%x3l`z&?A87qMB<g4`)cjV-B=Av%hgVUd(Dl+0tiiHI-Y@P1kXS^oBm
zIfHh1&Q@WLQN_1=pAn|6k}T_j1J$m-!<-XsZfnjwk26FZg24CVyoqQZlHvFv^9~M-
z2|Fn+e-!?4;05#v{-4C~=~?-i68B|XJm>f#lDd_oRJ`&@WhJ**V~XKbp}4<Ww82uC
z_)|LrndLniv?dFU9-!oWzDAV(Zvqmaby-flC0Go%gL4-)-1guPSc(XYKBFx2Z%~d=
zGAv~J(8-sDfKuSn_W$w!?iaHd&S*yQXA{6ls-FXPp^0BdI7cyGZsI7W-z*Py3tt&m
zJJbKbd2w%Jc2WFO70Cao3L>o2C%VX<A+Gu`&YlTLgW~JhN_ZHqP9X&arjBYja!-^7
zoZ@Fh&BlNr`q50zz>}Dh=-z}0=j_SNI-)sIRS+IZG;`w<{TN2|pZlpE=EK)Ngxs+Q
zwR<=>_|sS0-{#806*71>|3)TN5B#_$Fb|Cp4kxw6o+&HyOs=B-H4s9}`DNknhdy{K
z*m_kB88PEg!KDypzs>If_gZ;bV#Q3UrHnmf_Pd*=*5!Ij7P~*EJ~_C$hI&-;H43RR
z>5tKb#r!-KJyaq2QfEkbU7{oRQIZ^{zT@}_k%42EIzf;#0(~HQ6PU&^Ty^gglD$`@
z6muF`51fn<Z14Sp&~728`+8hVMsb7EC;|~zyT=xkrn9-3RIvIz|KD_|JYr(O71p7Q
zUsu3n9VW=ON2Jc~@E%~qAI{ytGy~TC7jp7V0(BHv0lXP2zuP)<g=bN>?&^G#5^Nni
z#dvc;uT8HOO~W;M8{~>_wwTy|zdTXgq@T%hfhqH>#y!Fz)Rou}OLbJf`5`*m)L#j8
zm_z*1C0Dt%Is4oex-uoKaQtees$F4kFGNULEupwOGPu3T?bpQk))n>vuAiUT93Wg#
z02Z<uE4kVN%HzOPq*fcF&%QZ2^CK3o4&u?=p)B&a#Uro70j54Gqr=VNY_yp|KL_gW
zrR{7f#x5}N(*X9>!g^X|95kE<eg7rcG8FT*o-kvH1B>t`{Y|f9!`{KM6PVvfGmGZE
z`hPq$o)v(~RaVA&$y-u^QC^nY;T}NynAWrz6Duq0WeTO8C%@NSSK!T6jTr9r57@7K
z75oJBU16jo8&x&Hz}`g-SK$mam_+Lv4`z?WE0M_VBEP*0*{C%~9R6=jDxJD7*O9=P
zHn?c_$kaaa(Vw^jsxow3m+w}*aL?S-Xu2q8Wx$ElADl?veOC?vCsID>#d@o-a1)7*
z&Xb~VTTYj3Lvsc`F;)^aDzlDz(&V?O{Y2rvDIyg{mLp#`a#w%)+|Ax3_ReLRUc4Ra
z!MLk(WAQh(=zKmsZRd(?X`w#_{xuJ5A|UknIJmh*w6vVO?d9jUDwr2?*c4^x3OrtJ
za(i5D@9r*Z4tAHS)PfHGellLWNh)S2f7TK_@F<)c`Y!ce)~^=EB|jEr!RLni)DJP-
z^XBlOIK9mriG*Q>q$f0Qt+Rb0i<L|y)R_-UFrG}JGj}6{*?rgXYt^KZg4RzdR3<lj
zFLK4>KTLnF&T$Ut=2!Qo^`Gw6{N8fLX9_O0oqDZC^!NI7J05RO!tevzv$a+ZrLfmn
zL3K?m!u$-|<FE}Ei_7hM?*{7663+=`lq5rAGkK_n+7rj@Ju1od(>0L_Cj5Z0T7_>-
zh>|gcYAE}8Fwoj{*5M8&5}OO7MP^Ma^yt%*t3sXkOKhW+QzEzqi}I{EV+O5HQ73z6
zdETy6qUE6VE~{A6;8mLqR4kpW40JX%N!zy0O%59yz4}T;vTW1|IQMXW^?k1o@nhhJ
zdrrulV3@;e?klARPg83Q-Q$bg(5o}bAfFe9+~A(P$pXHp+Ox9~`|DfUjq*9MG)BjG
zlQmwxr_$~l8@@|LPy{-rY?>Kp7}?<BeAN7_9JZS2tRx<ZkaoT;1822+A8zo6PyG`0
zi;C>CrD{E`O|_GL{f>xsSgn|Jt&x{!A&!A>dPIC(MT`dVQ=cKagI{k-Iob@Yj`@98
zhYwvxiB3%kB}bUNSHJasPkLAXZPu}O1Ga>jN<I}mgU3ar2dS;v=0W$IpQKT?m8<4r
zXB7syC!O6D?iT-A8+D~!3+lW>$(CTvl+R`092id68z^B{XEl%XSvlWc3elTo|2ry~
z@zAxF#dI^b>m5yaF={!U(|x|J<S3q?rx;(x&P9V|1dOF~`HrO$9l^bUbO76h62BV@
zd8CO4Ncsn<W#^mrGcv^KtyS$k7bQHr;3?g7E?`y*Xh>WL5NiMeFagC27b4of0R@;I
zL;Z+|9a!FTw|N7yTcyckg%}S`aNF4Ma(_IHOtP1Ix#PgyxymG?XQZu-BPslHoJMn1
z#pg@)bT6-Ysv<W|uUe4z5v~;@CX{xZKk~HgDNl@P=Ut)NVQI&_;pw3bcFQ_Z&wh~G
z`T5^r&E^XO<RE7SbmW5ss*CIq^U$1H^=#XwHNNeevB9Eo1M<rGGl8Ff2j>;jJAB?Y
zGB&$N$FDKcsy*kKpW7aZ!j>V>(5Y(dI()w;QGX!iG4FbN-Mf8;i`WA0h6HYeIG<l*
zD4aAKhmFpNc$}q8mA{hR?xUTjXd7>^TQg-Z-19A2${d=rT)|-8><9S=K%JE6xUMA8
zsFh|*4|7QU^`H>&b!ibeZtiiNX)-?(_;8qB=6<>RqvJM0eZwuY1FMc-;I7K)p^Nq}
z$gW!ux$k<}fR`|e=8Y~LT_d#>)&;LYXf%~QXma5C3mYVEKz1<T)UXAAr+jANpq-!>
z-HyL!t~6&~wmN0A|HeFm;{rwa_fC>^?i`!&dEL!=n#C{Hs>-kp9d*F_IAt{6Jm(#o
z5bj}r|HO?iH$Iv;>Cl0a9kjs~^eP^NX*1De)nM?UiwT`xn~l%+8QTQ&H3*o4_r3?!
z6JPl>s)QnB?2#YDN2$rhTkVecyHMViWY+O|uWGH@69f^AjP)KznP+2nMe>r*Z?5h;
z9_u|1f!cRAA~p<uD!R_q7g!<5W9`8Akoh^9O^z4|lrAFCm>I!A``I1u5_!EaThY;Z
z7<Ak`t<Yj@wL+{@ro3pY6j}5-d|=Dw?1TfB|Dtlb(j{j;<uiDV$a+@TOS5balw4{4
zv2KE;A@w@0YP+$Vh|4K9D8ej-3!$ipt9Os~udV*2h)#w!N9fj9+wfX8F`mkX$>GWv
zU7h38?sPCW?e;s?L%CNxoqEH_^Yli_CFU-xqpReFMYVGFU5`Z6jK2jm*y=)}#KG0>
z9|R~F1K#J^Y69V<nga3@D7L3DzK5yln+!U-f_A+HG(JXvXoG%kPwLVBL<N@T_3;i9
z6QB1zN=Fhyi1p`B#;TxD;-ps0Xn#f2`||@<N`}Aw(Xiu2YXL#pDQ7D0PRQTps|+TK
zjYJz)fjcRwkVN6lhg&CIr%YF(g~U2<R-xFArjS3N@ftluBq9mVZyVSnAFp9W?v3`3
zNHoQfY2m+plbA0<g=2wBZ@Pl4z$6eo2_0qf-YzDdwp&q(_Lm@bCq?-d7q`yM#!mx&
zEGCU(^z;2XlJGHulA$jkjWM1{hI-?x8trx9q4%Nh6Mn!aH+_@NIbKcgLkhD|gJ4Se
z)W_BYVqH?1pVSsJjm)m6f>B68l>VoL)0DDwbR+Cpb|{&1?~Wdi+=!UWM>*I|1z{9Q
zSdDH3tx(xZZ;wz7KN)U?4rCd$dAhVdoukSxx%T&f6ys^MlzYDuj`q_Z90syLs%|4^
z@!=6XeiXg>Rl#IA^;N!V2lRnnyPPY5TsDA%Y;SWUBgsAjhwvg-{Mw|Vxlw|6pTz$4
zR2I>dJGP|7)^H@>a(|*mWdbkwv_?XR4CNSQ0`sxUR_4Z^4YCV1fYG4c@U_(=na7fE
zwbe6VdC-_dX*=$oy!^&kB46c8>BBMyjzbk{!=6gx#dqQQH>nW$*paf$<!+K?4^OS6
z#m=vY9dDNe@jWk&%9e7hq4Gz&O-fNGxLaxqN7n`~DyDL=g)bMDJXcTkMEW;f6(5gG
zE}D@$b{DU(`3rHLOc2vzW|PghpUZ4#C8zToyvxIKD-}IRtBls`*9VZP@K*%f--r}e
zJwgU1bMxe%#w=EIVMxX7kNe6av4B?C;&^)p2jN||!FZ)F{m=KWA03Rvp2AJx;`D_#
z%9tprIwH(d;afJ|j#627RXRKxnjBh++^H01bB0Rt3f#BY3ob7$K6fz3S}rK-6+X-~
z6>wOi+EpSsy{J6Q#QUfiZN)EKbwrrx$za>@`Mn;}i-diMMCcam{7cKJRAVgOR7Rmv
zC|CMj;5`&{rhrexhfkD-8-Xz2v8XT|zmEAY1achuZ4#mgcylp}y#DbJBrYrV6#Ozj
zRK=DE+L}CR@BAIb8PRmK!01rEq{}<E8U4Z>I(pNSY>W6*C-3C)L_hXXM=k6yw#n&s
z>DukoRS=G1FRWddx@xtxXV3}jV*h}Z&2|P`uh~_+)&1gY>Z=JJw1D@$KhAWK$!%20
z{lyKdBV}yReY_|FZu!}F^5SL02D4|7*VVWio&#)>5#eW&2fAJUm%T}B5KG}hvTYP+
zs7<2TF;vfP?}CBTLE#I_=$4wt?wvxWzXhnw*L@at+SzLFkCckbvfl%Qw6TP1N2qEm
z4=$)+(>=fV=;wS*BVR!rnP9MctoQXPufnQlD2h>cj(FK03tD{Z>3y+Czz2bl0SjFx
zvPNhod?yS%oudV$aoTP6hJR{w+Cw^BX$fm|*z&xeP#NM;&lh;-kxy+JcQBdUU;UWP
zhGq5wiXrI95e=JYaxyoIH!eZ^q~>MI&Hq~eYgbLH5O3*5^Qwxq>e4;YcXtFQ-1Lw2
zUthIvXqTl(Ra06S8XnJqiI!R>I{np_f>D;gy8D~e7OpMIlWxAu+*p?1^+jSy{iH@s
zAeRnI=XN3mz&}gX>lH%4c$VPSdRI`R^8s3=K}XC&jY$ht^Jfp=CK%;AD2(yEZoNKc
zWbx6URNgxo@y-^9@*t;1(_ze?j0S|)FhWW$!=HMHoNly-?WhkwTa+*;;)p~0P4_a}
zt+r++(BI5bUF=?_9%T&`#T-Y$<HaK4?Ir9zD;n_~sb3!dWzCmMLBnDBD&$);JUofj
zjn2?@Gw>kjahcy*Sh8Kz9q%-D6r}sKTZu9K^zJITg|XP3%(egMex=22q0xa19-aCf
z6;<rDV83^1YZF&tiWNw^wRHU2qQrOFd+zY{^`+%pqZ$o^M)kZptxtEwH(bR;-9?kf
zobL_KZ|~b{qe+m_>Tq`mLKyT1G-9i{@_|hLmcn^v&I68^7Vn4O-b7UFtHVi*sNBE5
z<kHX=TSOX-F^Z28#iFyX9a==G%m*{a7@4#(TbHp--@4n<^uMMSUMyD~td{g(o^7vM
zg34-=#~_Z_A}pKhou=7LmI^KVot!pOSj}a`y)HKUcKC&73oc8YZq@oOq?G1ASLyBM
z0S64ER}mg;$Gi~M=<EI^CI~N4A=>K`8wqDWAWcEEh^`+3oxhj);ocCsjrKXJxPSGi
zUUEc5!;oN_B0Xlnh)q+%f2bFq+cDek?X_5=Dsv|)AiaA9$#nQvB0^rkGR8`)1hEw9
z`SXL*_HYuiUI>PJnR-RwHa4A0DqZ>A*&2wihE*Bt(1r$Z`)-a$S6v~?e{0Q$?SAEo
zM;b3w8%bsIHQk8tsQts>&<@bPDI72t0sE&`;7kgkL)Mwpv(Xba9DDb8b7b=N`ht`<
z(!l<!Ss+@-ULxCO>CY6+a%SQ~Zx*JR5p?}0`j>ydR6uln3lvMOQHlQzs3DJkK8WR-
zfMV|Vvi!Cgq_Rlzoc?`#K&QAruiIjsWe^^xEwRtz^*|ym*5U_G{L}TR|Ilh6Eu?l1
zk*>LVtSopBeu6Ag$Vk9!$iL@uJ5BeNR$XFxQLSMp2kZn*3ysOBPj}D=4B^Hk?^i33
z!6?^j;gGskb5kZWWQ4O1=#YrURaO!k9$%+sHtrQd*k~{vNoMYHaQBo0BVj0Fl+ViK
z*GKcn^Obs{fMMb}Y_t&x`LOe4{s)}{@xU*$Y^EA3`QOWI%N<YV!4CcW9F)oLWqh?e
zDwWPvlXW-~!VX}N_3(EGjm^#nfgNwJc&KC&h-g9Ox=k!|6*^etkEOgNsMT1|ulHNX
zmKQk72NqHk*YgH`mSyT*P3O*nV9OhjpwOz*9SCLcqUq`B?I0Zu{-BbjQ8G{~MzH+1
zG`wP9U$R$rP69)oPGQ5d)wY43pMNadDpVfuxEw+&NII8+f7uF|P*_ym0s;Fs)i}pr
z60c{Hv4n%jx-GC;ZePd@Vd3I*GG%grhz9FbROYbNFCO&%Ba`p5JFk$w8#wX-#n?v(
z9tMMILkQQC#gGhMH&U0w8O`5c!<~bV8d6;>0b~N??0Yzi5a0FjfYSi*(*T@39%aSZ
zfFHl*6giziJHEqa?}u%D7~iXWY<L`HK}twJN@wI_)PS)JURiNpz;-7!UYJk4Yt98S
ziEJk^;=#^=66;uyPJ3gCm`6H)tRYJ9@r8lfIOO4KZ|ml0K88t`Um{v4PDrTZ$CFJe
zi`mHck9ZtXrms;rta0=@98CyIcJ+r-;NLGnY>zsTn%dcC&`=5X6c$MDCfOtgg8N+W
zgDDCi{0YQ)eH9#iJM(*+#-zJtC1|E{4ekVXW{uC2>qrVKD8_T=rP5-N;LN1JiH89a
za8@y3Bkb6%#A7~uqL6`pTukP2$lX$CCv1}F97CBn+y0Xh7YKu}IgulZ&GHql+2ts7
z&G+e}WDI_d{7y3fD8VM`1RLxfRAt&UP?jA-1(=U#<^TrHPCSXgpk=a}qZ5n9ofY6I
z^2z~TdVwBt+t_Bh@MG~zEE+|`%I&l^Nahz^c&+*P=FV`^-0uUnX#XBk`{M;>Hk(EC
zG!7e)2=veRNU+)midh1Bya~>YDf3Wsa9r;pmrL9?x?$KXr=Z!b=EQ*N*I{P}1BBg5
zSO9)2Ym^E27em%Wpuq_+of|5#n{g;2P8^%1@&k*c)kpe&2JNJUZufZ#1xzGcre`~k
zREox7t@)+oY;zM$H7xt70`;5|x>Bzt6dr@d5QvTg@nm5g^(NyOPN;d#2UEoCA^$T0
zJ}NPcQJzF}r<d>R^Mg|H7m_%Mi>@HpE{Dc{n^K4g?BhWj`G+K6pWj7obUP#TeS2|>
zBH+c;X|V0vpUguQ@OHmSaVZC0J{y%uI2$k{unn0P8|>^7(5@BOPZTE7SXWHmB-byt
zK+5r>cR?(sL-&3`uFE28?LN##-Oxaw9*prSX(<HkYytfo`mO^;y2EL!|KkL+d!by)
zM4lvatLM#-9mdgr?$nuuIsxzF_g+(@3hYh-klPvgiCS~zS_L<Kz$T67_qy#oOa5;k
zd_^J7x=m%bsy6HjB%U_6BA^FiqU+geMWV}`(3B-oNGG8DZ08EnvjJXp{KbTmbK&}6
zI!`D7GTX%PKVuU&lA!<vUKBPnBo?E7?^1(ZJdGMn6fV0=2Nld+`25y19JCPtRRrM_
zGf$fVGpw^iYE<dNe0Gtl6{V5_!m9B8L{1w+v6GZ!*K$m$INAC>+;WGX!KG!DZW9_1
z3*j*B#bOBwJg#;VTM`5N{I7}GAbo}Wowc2z(}3ME!J-$52l?^{U_NG(^P3}$kKtfX
z`3@HhhYI<a!DmQGEF1_u)vyTU35Ua(&j1uqX0U#}zaTT@$<~<!Y{~|?Dkfz0t@Xpt
zF|d)L>7m>G-eTg(BxR0%1Dj(LKrK=~DCH+`#fa~I@#4KcoplJEhS=(pp)|ssLshC*
zXgl(9V61mZvz;i|r^@qVu90?{6VP8CY<mLPqrHMs^VgPbKb_WZDL=?AM}N;RxB<k0
z{=QVm_u;uAucc~b`5ieJwCesXzZz9k{)g_c!01ljOyo`!@SAEJNI>b67)$4_aXZtu
z)SCQbqho1fTl?Nw9PIEt#7u^r?@IQMR@=PVzhRdD-z<P?JSQMr3DJ<)xL_U^sv{4;
z!d)NEA^=oF|K-kbJxn>ya${{g8~USw(R+M?0`O1wg1(n;vYKa*%izK22}71VU2fXp
zG^-|RB+LekYIp|(h4vN2_*3xHx*oKE-zy?dMgG<ju$^RRK2T9nong*Dfb&DNKP<HC
z>2i1o0@h|#$1D7=$Ux`wG(Hc<<E4i2p2nMSiB2Vy?7;rMK5+)?N#<P4akx0itTScZ
z^NIg*D^UF^7sv$oE`XDi*=XwE*(eZAcxdoFm%-moK%9;7=*sk=83FvXhPu$<CkX7E
zU|zS=bpU8tOk@YF(ay4i7Z)kA^9PmUw$E4K_K@$Ns1&OEa)7k2oP%QxTMS!>TI~-G
zrUgAcCrb@vW+Ta%jt3T`LLHEGQ&cBm|3!lJVz`MU;giLBa4Z4?(gN8ei8jxhZ3afk
zKCxIGV2R-Pz&?InJx>qD4je-30CcpNElu2RBoy@J8*sqpb+yKuDK7)>VSEe-nu^2u
zs#qY~ZZ)j=BFr(VmTRI~%~yV$GnomRhD#<DB6~M=VYAY#rk#~mc;pld_mfTw4e-dh
zJNVe@eO4eucM7Az5%AP`hc35ze6pM_OjNBR;BoGpE>a{wN9DH$=M^S!u<R7lYbXJh
zb^$s-K8=G!ui1IZIO-Rgs+{G2uJ0@Ek6!|W8qp5lmuKz7m89`F5~%Lr4?(QvV+%e8
zF0ky7e}w#7?g2NY{P?c1Rx$DpW&On-Kp?Bb<xc<QdRQT1<$S6AwGP2b{Wcn(>hl)o
ziELr_x7RkPD1Brluz9S(=BYLpY5)hNd+dO{@vJ;sfNG-XO`wov^18JKrxVc4{y*1|
z=tK{If<=-7na6As*UBW)EMCrCi%jfwp#?nMoyo3leO4*VwwTQAa*TRqF9yQ+I`HR(
z-=>j%>^VoX|8}%68pJ757Mmp+pxuN3u~_;|%zMoH<&j+~M$%0i{6@Io1UHdPwIU1H
zYMlyl_!qtC<Mly2wG!!IBGK{z({(R+DIvl-fijt#^Ki2-G9HK|{#IOZI;H=EN&{^Z
zqgK$*SOEajBHO+NpoQVJrnf(s&)=xr*eb4@f~5c3&H#NZL2Nmdzs|c2AQ|B|x)PP5
z_dd*)Z)BM@Q_TN)Fc7pmOp2}kv7qlDv}Y4(H57QN^IR5yN(q%jB>0_$7AH8@azR2q
zFMOLRj(4_|pfZmkf4JP~$&-v-KRi5?>X|cfAqILL9w4=olRm4w2Ugo5JATQvc(zO<
z34hY-_PEBTubJb&3?GFukjzZRVY^K2mY|r%@lmVRY_=sWaPs*u2gvv^5ET;*NhuRA
z;1?Pp6{1)w*Q$-9kj4CR;C#6)y*&`$W35jB_ACo%@5oMouNFe^k*pK=h{Hk-@Rj+N
zZr)xO`;M;BoqlgGRF;ouR0>f5eEaKSaMI?~ssW@CNvDIQC$WQ9un)Ng2z3R*INYBb
zfibOKiz|%VFCt#o5r)S?Uht2G_5u6P{&)^yfCTXZsr}2N3p#_YrZy)&pZgcvo5CxJ
z|E2z4aQK40x~H#NO^y^IVMsQ^r9pCfsjgH5a*3zK$f(&J1XCw;+I4XG(UY5;(7+}{
zZqFj%_nfiAIQU=s6h>L+odaOpWVMm#Ou3c{k696!-v7Rx5gh*vCZ<gR2o2#U?5{8n
zp!N_xkIR9i%$KJ-GMmM^*QsBXs(^plTSEhP_K=;i0vIvm$B!R9fl9ezAJPSUv`>}o
zHBqiRg9spjW$Oo!%J%7xYNMX_Ef+BT{*C*8%QP4j)gPBP^2LF#`DcAH6dF}(FmO`c
z-WyHBf`LaDio~QpKi?3&f^q&2TAV`$pahO>`sr^^4ubtjg18ff%V`%emCZ5`paQxl
zF{~lpH|K!!3>b;Cgkrkq(*Rj5ns4vVBml?#bxr}wF%SWAJ6Iu^i39)^Xu6@>M-DTk
zY6#$Jt{vK2FPxzmwfYnE{`a^D^C>*8gda!#{#gPa(lat@Y?e5l?=M6*<zV9;YgvdV
z=54s(!O7Qsob98GMW@8KzrW~pa^22WzZ3F%o)gv)W`NU)3OLCX+8+*qGsYU}`NNg@
zAS0LEDqj?lpuVm2wT)hL`WfnZM2wJM`4gdz-qb6~x+MA;_crIOv;I4bY;Y|Z85#K^
z|LZhfq7xSGc)97*M_jf}pr$EvyMK7cne`P2>Q(v7o)4E4_edZD6z6gRL0|7I)x4{3
zfY)Jy!7!%(2eZjoLjLY-;h<`>QQRi3#Y#QiU}PeJ<Y;B&f3*}cbX`h;Uv}xy%zt~~
zh+f;n@BX|?^brVCaokR3;1~>`pP3<V<bCMlK>%2zGU#$CMqi4y0flkqrb{rG9kstb
z*7A^_hE$-vS~089ZE{>_^Wqfnc|=z&QoGY3BQDANHeX4fNS0bTev8!$7%Yy}TtP0C
z?E^9)U)&u6!>tO0%h@Xb5@)g13>+6VVBckC%1qt?v+>MmpeP?CFa-pf&gX}#ol`21
zkWj2L;6LZ|Kvj6Av+c|c_(Z29+;Y1Qo?4kY7zr|)jbLltg7egLVf1ztA+KwB9U{eS
z%&$&S|FP2%E$Dc&u$Xk9U=D<WJX`BXx+9o62EY>w7vdnL$-27TWit@KXDM$EfLjS6
zqTj8tH!}5QZ~{9Ad4HE6mw(FfLhlqPvkCiU3)%tc%wH@EfUm&-?z#KqEnI1TI8(CI
za1P+B>aT-nUq_Oa9{re+f@54BGJb(k-CjC4_jcBR!7M1#kKZ#|$e*~gVgW!j3F^9t
zDmjW}U@W=?0M1752NodYpcDH{Mr%s?r#IFiGGp){%tk1xs9t+~5zlrxnxocjtiJw$
zstQ^+c&P-U_dRooI2p<eiu>zxllMF}li@pe0;>xmXiA0j?|{1HS(@JZ2pH}ICIQeg
z%MLd6`~=KFV3Xrcg0n;GJQ5+FCayO1n;Bbho=8@UTp5rZeKKGUx(>&{NyAqJC=6bp
zv;Nio?yv#<(Rjl)f1RskHYABKz9C>S{EV^76OY`@_~xAXPQ*VZC5q?>l~fG8t09YB
zR-z<2{ZQ9^_GU^{7C%(NBsVoSJ8Vg`%`QO-@GJ_TX4iAkUmYwo0qc1CchGRE49Iag
zK$sz1{lvxAzO>TpLYo|=j4VoKNCYNTo|3~l_r+6KOvyuZ!;lCj@})ni7Jm_L?{MTq
zFF`$^4W0i7b2=UVnL5LZ-w|wTCi5|T>xJqrC@NM1F^iOOn;`9DWC%xOyaoqHl0yA9
z&)a#W6|!i5WYl%rY#^y^oprpu$11G@5*<)X1GEkI?a9(!rj#ce5IeoZ*MXsN#Y%ah
z%Qu|X3rt=oX4QvS4je@@9j+(Z)wo53Q4++}sK9Ov;D-uW5~#R1nyZK`)2Na$ZI3^U
z+5l?EAfzE6K(Rr_fOS~j-PR)|D`-e)2<<M{<Aupj<1~0hCd<HH>YSZr7kbRayAXi9
zHi_f+a=nAxw$JaOqz8*#?+)7b2T~Beqi3odaC!pt-93Vg$%p;+gvzgzx#HNIw)nrz
zMk9dN+yrU`h?S2knG&6!AXN>|8?4c_sW06G+)i~(>X4pTA$Na1x}7dx?5Vrl0E(pu
zQ04Q^dF$ZPM{fZ^f9rF~xR~;lTpGuyN@@%^f(Hsh<^;W=+`xgUmrfzzLKzL9x2LWM
z*={z8S}E6GJ&V`%c*2<h9-Wux^#VlHXulDtn!r^4&qayH_kn7?8bIZP9~5_df6-}8
z)LBkvbNqaFbapUZv~wB)z|qagJV`JfT-FnZXRtk+%SDBggSH3<5RfuJ3d6!86N&*P
z8k5BY@gRkquaSP0HY6ql&5fln4^U!SwqD#{Z26z~+sc!zzXSUWZ1PuH(e2vUnIg7O
zRw{47qshE5Z;k|41u&jj=Yk7H!KAk>w2EOh<5vc6(=60=n-C<bSOVVADua%pSbm!G
zhi+6L)S@%$>(V@>P#FOl^)jAbn;nY#(=EL*D25*6aXgb>^m>2NM<tpe4;2m_+8sIf
zckG-T@KIp^UKZ5K*~%kfG3aJr53b@Bkm>ruq46r`;)KBaC|i=t{2aip=)nzmC+@GR
zr5m^91yiBERBr`4^+>?&C`Q8qS;yMUj0|VrrTY0s`Y#rcKamd}#lQiPAH13G7GS~>
zobQ3fa*_@iup$BA>l6+H#<PT){{8tzjqi&m0JwojrxW{+U4IV_q95a&F_(yiV7dTP
zQ`qrXs#{2lizCo`QEwlZCBHq`Z~9>%;jrX;;rGx~F~`;<(Zn(<%<iOdIo;F{I=SE;
zaejOv?<i>9-5k9c*G*R(E`2U#$h;D~&~V^Wo=Ly(23RKLl<`()_cS1*lynw}5_#p?
zdH|61_H2z2L(P0FT^Jxuwl!z4J@T=~xfQk+M)e{`#=T3+4_r!zh7kqdu!pg?Mt9if
ztVV{b3!dku=<4N%d1LH4{UHiRHLXF)dhM2LQ<@`vcuQ(VE?<{utUXm-O59z>@=`_F
z{otW#IGnZz0#r+sbH3z=C_9u-T~7+2aQQxS0gO*5K-RXneNTS?5-JE_Om=RLGP9x3
zAZ(Wu(m3c(*E-I8PiQR~;Y5ZqfHJ(4+Ghq1nhUb^EMX{C1n7pEc7{);Qp|jJ7HNQ+
zU%vxnpMu&R;12^CQM%4-@LAIRcL+Cf%Hw2P@jFrD3?|%#C3|T+EMwGvB7g%ov8r70
zI$09<GIX=!&TIJ(g;Cv-H_`<3vpdu7Y8`-dQn)!2Z%I?WxDo%x?M4$V8R3a##B)&S
zF1Xvm=9a!Qk}_@UT4+xa7Rt9bmhod8^;U!axcEpdI$n~UgIzzfteQtV!o(valDW!l
zGunm4nx>z}JE6aIZpf2rg$r}~Pet1g)~fp7zBWr`mvj9#i;P=MZeZO*lOTGFLj&td
z$-Sc3nNSV?tnWFCTpI16V#qw4|Ku7dUC|WMxk%PJ^xM5ZV?5|i_&#2Lb{`{K8!bNH
zt8|Nr-pVn#cfZ*5c#Nh-<nqN(mo&i#gkt&-`QdZj@EU!LqL}_i)$|Js`W0>5qv|<^
z>59aLUzpgY@*v~mtb*#ADEPGDvZ<&o<%eVmr8BfBjd9A?UjY6mz2Wsny1@BKBKkhX
z;OzrG7(LtDyA`Vs7IC}Iz-`o-;-qhh3HLW{pdhWGFw02=BkkN~QxMPig`yXI8~)Do
zrYLsqF_KPWEoXw$j~DWg0;1l%^xz!R^#Noy50kV)qpa<QpYyt{O|L1DSC(3d2D5!m
z-7?l{__64OGtaG0mjvT$exlFRMi)D<=lo*W>3)|u4Owll8N4oMv#^5+6hC$#w~#+`
zQAQEPLct<u3uWp`CDFiS^7{p#4<-yH5ebyf3==(>^JyS+_TvzF9}1iE{Vw-3EG64g
z<ql26o(BG>-Ak6$q6Z(9B)i3$8H2StyAlSc3kg0|@Una{RZA^!>c0{cA1&cS^|M`V
z4{CHU-nzU=<8nTRzT6#*wC$e!7LG=tis^&+6}s^zIXzcmZ|oS2e$=u_v7jCxBy?Nd
zQGtYiDOS~ktb6K8?5oWN84&P}Kzq6&^0S`+Jw}7|Us5a$s5}D!%O7m&pQxgeUO%Jt
zMn~i0sE|qgl-kr|f$k1_Dwx=Nf)Q^EE51V#_%U-e;Qm$rhE7HDCdwogzKm=u%6Rx*
zmQy#Lqv=PDCor2@?3J%(hT{i)fmTF=y2Qq1t=3&oL9;h*)uD!@iGgZ54~u+=8ZTd|
zn|xw_9y22_aKZ=FK(uzT`$Co8>`s78C@B8&{^LZ2&YjTGg(rX)qp;3qWPj1cBh5CQ
zOPCNc@KNs+H-~b+chO@l-<)N)!*8kQ3c)-2Vv2pM&}xa3iPxtY#++$nW%);nGp4W3
zrt0ucR%`Ag>2l7Vxnkb^G^4u47K1}$Oc(!3C2AfAaQ<yxQny$Y9EG7SfF{facT3*Q
z$WOHlMnOD>`J|BCpPP@ThrandMWrh!p)v;UpS}4KizI*9<V{~IKJK_@e0`znkX-TA
zs1o!Af$Rk1n6%v=bd7LBJk(OMl+?F0L&qu9R<YGxn86J#^mC4U)6q92&;5Dz&Q_M?
zJ<5KnFLOgWvL+r@2Xj;2I&;o+{Ty^5(Wa^ky83_;n>GYugRjEq$Eo7(_j|bze7-k;
z<9k4dEIMuLiH$RwD3$lB!;nzv@AKZB-FyRAcyA`h)%5lu?(6=($=o(HxVurbuOMN*
zob}nwMmRp7I}E=J@6ZrXmQiDMg3aph#HV3jq3rJ181;(RPPH8D{iOo>Z|XEf2b_?Q
z_lC>w=}py}th)!KS>tt&&y(4`6taDwp-qx~_}Bo#?EKD_Mpf$S@7Gdvw$Z(TJ!iVb
z5)~@Nv5{HW8ZwhS5TQF<oFG7nhh$<NsK)-hRLoeKj`zXiHAC!=!io08kj-^jk&#NJ
zZs&JVx}*3`YJaxzX+*n*F!J?p#Hle!i}|>|NY0H4-qyIuCaV!ZwI=hp&<tFxHoI7n
zE2J-E>CDao3<ffUcFTA=_orI3(S-S{&H$(tx0HK>vW{>1KpoxeY!Ue^O_t|TsX36o
zKd(?^(Bv!yR_t`WuL+%E_(APycFexOEhl9f<Lx}yBVKtJk6qq@P*(1$j*0YYkR%4n
zy^W6HsJy?`w_)dW7ff0h1-wOLNAE7UOFzxoF5nYC9|L?4R-ItM#GjJtuiutPg$yQ+
z9gv)}tRvqET7$-u!#+D<ZD#q+ks7pn|FIF8-?f3ZU2P2q)D%zU#UDC#fM5bh8Q))D
zyaLR2h7!T`s2iemnql)&S0s(|Z;4JbxHzh00lg9WUWfRYb^@sov)FE!UY>fZjmZq1
zWX@dPbhQ3jknN!Wj01G>g59JapD~qGAcbs_G!M-iKANR!nQ93s)&TIDL>4Vj&rVUz
zwjgo;v9>gJE6O!5^(uV=Vs(&@*IjQ!h!b<049i#Pa*T%<uN#<^_$<GBq?5VkT$z=A
zY<^eo2?8mUqb8v3s^$X%{K}4c4S+XEwT{1YO)z_X+5dfNLQFE@HmqmeeoF|<75nvy
z{#=H)eNE1|`TW@gJhbM&uAI)olNO~mX{$l9G(b*@mLeu*ay`~OxD)>THc%@Mn+lML
zA(_C`6gVH`e1#0$e7GEY=HR6%0764T1po?b9`J!|V}lPRt^-~pbpT^pX3@fvV9o)6
zFZY{6jMa=Kh6A4C(x7J6E_euBOtwR2NM5AOpU;aC9jkLe+h<IU+XI)2ls<YwCvep_
zT+TkY5LR&Z<`_-;2ksdNnh^p++Bj>4E@v}@Ta=1D^$;*9+`1NP>UqPK(2xWH$bnW(
zi`~OgDeOdZjn;72uUsOq02P30#{CX70INSK@(|C``zeY#iS3=voIH#gbaV4^Anm5n
zSr-|T3%q$GvtFY;M8}IdH0`cjmmN&~Y#lw*tP4>4%{zvd-hk`ugRjG$wi(~7p!b8*
zcpa%#cYnYe0$B6a(XYm3?W9BiI2#D)3F?!&`BY);n9)Lp#Bl&Q4OGWE9P*LguURan
z5!ZtzRYPJlK{RUR{?&p4n^g`kkWd^eg+Zh!%%ToI88t8&jyM;*QNwxb0g)BT+}wO5
zjdNTWb*{&i9QyuM^o<?>_2*bytlGPFnmou1qb$U&=ZwW*Opkj1_#()Kyc|oo(rh97
z@-A7<XCCfd96?F=p9<|<i&n<h!V$mQS6W~iEG>!8^DV-_-vSB?GNyKANAhl`C?eLW
zT_`Xz914$3ZTckf&9r@Vau!2dBwH$=Rg>GVx%CzR9G0KcV(ws}0rB@e&v-tmAfSQ(
zghPXK%n?y!$m@hpODPMG7S?THwfDk+ZqrJDYTM$1!{>2^%NNWJlBc<Rx;diC(4Kb4
zX`{K~=7oB^1BGdfS6j#tfl}!Bm0N_?UmghVA5SSBZ|h&`B^R3kCO5EMz3v_nf+u4y
z<~_YI0t7B=ZzPstyX<1~9zfP~|Auj@Zg$|Q7`Y<@TE|;^UNo{^G$R>o5um$6k;0Fl
z;EhM8V1ORH6#CK`7foRyoqZh$Nva3L0DwoxQ%*r!SfYd?4MEtc?Oq*eJMJ|>Ty{j5
zf7$~jXSh{AZ=0l_rhw842zc=-`U+25Vn>to*JnX?1M%eW?|({%-YfN_!0b*p5Mj3*
zF}UjcgoErlavgWMEL5mb(@iAK<KBRN;Cr84oNZamk6ScF36_j7lD==F4fe-lI>878
zkm4$RLKVjsam>PBkxgd@i>d^_xc7^f|811l4lZ@?AZ&3v3e`{~GM;NR@~s#u+zE@$
ziV%5l%#-pRpFfqK=e>6tq;a;K#;-B%KUZmCNNd>Y+M==`dwvhCrqyo7UpQ_M9H~)B
z=30(jtosK&^!ktn{m;h5<__&OHXB;Y9TA$2k!X&6J0JFB73^kT^vRC>((Qh1q7~=*
z!wu`HhhNoSIJ?zmdhF7HcD=<s-4f%=nKEZJ34)~?{O|Tk#)iME$v;wWsNaLB^@3YK
zO(3_C8HlsBx@(2~0?|*l_49{Hriv<V56w!8(Qd73re&e=h*heY^%7gw#d5M@2#_Fx
z0Bu{=s&uMAR<_K0`l6(@l!5T`h#Me_9J~<&Y%}#2A!{GZI>=k%0PZS4ivWSS^XT~Q
zF*Hhe6{my#--QoRd;Q)()6g?bIna?p`1zKv>Ua&%XY2v)ce&94s>Ssrq3(;SzREqG
z+SX<hd>J$DmRwIiU!i3wVZTcGj?3`IMU|zmNvUP(K~s&5l_d=kMp(Y@P&^qdP^P?B
zK0(2Xps9+X>v^D@{dfj5k|n}LyJwdIpW=@;X1dDP<S=APxY-*;_+@l4KjWqaX?^3z
zo`qV&@2n#{?kh(0krbQ|N7D-@y#TGQyEDPKG8w09t0lPEq37MzL?6(Y;eAwJqx%WR
zRH0PFRET4q_({Zf!r3dQyl+PW?yp173`L_?AQH!XgR?pCg@ggA6-rzh>Lb`wYJc1g
zH`$$WdH%N9U@SBn@{vZ|k*FBy{H0|Fy+Xl)+c~Dc8e}-lV^&;M0V+y~nvYsI6xJr+
zdE3{^uj5x1@WTLCcD~*PRcEs_$2Mp*-e9`|-yWOJ=K=0)iW7(jE6=s==iESOmfDy;
zBzkxi=}_P5kG|SIG6E>yVzn|%T|K>mi@?7_V&n(G+<x92+qDCj-FMA1gFlQF)Tq8H
zqRnEeE#}A2CuK&Sy)s%PaK_6YURBA+J!x5NGXIqNga9>McwR+$@_W^<s;xJW4B7s4
z!z(`=57iG{D#{oV)~S1~?`NA^&CaG%D|&M_0QZ>J@6b=1iTgkPb^HWt+b&TYN0M(_
zWOuOr{c0ME(nZw7wb6Kf7R#O7Syb(JFGh0?xqZyd(SxaVQY2lpM$_gZ{rtNNQ6uZ6
z`0mvF;dry9EYM<y8?y6f{PDG)4}~jLNwfmKigvG|rl}fPP(N;$Kf%9viG*k2hv$X$
z^!lCUUM@Q=Op}CZmHD^h*&4Te*=-ofg8OVN8Mo8&>_qU!Hhv5-M##Xq{&>mi7#<dw
zush7SRjEl~7e^aF$Ln#%U(WM_UevtD;pHO+hX=nB*(<u1xl(O)2B8+66$3Oc^tC=*
z**{Du2+nR1g;0&ZsyH&Is_eI{HpMclnCDo&r24Aw+pV-=Q7L30{Dyxyc{>&D;mDwf
zo?n*Evp?CjLBwZ=_^<_%EK-+Os^1`Lbb5nmuD1%`pUgGxj%uRTXkos!)27FG{k3a*
z^EjGZ`(0tn*9W@eQ~CyPU;TWQm6y^RW1r&C_SkRSs;SX>_IwTV{6QVWwx?@h?@9AF
zbl>$T9BxEF0TZ3eXO5>-xA69UAUq=?`pIZ&{Umrnilt1=?U7%pg1<bTQaOdKez5wH
zHLKuKjupd1|6eQFLzLxi_vha1Z)VZJ$k(jR;f;=S`LCPqpGz6k6f&-)jM|YE1vf79
z@-&||dB>JL!j&mxvo$&bIG+4}e0*4s<sY&7F@g%ifqL${mPJ^J;MDUaIaOArk#&?U
zi+PmS$ixFfeP@O{7Tqgd!&g<wh}C16I-T2#@Ikp>;<p9K#ny$6_q{uzfoMq|+*OA-
zyF`p)^pf@hm1Uua2ha+4+a|+0YIR&y&z4xF*-i94jtwF=)Gv?-l(e-n?uE6awlf=V
zD)QZL&G(tA`WbFMHmT-m)X&0m3x~yg=zGkHCs+)IWaadhf3hvRG7^2CC))E(SRrNn
z0vO|uNHuOyA1>l|xH$_)kN*u;EkuA|^pwvT((2D?->i<?#lx`1o$U5&UQzMHZz@Mq
znXsC+rS+4i*nMWt_hWi(;T`>VZsSFeJCPWjut7v<0>Aplv?9nqY?4@G$7XHWz$$bC
z$NJ(KbtKgA_r%-o2>DSX;iFf!u;HwX?o^MlCnpM|<q-+^;AE4k&cF0GPyx+PB7QYj
z!RX}L-Z7m(z(e70zQeo+Iy511hb%+d@y38gfnRsFrz6^p_9PtE3;w`oWN_1g{+`k4
zuZVaYh-4B`odDSv4-kWhI4saG2$;P9eVD(Q8Oo?DmBnF`s;j!k`ZrZ991W)R6&4Bq
zdo&)W_VD3JT?%U<nPP&JYr%;5UJ0>)(_6+0(EK7J*ybaAPfC_C!J*@cniyj2qg>Cn
z;IIROOW)+5DF2(-&vE=NRa)eZ)%1)25nblFce2N%yTk2^<rvQKNwO@iA;-fJ<Y~ui
z6h}w~3ca_*$g&@@3sz>&qu_56phn!UeQmbR3(;hg`s{70fWE`~=7eFT^Ob$+X(T*1
z`^zPDG9}`Y=iP}FM{LYBFC^g_@w0wbWtT9x7bnyEB6+3nH--X;eLgtEm3-Ov;zO8i
zM|{M#Y^<Q+SQjjp{SmvdIDAm+*yKb@er-!&iU+J+Q#7rj0C5}j5&BjVq-}>L|1t&o
znA1^yPj_By^<6(LU2)Q?&-+yRmMG&AlWhI5`+L&HR<lS~j&KZV?X*4nh(OdGzY3{l
ztVbL}pSstg0uQ8|;!}e~obcZ0SYw_T(`c65orqxzJtpQ?o>y#~O-T_yqlfmBO9Y3!
zR@(k29BJcV(H4>L9ht%=o?_Lz7(S0XwcE0<smTL|U5hKq{H#dJl0C~GCb$$KQz*XO
ze<@};Ypv^K%$m#-zlt^|s!^t1-<wg0bi`SnN(i5wKiL!()@QRV$fZNw;-@Ek{Jiz7
zo#I19cN>eKogc2IlIJ<s$g)<ZxHe(+BG2c31yj;@Si@aHNhr}nm7d7g1W+}sNqv$I
zKqi;CiP#BNbCavx0j_ZVUXRycytimNb?DHa<!5xTW_7VQ;~@2eg?r2Zji^q~%VgvF
z!d@W(8%jc<^YCNZ#p9^X_E;vyR@klRqN=U>JNZQy(K)#T2hnn&Gk$0Ey<5AU&<b8=
zzXKV39#^!%A3R1<XAIP8bxcEZ-{{Ia*No-@DefO{Xx}@Z?XPzH5UH8YxfAlxua(A=
zBV@;hx!Ue6al)GL+PBa}Q$H*bU*q{>9T?G()VoXY20v1mYf~#xtkaf22WoFcTN1t6
z8HsV}7Rid(H%|O!(l88vP+p&W?vO)w%c$2DgPGxJJQ?)?j*(8pirp%*(niO4Y}0fg
zL4ZI4`Ov1u=1##2bshT+mCNsuuAKjH?wqs%v`yf3D|`H>>UEeuVD4VEUxQi&KP)Dz
zw>Y2;n2Jl5LZ~!RZjab<OCa_z)weQ?4b^bA{OM{@a;)!_47B{CGO{Vp-dqfwRsFAi
z!EiUFpK<^980|z_pfd|*f2E~Cy3|HNnszovp4NJ;`NPQ-sBa=BC0MlL%b$|vE~|%g
z3RLy1w(sx-5xJU!EXh1h9LBxpOO-fhBL5F~L5IGo=heGN`8;1>&MSLMg+jAN_3`D!
z1GLE<9g9n35lT;`U{2`s0qGK6BJll3BJ8~O3d6g%!RX;b@aN*C$e8h$vee58e2{_k
z4C`E6aAo@vi>3bjdlPmXI3^U`S@GxJ8>LdVlTf~X6a3_p7OWD=;SZ=)zBq<-YVKag
z?By<4v}OyQy?Tv2S<+)<w?B|1q1)m!mAz9It;U`sr^F&c3~3cVq;qreTn{Vi--P`q
zgg=a|88COndc6Jc5fzH%$JkzN<-3o210pbC{tB$$c>wQ3o{|kaVaMW0GTBnwJ%s}M
z*tyGCGGPD`#fyWxLIK{m$8eM{njd?Pp2nMZ;vIj53%&~fTGf9qik5vwW7e?lNSD%C
z);n|QS{yuiUY1|#vK?FY9zm&sxiNnJa?Bpq19`J$fTgx&1>X4c=-CU2TiM_8A#+-{
zcQ$VnlUYt(x`z3yHp+Q1E%UFy`iEFJa+y66ks`Xi1Fze$4;QcBHq9!}o7O1jA%5%}
zHyVP~Glq$sWMSOX*xrfrSCA&TlY1NkH-xcA=Cmo%tY&!+@zze7)hvfJCF@BCxiMp!
ze;2k^-k3((t9|?SVfpgq#%NFBzyJOlRHl+zrfAWk(uWNxdMGa9cR|JK_3PK+?CgwQ
zy?UW;-MTW3J8@_&oj-qm%$zwBZ^drON|h>M(xgf9I_0}0^Rs?Zz?Cak4n2GJMB>DW
zrNj2Yg9l^Vwr$b@H5IpXqMa>UHk2(}7Tj@1#jX;jJ$(33j)m=X6T5l2JBn0&l`B`4
zid%CIP+7;FJ9iK*S~RR&xl-;1qSI8?r^MPAIhNeHbE8e0Hq!q`p+beEwK=)jW5tSv
zKEn8;X3d%yGiD6du3amw!iy9sB5|Q*IDHlzJa`cFDME~C?Qc%3jfvkzR7lgZoxX+s
z{PR!gSe#gKUiR$S15yKVobB4RlS*$=omHz=4a0{Im$YY`+v?S;qkH%6vaF)#GuIbm
z9OrP&nl*C1nl^1}O!6{3b?Q`!(fs-IW!_SyN@4Zt)$&>~r_d*cty{O2ShE1<jCgbG
z>^raHW5<rU@AmSt10qA?#*O71apy#$M1HqkpEGkBrj&_evFRp4G0SiLZpEEt9x86R
z7Co<S26;Z`;C&pw^9FOpIjLs)Y`Bh-;`^IMa{PAhSuz<Sityj39%F~lmlR7K6VvzI
z#e}M95L4(}=8E$CuBW25Vx~B<ommI&VYASceE9kuRpy>CtEj#C{G-{jwCHZN;*#vM
ztI*%<x&9Iv6GlNVq1Rb-^r6%@#EcjQy-OuSQPCHQwSF^wtJp0-rSYxjpM)+YJlYjW
zf~r~K$#IzL=WC&FXt(Mz4hp?e?8srzxoA?9%@7CdPg@b+zK9d%9=ea~Pd_4^(ASKw
zk_L%?>XxX#pmjR+OY~(>CVgylE9NZU&!|TtHk&TE?Ovc|`-qCEq-DFcG>*YLa{As|
ztULD<gUhA#PImV^e;$5-7!gCuIc4N;sbXT|1wr*E@HVH=pT7UD&;=z#<Da`Pq6nou
z=Xs~FxaYM-{Y`&i_}MI<+w~nm;Lu|ECF~bEw&X&Azw*=*p?6D<av9yO+hoGvbGy)Y
z8FfSMo%<PRo+}pqm^EE1^i6@W;klbu&W080svHx6O`A5Mbh_9G6Uy=>by+Cy2MOi;
z-)EkpfLJyB{^Nh*CUF_3ghIGgfkY@PZk$x4FBP;pl#hajrP}=SQrTZZD6pTu|0Zt2
z58VO2E3Cp5pFUR;FFg9-p2om=vGwS|>n}2HnNZHtN}LL2Dzwi%cn?~Zo0C~W@!fah
zO(cvK4lRUrIu{+v&pr#@=qaY-`26iZq11klRx7XIo_L^UP83-_44R4s1=7=usE`7w
z;zUHd)z_qw-&`HCVhew0HNXWAE*j(-cMMba-pBX<{wH4~nP%?4X;oYZ`1tid<V_mQ
zJ&6i>%GIYHabp}8At(u=h7$`ax6onrRh$*uraBZ(;;zi+BFi9=Ho01AL>4a$O85}m
z)h@eunQ+@Z@j@($a9-m@2`g~;rTAC*)sZg(WyFpPT3y&e){75^s^v;qtjZ}-LhL_s
z9Oo}zm3gRiK6dIX_8d4Y&uOiF<@znDQ0~-s2qw>2AQaLyF|vP8+!a>g#VXg4%6(=&
zf9VRk4H%BUR;@)*;dqxy|9Y)DB92&07}~q5+zws2K})H~uTqXQAYqU>LptQkm0h-_
z{=|<PN0iSmPs^4piiGjvN~`q3<!j)fu*~Yzx*1{$>z?XOT4U4p-6&nGu>7Fx)ORqR
zJr^IXD^_EPSd^$*z6|o_%!U-sPKYHGkku-dm35vue-W;C@2Td5CFmzP=8qpg3x)e<
z8Lrc!J-ACCN7hVM+jg(SLV$|%K10VKY1RU0(7H43J$PvLT2<GtxoSwEGe#^bDEaZ@
z2X49E6(4h<@#^(k8J$8bmMBpaMVGIe|83K0cYQH0^=ni?vBGXRtFl~-2`7{vBPY&4
z`8rMUN0;6>bK#OomWwsk(%mzf1gcjk173^%A*AZanJqKI-8+m!Z8M-r?s!_6Z$qF-
z9oI-;dZ0~uM1FV-ImK^+AH4o8dS3k(;lP?#j*5z#A_elHTB3K@*f@oF?}=ql=KJrs
z9l3Tcn3PE;K6C}<b(}xRU0=4cvNXkY=gytTpCb#RMe$6#qVm$x+C?bN$In}i(uMQj
zy?7Qk={W+OhfR^n`0N={W8B>3pwgH!St$9LmMDH~v~O5lT5A`n(@rY>Ne5G>(;yr>
zdkG!IqClmR1u<*+I?NXf1r)Zma^o(H5Y#YrE0#n;K}k1r={l_1v|FYd|K!Y+7Ri$&
zM6Br1P_=XsnPjc)M%BuKR4`qKPm_5&4w;0@H(XJ>LJ3qVQAk>zXG)U-dBkD@$9P;Q
zs(Dnx1%Pu`ZpzaKPo81jj{RsoV4SpCj};@T)p4sj%pLh&vFFHX)ayJLsf*Xf7-0#|
z#S-<tVeV+I9**I_s+m|jYow>-vPJXZ%*F*$na}zToH&Pl$IhZyUXSl}o(hVH>#p}C
zRZqWmjpcl(JX{RuIc6s2jqHUekvtl|+;F`w7f4$58;dHfdSmdEKgIcbRn+W}brjB%
zO)9$0X7EAe=FOX>5|}<!ED7&#R7lezdcc4IxOMB6RFHEqfC}>}RjNn|EYhY>$zHEs
zJxO80`tIJni=|7Kf{OVD4H`&=DV4Zt0fEY8o>QsHa?IPTSu@!cT1~Ug;>C+gr9Z8?
zJ9OwE(-J03D3#8i#XBzL<HwJfHER}VIK+9QQk*>I>>K~nZ%m68Eo5HPKcwViAE<1n
zQl5@dxfr07%wCO=G>D`tqL31dV<J5l70+B?;J8S`LZv^c!l>l0QKJU9z)#EcIC0{*
zE7fV8Zq8A?S7V|wo-{Pgn>UA(lM`acj*WBY&Pn<jD*KBGnzBz~CkYot_`XA>f6JCF
zrQ)1N3e%4sJu0bc=&U?v&YUvd(jVsGVhk5wxPZgCHuoD;apvl|cI_Gp7cT4}DOIXe
zxNzZuhd7>)qV2sfW-{hyUW&@k>xDZYjEgZcgOxvOA;}ic4vy#Z`*+BaIEuv<sZ3pW
z_Nh#N@!=~-F+(LbzpE)OLIF$z9@3yxnsZtjO!gJZ*!p63JQcb<)?JhN9t!%L%@?1c
z>)NYG6fL||=67FvO&a8Ju^_*o%$>IHF8(?51lbZt74#|Lg_1Z7D*RO1%1R7>ER|&z
z`ZgT7{Wh)(!@%;Q{DGUVP;%NysTZk{Egn9LzEZJH@v7hZMcJle5hJP4rHmBHbE!*N
zbw%h>o?GlE70#|NKcSw$;PuDvXua~X)F-hm8X&U1G;t$IWj%E({WiLNFY?-}Lre7Y
z&jkBJZ26zo>D|`3{a%-yc!Vi??xMOF1F49p6Y2gFUu$V{pcej6&$Z<EBa{;Ol$&*0
zEP%Z7@$Z5Q9eJHSTH+^<6;T?(nl}4HTnADYl<Zgj<U%1&p-$(GSPY>gjrr>y&xIcf
z=aWXZlmtRQHfZxrwEz1u&I(LBh(!`*=qWRp{K+I1b@2U*G-feL(x)~;AheC@(bYIO
z0w<3iLe_Y0KYD9W@WFGmYc52O5L$9~@utzUL{elCTXHBz#Y*#*OE1LAU_z`GihnAi
z|8AO5JZ#)<KBZ@z#;u<>6l?96rtKD%S+T^!i5BYNgmroDq*2AH^h?uha)ptKe=4Kr
z)J`YUxtLIB@=3`(Pab%RaU8svQb_+Zt7ckZNfH9J7Mw>?vEn<cRvNjryIPhw@<urG
zzyq=H;il*|)*&k}`SbeIcdQZ*rX^y5<EChXmj6QpFI-ho7kvqoopsivgxe{t<@cIc
z{HvQVcR?s2A6g`#h*vKLDxSwX;$cXsz3_tEIq^X}bk>}GCU;g$t)2$tyypT&13`Pl
zT@^;*qK>k{Z}8Ve5PQn7^2xp2=AJEnSe!U=5aLbTSf#JRr^U;1o*Y>%mMvN+A3}u+
zg@Z>=AbskTIC}E5;4qCOmHdA+t|u(*kD$0v`jV%8{*smG(z-eNbavyiEK(p3(&sB7
z6!`z3eT&Aj4O(^IJh=}s#E0;zbsNPC?`JfsQ$t$V*REO-{f3XlqbE<%qG28M88!y_
zbLT+G-wJu?U!nx@L52V9MN83GShFV=3i+n(yUTta-nLfuk1rj$OH{;6Vc}11_qmfs
zBX`M)Xx6bOj+{6JE(A2JRZZ5LDScYOCHz%b@Hh0(j<pkV)GwbqRa!)O)wXFpRJe23
zz-(bbPr4i`*c;TWf~NIq$=wVE%T^OOpD`^a&q;MNutz5kX%&P|4kg@a(7IbcxiBz&
z%uun5BAMhzUn#IIQ~Wo%NU~`8YMK3mur4$w{YQ+Ki!EG~nX%w64_Qd(L#)<s-GNjo
zl40rUbz*^J4YsdYBDnc;%f$_A=^iRGoaD-u1?SG6moI6yq<P>660NjIi8Zj~uSE?*
z(>iD)1h_~!phQBXIQ0w}(x=BA@i9t=yZ#}1Ui|}A^K}N~ys`&dHg7`Na^<nAae5RG
zAJAF?2Z36Gk}GYT@WSG>Dpsxj8|5ohvS~+7o;ZQrnNoYJ&Q0^jO78Xi*)yU>yZ-p2
zZY2!u+(KH&Z`yMh=Y_JnK=w=+HG2uFlq!s+6Ww<3mHjO*GM8-x7oh;BVwp<#fgPG6
zN5(X=U0S`<(!Oo|s_y+@TNzU)N9U%s;8d`xq=up5-6(}~XGO*|$?@guH?*!-MeHK{
zX0`1GRmz}us|FY{bpe|99))d!u4nFuo>ECoOY;5`=Std`5(RSN<OMgyh_OA>FnuE|
z8rv5&$`<oh+ot7Ty?KYf)@>73^&5oJoEGgVFs^qy)E2uP-hKFlNEtnj*I53f%Y65p
z1IBnuVJ>>`KihOFR8<;RL<k=aZ5#Z;+)~-WQx=Fti@0c7{g>7LNAb?spzBa+flt2!
zrgG!X18MyqNl*pS7-RJ8rIHHh@akFaI|NM2@H}_M)HrqOq*Mr-OZRRZ>EH6^&1*IL
zs8OS&-yT|FE9ECGzv<X|#*7*AorRX|d>5);zrIu$Q-Mz9G==e85X+P)leGSh8#gY<
z+h3wY3Axje4ta+T9V+wE;+huYwEF%o7HhZ=z;p`Rro#I9^XF1YPK7%c8!A+&04_{W
zsn7Nqrjng~;z9(4R{Eq?nJ{63*&TWG(?Pnllqpl9W5<qCIZfp-7Z=Fc&i68FW0Z~j
z{KH*Bw1{V&#Fh&Z+`&bQ^URquOJ5*V`m;Xvxk-~IQqfN3KCStwxF!xp@oG%W%Xf7y
zl+><W+r20$WymAXU5K-1&z3tZPKbM%;*00p&>=bJO9>Z1*f;jyRM-ZM{kV9-ofBM~
zXxOl!oELJ@lhdBqn-eYuaXyX0#Rb!a11jmc$YLtW@PZ2<O6Ws{V|@1PnTqn&k^Me(
z>Xfg_8Vj+G>`or%&X~?Ht*zIcdnODTzY3ppZxCMSFTRO;B*TSM#1Ne6@6b|EpR7Lp
zMDm$$ZIfLX6}stDgdsmh?{t;c)~rj|0b)k2biyBAT)hACUDnY!PeRnqnLz$eAq)>g
ziG^=kb*GOXS?ELlho(QRv{_z>wX!Nt+IW%R(lV?3JyXZ<k64tLDcWY=QV1*W4y&&S
zgUioCKk(9Ozo!ZXI`u@eg+7MYUo~Nrd3*XNNdvR4Wj3?@r2b*ONTV)>`iYu<o|D%2
zVTC_G3N7}#mvEM0QKh68;t3faPTdo`=S5o;Gr6$~@Evjdt~(ecbT3S&a=VehOzM%G
zg&s**rmM1+(qsmcKUanFoC_|i#R3W!8cL;&Eq5_+zAMZ=E&E|oM4lS^uF|O!Q#v#i
zN2~ekmpy+6d!-BF+=La^fQz{1ilvSv!z+YAEQNX?E|gGe2~4R^lDZ+GtK)8nX*Jz+
zW-^1BKitv4`$>B74W-5Q2!ytMmAXbpMBu@_yC^7>v0h2>)G?$opZU1$YFdpn_*+oB
zkgA4?@uPR%iW_iLSq+_gL<}3!12=o&lremz&~G6Y2<j}lh}`3j3x_E&!G82D<t*DV
z{vq%C+s{AbZh(HmN?!>#9#E;S#5fKWS0!^C`2gWLt@=re!v_$BM>?V1Ahv2LF;|C5
z;|_rpC!a{}YdWKyT0OOg8+LKARZTf4QDxLIilEXl6>bNk(6Zksp~VBwsN`2^PsDBr
zO0Dd1#Ug_ExDyZdp6hSOMV79`oa8P6ruY2$P-NJg(4wE*`dslIv{_hj2&X@p5?Z!e
zi!ZGBABm4JzY3ifQSlvByu4WwrsWe>?4-oWo;f4Fefy4{9a>}FUn}Gm%>9RtW3-@i
zxqSVmeEFpkUrDOu$)vLS+KpQ(N-EN+%%_wuRU8R~b$zL7^@Y;Dpj0^b6$)!w;QLfa
zw-Z$?7W5lF&b=tD>95^%Ydcf=H0a%_4Tg@Jf@o2rVCJ~t?)kh55F=BmBNiXj%ehx=
z8_TZUv=v?YiEY*oA4%)?rL!la((fgu(p}~2->n0lzj&oiJsi1X!IG@svI8Ca48}`g
zr8HygFu5qgynVa0lbrh8O%X0^SfPaYDywjIO6;ksZNi@bcUkP+u*w5>zRH*_Qw9v}
z(+yn(<qkgv9|~u}DRUOVAi>#Bx+1UARXwK1NWwynI|^uFWh=TCT5uN@4wWZVPJvq4
z++I9Q+tuf`N1$HzI0zNuf6}QMt?6v7)YtQB`#JMYjs88a>;tJJ>es1-U2U=;Z9I>b
zx8CWa&SP5yvWT6On_Hwu(TbI&0+EU|Tgk(R_mMA^*9(5pqC}F)e6}0=r{i4G%}{7r
z9Z^_9lOE^x{YTO=`>A+#b{#QQrt?oI@nU*SSb8^zUIq)s^uf|~+a)~=t-q<zH!6i=
z3Li66=2Q4S>LjRoUc7!I&#nFOs%@6-+N?I#Zr_V#>v!PyB6(4-QYmXyT1}%uoL2D^
z@3d{KtjqAJSh!}h<m|6mt{DCt-N$`_gZaqiKYvs&V-XMG1Y+lcZ-jHTK(ux2>}4!o
zyA{>i_QR?5e_E{S&V$EtC&H|u-Q>H0%Gh<pG#MP+v6+hUaI7g7IZ`D{g32Y`@R_@E
zJvt7VEEh1WIWSHTHx@3Px$U7Y-<(i+?A7mt*3YE4;X4$CN^ic8Q%FryzI=HP-$hi8
zl0rubX@RKBr|>$Zr9YMORN7K$Px=`pw2bF}3Z0#@o=%-Q$z1`w-%_|h#4@Bup%R`7
zebThhvYv`1E<8}Uh(Jzt3YFU2&B?`ebMI`lYUj8p>?3J~XthtFrMp*Slo@~iAq9+j
z-*$F(ma%j|Zz}9JcN!3XQq7PqhrSHV`DRkeW=msY-K5Z=FA<}V!bpvm<KRL87h{yr
zu{ec03>4iAt?gC+RX!fM1AueE-2m(}7Y|5f6jr>aE8(sR&aF8ig^#IlhXv=wD5OQ=
z4g*tByD@mp7=`(bN`7OScf*_qTUhAx9Wq`-Hywecwu=az2Zaj!V?sYtNSu=j_{eQR
z2_uyI{G_K|WQ<VUM-c{-84Xmtxqj}*SGo&Y`;!TQ5;<IG)X5Pae;#=#>n$eq6--l$
z0Wy<?Ki(-zeFuLVl-R;}ka`AHcbq70_XaA1sVlM8Z>dX3YobdLI)x2Fr(j8_bD%Px
zLZ$@@9lgen>UQnrPyTBWC5fO?W*UX84ir+$9CLMhHr^8~57e8G5u$})lBgq$HFJ$Q
zf|6-tp}4?|&SELn(qx98f4GRzPV6wC@v5^}sGwzgEin&9eNNY6&hC9OmSs4-FzhrH
z>UX}G(d{V9dSm|2KXpMr%;e`Bk$#7Z23!c*Cve~*2p4uJPB9_~J($}fL2c3JA>~p?
z-B1sgYjQz}J3TBp#%1}qVj}>bAzJ(g0j<#YA6y--*4KhFG;Ct)4<>>VBZAwFoZ&aZ
z0=@90lMIRnQgpEr9Ya2xWQ6$hR#4nf>2IX*i7sw7-WBq;3ql&1oe#X_jWWRH{Byab
zX@#(SFOen&7S+$-p*mW{E8$isCDjGLdpg5MxN#gRPjtiBxF`@`(A5~_=Y7;Y%3K{P
zafjGFz#DIEu_K^q-ULDc;PED&N~}Qhzjs9`?YUr)@26!TZ|H9X#@u>lDzU|lox=9{
zP!1m^grqCtu8E7{p-24qpscg#g3NbSaM7{uD)Y`^eC3oP{6)|zMMEyJ6N3Ng%b@g(
zGm^v4nnFK}SQ+NE<X8Ff<%?9pdsR<K!D&Bt@p8$3pFML1$-6#y%p}P@|5kj2kn{eu
zP^eQ92<r}`@I%2UF@Gxa85c34*d-wp$A1bA{q6tmM~7Ym(QC*k+&r<*BB#HxE@SwW
z_~_=x(R+i{d0xSfPvyOmaN*ppHNLD`a*(F-8jmkuc~-8+5WMe7ywkSId-TK^sm$jB
zL;L29@q3Bi<jw+>&SU9fzdpu{@z%l<XU<2DK_ldX$;4p;z+DSUxML!Bwk-Jfz+ps=
z90~aag;9yB^^h)gDwW0kIC=I0;fOdCdWyOz_r$^jcTp6rSWEg1Am=`JJSd@K+)_mg
z;rQvZGKy)|(p5QgM_MZJnzV)R1Eo^gGH9ADYJe&8Nh`+RI4QHp$7xM`lGj>2_L^2(
zETFvn^gmQ6TN;POVxUr7__e=!Uj2%o-^%;fys}sHF;t^UB`m6yEKrpBKKV`hB(L8P
z5B$=!<XhJ)BWhQ#D(L3UND2;H^oonOVZ$WxvZ^TJnPn<bB71(Xdo6tKC^6F2kUl44
zs^ob6_C2zcZ7j1CFCcdQ70e|aL8~%6MvVAt3q8-3vV~!5;#J!$OXZ1r@1R27C-PFz
z6B>oq=Za6>J8c`wZu<9-bXwa`>>A+if()sW8Iz?Idz}g;J;eFtgj7Q`ir}sY`Ul{D
zj(_ajWw>(lj{9PbX&ciPuMzpex>Fi4ojwy5{j&v$h1Guf<~?MQOE+BMDi(7}H0UIG
z`n$KR>!GST<x7H#7}sv!^O{PesQ>^#07*naR1j}^!nxr!=~t20wLu|kuD(~|ozk3w
zm3diSDS}m5Q-y0Ju}jt{>N<V-@}-QnR5z(=sHmr9JC*fRjw>ZQ70$FoU$khEq_d&b
zxU%wRe<@THtG=)eTEf$s+nkv8f$w(e{mwgMG}cMQx)O7Wah1k=TyP+VKCRv9D7QzC
z9&+))v_IY%6LB^w*V&HI#{%o1MZ2+23LiBmre#!Ld8Fk%pL^V8K<Xnd@^DAAQOcAl
zV?6aT%>KB#x_ZdR#Se47P`I;$!bKUi2*mer&X2kL9xC<Xgx^~VV=uB=O(rE!i2wc4
zGnmU9Pv{ZSiSK_Z^N$K8b0uNmM;e(uH(pALldQtxozys_N}*C%DY99YGT{5ESC>5b
zMp~31<x@%}J2~{prB6DU$-=2>jqLHwb;U!RS7oX6HwqaE%!yb1rXI*tmtv}nHm%p#
zw((k3`63E`0ZJ4$JSX=)h05{oKlMK`1vfpJ8Q6a=NGxxXNm{S}D|`dc&ms*(7dOm=
zA}M29Op}%V{PRrgw4gGdi!0=?Cj}Ab&nU{bfHBTmnDb7$XQObD!c=d>^w^Qy7B#p)
zk?-f@QR#wI{*FSSpH(c#929;8NKctd>>{Z?|Ey$QP?S7W=o4G&hByZ95V<9k{nixv
zq42|fao5;a9kvYu9)EGRsZ-bDfC$i%C}fCV%5d*MK;`f&p+Mi-I;*?YwxVZYc{;$|
zbnykA3fi5x!g8B(MBE6?2`$Zy3V5!9k|#{bB=Mjk&Cd8LspJbdsc(jFca^&W6n_}=
zD&8>Wr4ttl9es=!&TZ)p+`E!EdN}z2Hs+_tc*R?9j5inF+9s<z-42!KT<xwaocmT3
z&TFYqTOd}|NA0|0uBh0~76F@EW%ZDjT=3;{HJ_{9N{;+ofi`C?m_+oSa$l@YE9blP
zQ$XRvwM`+97SMc{L=}rI+)_oFq<V`lVqJ^O?z=0vFhlt!9Ee#H&ZAF12+h^U2My_S
zyb>yG2MP=9<*QwUC4C`e5mw@~vL7~nilm5fN}Nzyg|nW6$4&^|`?Rvk8@F8rhx;3(
zNaifhP5;m#uIs(~(&C#|-<Pl5K+fV7Fi%kT^y=8!G`rRF<VlkVs-9<<H+i&szQrro
zfZMYvSHup2q2ngY{?j2ZE!k-uZ%#so3hiNG;L_C_=CZlRdDXUiMiCg+x0~ecC*4k!
zhArg|1v<f{-vMPYU#fZooH%{XSj;`#z5O3@r-6Hf2+$t@mH8tF^h7t|tk)=1#?yB|
z%}V7lykB>jv|!oa@|Z4l3S(9ou35PpvSv*09v|lHc|qUvNcb_xfp}u)1fA)oa88O5
z{oOvMnbGy`eZg5AUzYJI-MwNn0dr1v&1Cad)cpDL5c<Po{E^qAvcg-1eVOUD*k)OF
z&UvhDnnB)2tySZbyw>VbX>_DLanF71zx4*u#O|Zg!dj<uk;EQ8TvT2j{}mjR3y(b(
zTaiDA#mb0+k|+U+iN#`DN}K$GSAT^o=FFKbKCIi>q-s5{!E;`Dadhj{9u2d{M0P7&
zxi-0oEL7r?-=t6ST4>Hgv_1ZQjUCM86a~$M!#2**8kdd6BwvRkMj&-TNMDyXE$E!s
z&DuY|pxznWshLbB70$33i=?If()HUx#rUDsvye5N+hR7UZ)_o6waqr@jF)se<plN5
zKil?7>vLsAZtMds<S8G7&mJY0Z<v3NWx99TR`u}6@h@4s4fBLkSz5~H7LJ2SX;ZCC
z5vj~4M}PJTO^ro7g!iqO?#U6YhYThst2Xb(>dpV+#=k4wQ{u#OGl=+MV({UfK?GKB
z{a0Ao7nZ)2R6LdLUBpfYqudeiC-lQW$G}%_-9?}Aa|B&dO_VO=wp;Z6<EN75CsDlE
z#$ukrRHTwd#;-!9vZ+$mv`$*olUrR0xzfo|Px>0BDJyuDPN$^G+Fiw&J0=r+_Y~Gm
zet8O==uVk3Mf%+$2R^O!$@flSyius|r$bv(#*qJ=lr<F6<4{PQlQpZ`cN6*dSFBiJ
z*7}`!$$3wM9;NIiWln|+8JOg!G0F-o{Gmf!Dziy3#Q&EsUzXPM#GrTY-X2Q%W=tL`
z5GT&g&eE!$^fbz%oqY2Ya{CiQYcgWQ2uZ0!x*a+;=59th!6hvb>!y%KN!2MM41ehC
zn3O_B;o<-(JE)lFb;u*HCoXQ73aNN_&6BQ42^~6f0nV=qeYR~}h*9<X^hjfna6+-`
zmHZ|Q;z;*G>*$dcQ%V*A^4)(DIu3HwQ|K&~)I8+bkM&b&OPvFS{PVQ%HwwSmN%_)N
z?4oZV9HdfzQh(8T;ivF{mkNEqlr{E&VXyk_rc3eI_tW>^dRv!b+Mcy@;zVhlFR^gs
z>&7H-UQim5eug@zNy0I2Zn5iu4uyFg&D?(<!?xX$ifpnFSW2TKStVJ8e}bLub6~N+
zeyp&9A0f(@6ApXH_fKO>>Mm*u<@p)G&u>mRUmbr->LH1aeaYuf#XNlx^2i+q<oP!h
z>bu-@LGRs@j<~ul_NEaFN2LGhFMJA+a;VSG-3QBs?*Iz@7d#dm{1obg+;v0ZTB2|#
zhBXN*-ib^X5v}FdX+A?hEA)Mam(Fbg0c8nqA=zv+oupFkh?T+wVk^{gK`}$`+L^_T
zln!$%36JYkypI+9W!;M>#YW)>X5S5A`Sg<q%~&Zdt!Q08Mku+-Wj}I<t1($RdEG1)
z2ZokQF1MsT5;vu2LIJI|V5!Vp=_I!;X@Gc><jOcX^?9=<rH*%{VkY5qO!R%L@CU%%
z2kiSrp-dlBDW$inpaQ&A!NmAOZ1FlE6j^@>rzt&!6W@WGZ^?epkzUuLzbuuFmD1a}
zsc;dXtndrKdI}1DJf|4pFl6g(sr*-q3|-e=7mF6*#C9tuxs{A_GJf}6$rZ;31job0
z0&=zG5Q=-F@MR&Hco>q#hz~khcGq3Wo6mWfw(q{oUQF=ASrfPY6xMQhKaKtHN=UUs
zp7(Q?u1MuOAEu--aoM~b9a=V##WST#BbEEJ1ZO?zb7F{XsDnmNk`7dBSFK>S3BKs}
z7&H>m#Wr_x;?p54c|J3war@W~FDyJR2xq)cpFKz1I37R7!bfmhOgVh~lvE~@f4^N%
z!6zwp4v-sv!?vC1Eu77AVSpbMEjsm-)IwC^Pn)+0L&i=<4dDd0vS>GgSZI47sAI?v
zpE^Y{q)U?$q^~Jaq#$S=NSYpNVy$h{a@-xjHu;~9Ye`)+XYq23m^fXw*IQWAHx&y0
zvL%X|7M16w4uToUg%K{WBovl8%h#-TFBef*@l_IB|Cg>_NBdp_aCC<Y{@s6Aj{T2D
z_1vkJT6XRQr^Jase*VV7pMg=P&Rr;d4j2{qW2emq{SM^Ho*A6)i&w4*r@X&DD!fWJ
z)^7}Z=UO!8w-lxo*^mMKu%N2v7d1;|Y!T0`p)SocBAvL;dL=%|>s4LcO4@wIRk`JJ
zXjy0RvGN@QgkyEyQ-+pJ=BfC3;TXO7ifgExDHbMGNr~-OU!n1`D_GGmy>$BJsem6M
z`WKCZ27?C*KG+uGUf}kd^=)-}UW4bn@_Ha2@}AwhaHPGP!pFDQh+oS2<TvS)yk7Om
zo`1i_zPHL77x|{|lXMwW8u?Xhb-!2rh!;C1isZ?Pr7qi%N$}g#@_e*#^16D<9@Hu)
z=>6ivl38!veJHu>7p>WX!>2BQj&<oU*QXLe+|M39eJ&lhrgBbVwQX95lRuvnH}gjH
zLcv<C(P`*p6v~xZ^2>)87Tp)F-va4^QaN)WHl*nL61ihwuW(v>LO97ynbf0!i0ZFb
z<5v09k$m=a3f!S_4eU938uM0eK)b;c(0$Z&iBrcWHPE+hBbD9LF?P&fR1BUH%|uiY
z)HhQW{w<vs*AnADcJ8uV+#sD%YJnT2b^md4(WQ6m2J)Pi`%j*~LMgF}!YJXxg~hNg
zEsRk(bm}}(I3<$t?>~H!ShX8C0h{KJ!{cY-duG}~x$Ce>sUjZAnV#6{dh|{`<fW&=
znf&LZbRicyt)xlaQma-i$-hqOo026<N{f0bT*=q&mGE~-DjqHj(6MUNs8KOw$PlDS
zlSb|cpvC#qr%&;J_O1dTitT%!-+RVWOh7CoBve#H>{je}c0aqjyE~p;Ji9xe-QC?{
zAs`}Rcl-YT^W9~Z*=2XxWfvvp=rS|6=FXkE^PO|P1517=3BY9!fkif~wXuH$;hjEx
zx}d9pwKoJ{Gc4T!AfpZ_A+Rm3({Sa0C@A}bHo?^oN|_ZaR*2xXI8KN;$0*x{SOp**
z5)wiPo{IiJVUZ?H8ZoX=l!$MNa7BW=Zgh@GULWg-CxCbeBSwrMfb+1n2RM)OhVhp;
z>b2`h+xQ-kf(#=t^8ESp#d*fHfi?yY9xT9p(ER9<qD6~FN(TQvD3TCd7xYbt$p8y|
z$<lwxk|lyx37#6b9_rMoBl--=A1H({e&Ftxg5yTofsqNu+y_QBMd2g@28)AN9pEuy
z-XXXs-m3s^5u~&XgZ+>{bJ;N!ZZJ(|ykuYnXL`(rFe}bI7CieB?TmaUz&chj9srAL
z0K~BDf6i1sFdl>e^aV}=B?5TW<q}n0dD*Deqw05c=KDurS_A_I{m_mZ&p7BVV<|P=
zMbn1*8P&eE^AX>0%$^4{mO*xmA&f~!^EM9+4dFn;;Huww^BIk1&xYI?Q;P9X;Ydjh
z=4tN-@d+l>aOc<tF0|!pn80OV0T08=X3Wpuh+`U*V@e-g0;5T7frUH*C8OWXSUIzl
zl?7m^kgg4!!%@2*P+N|-R-EH5RLN;e%lX>CIt*u#=gQa)_K)l8PS{5p6mXCBboS<e
z#;+L%3KH=rLB~@zn~lJCCLauBetl3rL9r1m>f^W!Nz{mpq1sAbR3i`>`HQNClCO+F
zG)I83QVpq?!&>m=x98xh5O9NaFO`L69pev#y#a!|eP~4|VM)K5!G6%O)L@`|kDto*
z672zig~08}i#bsNzR%$oA1Ge3a+G#S+AxQHC+01yz393~!V7LNx2=mlVp2^vYP-gd
z%ij@Y!Q(%Q6#<evlJ;>(+tL!1I<p14-;<Bjn%j}283&MUzz<%CysVV3Yl-74;gUxD
z@N$9DW7UNxtkJdQKscs>dgH3OP~eNt)MC{wu`7V?Yg^<J4{#(L;|I4yD};bAzV)+l
zboe{M+8+6M@GoeToq0&LP9!pAf<6P-@5>5^QZp}!90ZmbQQ2A0@5qY&@>NWd0d6CR
zZLvc6MSn^cD=Yy0iasU9Hmot0&Y42>TXmuwh0BUGE4EZx@y|3hC%$wYN(WS_Y$-w4
zf}pbafr1C%xJ22~pjI`SIe)2`L@oE7CRG#5e4e=(gw8}QyYvy<`E8muq^mdm*lS>x
z_@Piuru*pMt%CsY&s?}n%jR(KF8-KB+Qx0WsB{0Jg2x{Hn7@2Ad6lY6A0OSOUY*)f
zjlbFmYjkU-)`3C-z&rXc9fP#Qehlp!`T7L3H~`8=^zT7icJ8J5Y_-33^ESbMudj}-
zbOg|aCq(nky)?4Zvl0Y=J~;WIII(ugBk~cv7I6<GIdJqO<zj`5F*1GJaN+gPx_du0
zPbZGQb8FRycvShu?8DqL2zKUdos^Z5R^~PpsXn}CCLDMuKyW-8z|Ftq+H<~_9mIzO
z^2RWTj(fSwMpmSh%uEYThw?|oC-K~vT*E~u1fOyI4XLay@1ICJHjr>K;Ek}fhcY35
z_=g4*qiZfbr9tH!C>ej?$YsgdP<CHUBxt{oUW%2eJ6Wl^{Ct?j0evWY-8r!4_U+rL
zZr!@_ZkW8ToL3SG@A!6JaUsm}okPv?*$TzD$rpwR6_oqegu&7(3L1hIe2(V%&ZaqY
z=WFH0%qHHmzwytI0A!KM_>SyNje3lrW_`wpDppLPGkIWV0qz4l248-YUZX@FK>z;j
z|6=NwH8gP2TvpyESGQwme^OqB5(U}HeLLl^*@oU+*-M7@?bD<i0gRs5ucH8$eMj}7
zk_|gk|B3(5!tnz*xNRAly<)QfB*8NezI{;Aly208JY1cGRlh_T!?O@kPU;7ihW0Hl
zCtt_dmMQEleovmgWGb6ow0Of#s#UQhb!%?kvsyksf(Jk&4k|os`a<zT9{|8Fo;Xm=
zjeqd;MW#h^R(J2%`Kw~Rke6u+z7$r;X!+zJG-AeL@~YNCv;pq^MLfn?9Wtrg|J5sf
z`}QU9=#O9^7(ZA;FI~D+Si1u>2RA!_UP(H3>?r7Mur2Nho|Eem%+*(|S|xbcp(ucL
zvyYDt&7M7*=@wH6EA=sB#)y3Z><7ROK)Q74(gdz@2(0vBq2IKrYEFl=ELpM$Djo#T
zm4d$l@C7KWPrzHhb?a7|F=K{M<^%ketmKV3MyXx-wS+WjTUvrAAIc6WG7u0Ko(?Tq
zw4mnAn+w1iUIyUn*OsV`N!K__`;^*(x5Bh((?nom0Ot_!+5?z}NadxaNYwKO9s@WR
zP{4p&ABut1t5=Jqo+>(%5jYv=4+;jvOn|}z3Yr21RLZfNH*XR=8DJoR>jQy`;Z1?~
z0}v=85Exn8%S1W`X{mGPzc3M&-=Js#eJuc2;1+q=+Yi8NUbgs#z@mI4bDEE4a2A&L
z5MU~kYPpMb@obg7ZZMDU&4!W_+4vEd0>qo=*d2<i6p^5PS#ka``OLgPL8DaP!lDrW
ziK?z_H=+svap^rKwIMB|>US5mNcZD8f9q8@r78kkZWvArq}@+l`&L)fv4U~IKS*X~
zVVF@jJN4djhbsB1unQPGE@<o_@&Sl<T<b@*7pR`$nSmoQ(2w(BC{I+@@dtD^O&Ea3
z?+SYzEI$`Y^G*cQ#r{Qfl-8eAF^4THWgKYc(T6nm#3P}M8CBJl;0*u)z4d;kRr)jB
zn`3-+Hxb;n`HCANUtmG4@~E+(UOtWt^81tfsFNZefrW=vb`k3W8*1Re*w8Kdi#<9Z
zz=jq!&KV^(Bnm5kVr&Hd$C^e(X%b_@n{XN~SE2~Db$u^@XH$<F7W(`>B_5`zi)ik6
zXh0hXK>(OQE(35#7XU8`h5vzjcXEDFX>*bx8<+Yec0ohi^1w0j`qKJ;Y>^IaEI8)f
z+097mi<EEd64XDB-+mR^OI>|3wjFsn^bOp2axOf1epKriZ@5)S3JOEz;VA(t04P)N
zV#8#@So!efhdO{UG{1OxMJ3<}oOUA#9XoR&s-gz-VYQK+Eq6k1oikX@luS$#{qAyr
z-2Z-%hr6p$R=KcL2UmL`Z!h|TxmD%zhIw(}W}S?!&9w#ldim-NrDMxHe8@n;U@;dE
zNDi4Z3BX^wj=pW|2i%wNM1Wx3pZn#)pgXK+jFIT>Ti*7^kDr8ML!M@-by(`d8v?xd
z<_cmbz@q_ykL7U+q;{jqm&(bP4ta{wrOQ|Jx^T<O8d`7NyeYZ9I7!V5W-wGNT6ul9
z`P2H#&!|~}^c;}sJ8k6v-k2S1XyGA1@^Ze{1mzBbZetzNqF{P{77?vtMPqqZqAqEi
zo&3T+h<ttYJcgzm+8<Pl&V<sKeL>_uq>xa4VV?l)hwKWZL!I*oN*$z0Qe)u_p;VYt
z*G)>IT8nN7YPTa@G-=4x1<Y7kc_EB$rYu0qR;@Lo6(Z%l#?SMLi^4feCR$R%PVkSK
zb&*6-S-F3s7@#P!5je2%h7bx4eHul{sIwb4Y(SOryD^yU6Lnz|O28xUUf?5o{q8;G
zaB*aXzoue0Ao#He{F;R=#38VX2iUG%&X`@k;eC8(N()%ML!#QZp*GQ$nCpbkfxs94
z^7UIarH-Mx=I&c+8~um3g2c}+rCVYd%GUBhkDicC+EnDs#u$kbMLpLwXTRXp|MdBb
z@N7t&IO;jvmVL))%kgWpVbf-Fb}PVR(3=2gj!Hn)0*-KJXXi*oVO@@3sj!xo^R`DS
zs3QUM2Y|G&x<}Afxghv1tiho~0BDQAsIb^a5YwmxfH@Qq@U{bhF3r2O>$Pp$mabpF
zPNz?wCIH9)iD5ObE#@4fysbAJQ%ua?ZNj~9`1oKgU#L(a{-%ih?I5pf>@msfXs=`c
zI8IRJI5|0~fqd<9rf#Ee(WR|HpAaM&3Wli00R6%>f&H0FwIdxPf(=iaG)crVkkX>F
zgpwn4?_4##Z*<zwt^>~bJ3I#m0|Mx5wCB{?<==n#PL~-M0nP!wez~A6U=0#wt~?Lx
zhU#jUjcPlpeuLsAl<^%`?x!}E5>-E=+P8K-=!Y<dz*u0oBacfUaHQaH94VZZN2+fq
zPbw4D65bcUsG=hnhx?4zz)~Jcm#8F&sZ>94d^CMn(A}pW1eLOO3<I>Q!iGNZeB(E@
zWq(xrgH?Z;(yc`>@2JWu`KFIR)H4Co+f-^Kz!CWG|Nb+^b4=&RNWAq@rZx{3B6aJE
z)-9EnFQGBjCwQ^t%c9a)>MJ9q;{lErhq(fcFMgI3^eu@Az`FLnwZ~gKU%m}%d{7TX
zM*!d(3V^GtO`bal#fz>yQ_}Ea#WiV3fyIZ!p>J%vuvC`^ah2E81r|&B*jIWxTWM1)
zv8|n^Q9#dz_I?23XLHS}r?l?m=<ZwI8oU%@N<6};j4aW9Dh`B}-^){PQ&HvXst6?<
zu1Ry{;`y^wucTqO#^~$=ftuDa%~+38_SCDi%0dSoPE(n?UY~%+zK@TpEYy65nGP)~
znS1gvjb?yem%!!oXje_p{Q-LvrAd}ZV?qr|5D0=kX$$zj^JLb1Ug0=Ey%Q7SZJ&ac
z9WY;MsGL`#_;_AH&G7Q&D+cqOV&d{sx2fE}>Y<4C3g`qsFwp$``E$KNJ&kP~;IRcB
z_mOr7kTF|Uhe+vC0+dH`v9nPr@#Sg8u7kT>E>Z1UURHZuyDcafOi?nJvu~*{Xd_9I
zs6h^?yeUgi9!a!ErqqjWSt+)FvGv7#n6F85MRi>pH;)1Dy2QP%y<Z0h2f{Bpg6}we
zlqm>e3y@b5fW5|m`6v%?S>ooG>Kj_3ABF&8bB<BIUu~rYtlcG!e0e!t>!u!)t~RiL
zUGz7)%9@fET`(W}K%bz{FjkBmFJs%Y=>25uPK_*wSat>2qTg6S0ND~uy`oFdmLV}z
z*HArOc~SKnF%0#^F^sC8y7p%%O?W`)!HNWqq=WVb?HS55v_E-XbisVIfpPHG({K4c
zq;)A~fV8c>bhp#|mXhbIup!Mkf;?Z>I_5-phbR&=BVhP$jhTHat;WL<`0ancg*z?9
z!-G=}>8FhV&yh92yXfv1FIHTWiYOxx8%6+1I9!wFN@&<q%8=64+-4%x_kaGGJx_>A
zaC{9(N;1`#Awa)~bA)mfL9}%N>{46OvaEX=N|z$Bw5hR#<-fLkW@{B#2+2kA64GTg
z^yNEyFR@pVTu@$DS$Uc0u4iB?s3%XKMYp=To^oDE?fs?eT*cY*%6!b})9C*wQeOaO
zt|<4fnFAPJJww{$f<xQZR#Ta7YU==xE%3NEb+<}+WdwANfS3=5dwS|DU}U1D+A(r>
zv9cJH2Bvxy#L5FvIu-2^P(9zxpU;bSqTLInqA~(;YXnsIj)Kd?)LByi{g@ir1lF!O
z7W%((sNhxwIvrg>W90q%%x+0{*!>7mzN4c3k=Ik!O&bA~23Hf1Gpz+x>6oz4k4`^g
zq7DMJ!s^~WQ$~Z$e`c5L8@KOJ4mVfk+ce+X#Nf`9OoJm}DGU7<Z{8W@N@=h&Ix^p~
z+btwT(um2Hjg{f)0jSL+F&4bS^AkT~B^`Tbf&K?Bu+nq)6>leEkX{}mMQRV_gI|A<
zHpLR<&E@jfXi&D1j4eK2zIiW8iQcMmUWwx4dHwR`3oFWE-Xcl4f1`KbY1hJY`qQUR
z+Q}ww$CQP>$sLK|3ZK7xrJMHxC_8%%SViH1XG+-^?skG}-clCDmTD(1p8y2qH6)<<
z0d)@o@FFm)Au*+_p>~wKpF0Akc&ZrcnP|sQfdrIi5*EX_jhmf<#l*GCL$T10*)wMN
zi#=xdt+XDGMgW!@=Ig6}{(!Vyds_rf9L>RG3`qd<=~tP1mA16WO$bn9u3%z3>)0dO
z#Q_^%F?9qY%MYvMOuD$rurI_0n0ELf9p}Kch^mfg>^+$pL*jNjeD9TrY<?}|J@ZOi
zalqV692nSIbQ3&xtC;I~J4c5PV?KY>ZNdS1_3@^edu-;^2QX*}OJ>#e7#FX|s7DXM
zc`KADNz=v-Gg$cEgFq@$u@)WPwvLJy%5Si&F`2k#C(QVV{^cOIFJ8T(!uh@F?-2v_
z^41?cd4@(!o=G?S?ueKJ70Z^U4lVu?T=!^e=KLkJWc3DN;anhZ9_rq%C6y~(oUY%x
zL+yJ-1h$p;4OBQYCybOQ#ff!GUDqiYG{L0E(qbP@Rdz1uo#2c_`&S;B4^(US6a4oR
z4hCx{i)|OC#@W?glj=wc(tVQyj2Cc7DaxG*dP^?0e{fun2!(>wAj)5T{GLJ7{}UG9
zSe|?Mj=BCVA9FzLlg~wRt=Je+Urbpl=ancvo>#nwk|)*ekz&gA5mi~`{*7vcjLXNl
zg4LwCBIY0QbNZn%H1~rZQHchf=)m&nR5V|1jXdoQQ_5=B)wTVIV;zk9X9XSNptEUG
zrJz53iqXK1P4%9?N%L3J#yy8A^yv#KS<s8Bmn}~9tNN&$Ubo{Q{j+Kd-MSyhT==P}
zdb#2>qFXCUW`XCsdMCeVMRh%ze?eNRonJHzKV`S~?%fkXT8|$;PQ{BCH@Htz${K7a
zJQLjd2s{gpckuextXWg=RY&zq11Ej{{Q0SL>C&`d!2<QJ{{8n~5s3K6kt0;1L<td8
zxK5oq>imSdVyXK;w3AQ=QR}$dtiDO5-K=SJHNzgr0WUsM<$o^GmbPA0h=aA7E8<&}
zgURl``CNDz=Vv+{1c23x#W0uuw8KGk?EWiC&2*rZI3R7$GLC92M(WRqUH<gq!&jPB
zFNc&w*FxUWfNl5G+tSL$)_K$hOHx{VTw3ZoI6jVzC5}k}^er)pN*jV7ItFm27#~^R
z`q>0P9@UPV7;M_w)>933MO8kk{OL!68PE@A5I?I3;s~(5=7MW<q+7m77PttK(q^?E
zePvFJ(N(kZOX3?1+x386fBMGTE_9s%EacbpOh?n}xYNGdFR8_<TeP-CE-LDrhGw(k
z08#7*mUj}~2#A7ygkMI7yL#&c#iHtaOy&cu#0%D3K^yoNA7$12nfw78L3g6>7+swm
z>DZ19l-nblz9J^2kDWT30exSZG;%OG+GnOQQ)W`BKkL)=<NG27{0$CyOqCn7;9$Uc
zXzJ*p^p=C8^&UEguHE#b)!hFX^OjKG;bW<1$JVTDC_tNb?50}H+tI--YuNK312w4i
zr$*mGpFE{0z6+>wxiT8*ab?3&*EJ?YF(C+u1(1B+smEl?lsHKkEFZQz(Biww&x*_5
zrR{aLn>Kmm=}3iaY5Dp5JB`>ANRw-1rJ$EzX#JIE)TXe?I|YF(NAC@yHbpX0@vNz-
zu2(wR<@b`t?GK`+th8Kq?g<CP{YGWnBB+FP?%a$-OPmu+w4v(|l=G@<RE$fDsdHl7
z-c2i|oM+R<%TQ^6Ear-sf5gw}hsMyjIAo^7t7cM8x2ziJx;9KHtE;{|?aSA%)Uf+-
z`ttP~O&QdMp1*oSea6qGSM0$ynYXnCK>ZlsWz?uzIU3TXDD6CSidys=O9&`it3pZI
zf8so~>_3i57s^X>$MogE!@)Fi<`T-xaf`aOsHfdroZD7ikLF*HmTD)?4WJ1fc6D{7
zQ>RXez`9XYGo@@)ZNVF?cJ11P<;<Bg2|<awckeEO0?(K+!(jDx?b?Yz!^MgfQ<uAX
z^(xh>Rg3!c=|jtwEu(GQwo%=>b?NNcvnItFP&XKlRxEWNh<4)9xR`S&|1|Ir0C`kx
zIN2nlZSB1ni!qJAU0=J?8GD&|G>C>Xh!2kbsSMWFn16+ib}yh4i)qb@hp$ZSIffMv
zz+#5(yw6z7ccF;*#u&{mRt#JUeoKo^hftYtqe^MtH$&f32rC5omUWVn#1h|I2|+CN
z9h?xxDmup)`3pKattfR8{0QJ9$c#9xF*bg9z)oOrXW#7?l#wI2H}FnRTdqB$!4;fD
zu-AnQV0qc6q!xuTQpa_-si0FTy3SVa$Ji1ck<`1Ew5RI1)6ui{-$XlOs%5o^WxFqf
z>dgvf6#K=&r()2m*23#_;Lb~`mdjRSJWf7%P1hK-he83`W5~_W547fD7!6{Oe#x1~
z+}A`jrIwq>E9IJ&cJfPQIfGdM^fz63N+mh1aekGR>Q`0}blY^7?mqoUIWt=3r6_i6
zfj_v0Eb&3Dx*p@^OY+;dZ!~1=WZJ&>0A*%cofZx2(wfcNX#CLr0-T@8V0Zxr$~(4f
zLXF#XqvC}N(50(4=)lnvY+06u2K4Ae^=ecVmhVHxO{N*+hEuK_=7ai^XU`WF_Si2Z
z71REdsosEg?mt9z{;aHV_75FDML&N0pdD+LsOf$-@7N=(^-*x;`pr}zuP0mh4-)mt
zl`2kKckL(NMayaFoGH}3eJgc)-@kvS5>@NcpOwnd&_3PN>2YbrQXKjg&~@n?HUv2O
zr=C0Y+H#-Tt-H(AJHJy^4{HvT`%Gg}YY6MohQV}|MUZS$0%a#l@;fd2D?2qbon=&9
zO}MNF7~GxU!QI^@xVw7@?(Po3C3tX$;1=8+f_t#w?yhHl=iJMG*6dkpx_5WIRrMHn
z{TzET#Pnkx7I^g=FAKivdV0;i-}IAMa9;i15w%iP?lU^UapMX1M-W>!fZ{n;vi{(Q
z)_%Irq|SDhb^)>hwI**(o8(|iV>%!8(Aj#U>X;XoT?bMM%!p}q8BgjTJ4*NYkNY!z
z<eyy$?!-b08SKld4Z10o>v7^7B_4XRy~JkyCCmloA?<8s7xD@|@g=6V<pEEN27Io^
zWIClX5^t+-5`ZhOr1pTBb1c)vRcg|(M2`Cc!CdTfO;$m<&F_o3*+1=8>Qh|F0vU9U
zZH@7M!&53xrqE#R_{M^bR?DYULD%S`V9+56k6Tby!`C8%Xlgl)!^@2EM&tE+Pl&_l
zr&?$MCH1AI`=5q_U*W$&yWEO~z`!T`-Pp*{ZMKyG8oo;+F{xF7UgIwt0AJ$E^jbJF
z0m8{WP&P9A{`N9DG$gqTur+dX1iT7V7PBA(To2KJ?mtl)OyFP0GTWT@5CDPi`%pXy
zbMX7IbpY@p1okpsQFMwaqdHS)O1V_|d4*(}#M%ubpbrNCs&Mu~J*T$iPRCVy)xQc1
zq`4Wh?mR$G3keUIE|a&IbRji99joE1&Scc3V>NEglaC=XlsJAnuDwsK?TV1#XZRO1
z=t_Uqqa8ye=<}e_$bPTH9c|3&5-0qXf$GOH^AJAPL=%p_3Bx|L8S4>8_(#zPeRF*P
zGZw-Vk@^M%=t#>5t~dxg>qt0<|J6loFX%NQCQ02IERWiwpLzL?QJ{^mDf8XTVuF=E
zu#2^@C^^y@e2HF!-Oe(`CtTQ*clO@+ey!$jY=NdfvCh4Tos69=`xU3X`}UAXxJRQG
z&q3%(<gdMeg>1Guq^*%7U1*zahxlH-&Rhhu@S*D#f%iBliHY?V{d-!oe&1u*?bS(u
zy0+x{&V;T7oagheM5;WJ%WZbOiTC)9Cv)n^Xgjv&fsS`nBR{GDe;Ld6oczwe6YQ`M
z>D}4!HV;(?WC+h>nxJL2rA8^3{@1wKF8_UurG)%Cj%MbH6W0CK=-9J(+PyXuJjL-v
zdyK88o3xrT^Cclu`Db@4>P%jrtNN|!;h_bB-<JK0*}<x%f^#c@LnT(jIqG4IE1Ln2
z(!+|`#Y+yGM{Aw=PK#aN&>L_1hvU@?97Wyq>?ajXiL%~T1O1y2%#Vgf{U9fX{o`_Z
zmGZWZ*4i@4ZlVl)MOVm2Nlq8-%?FoG1p{bhI<tC*U9+ux7MbrgN(Y;;<(v*}1Z5ex
zkl-j@!mH#px&DZbVt!1X=XS$a;DerP-=TX**i`Qlzh{pv)vcF%e!J@Tj=R-NyiO!G
zpzV>JAb3NZX7tZ7mrn@|Ii^zK4znn5H+oKS%R`apb<eGO5Nu&?qgp(MVW>5YbTN(O
zawUp#Hgt=Pa<WS0qfWIRR<238v9)xSAD{Do$k5c&;1hBJJPPa~ny>9?tDI$uQ%2qS
zTu{_ePtrAu$OFB^!T~889U3zY5bB$a2XIjPt>t$8#~eOSR3`97<zGL0e_UaXj$6n7
zqP<Cc5&fJ@;bcm;^-0uX>)29>MA)ySCxLqR-*M5*Yvbqj`FY;0>)bZW>=9-)ZBAI)
zMB#YlFVZQ-F_(arrmQV(<>W=X-;eWs&xH2;vPB^E&<DtvTLB2A`8qQx@Ap?;PKS*w
z7A}Chg&6@8l?g1;Xt%&)ptc_~PiX8vsGIVVdX;X<M#DixM`ao5i`keU#B&ZXbyfv>
zLFO4-+S}(MP)TS=Ndukt)oHutGdduvKY$LC82}|!=CIjy%|HjR$P@splk$DwH`w9m
zmjWPziNA-j9bSA+-Pr$u8H!|Dk>QrsSLT&3f1AYkx81ITF|N}EKE(;8a#f8r9R)(V
z>+G4GZU9e)ovC~4E1a9*#jth^vfq}cZYk8p_X}(TD2p*c%=`OjglLm~DVch<W=IJC
z0^~zs<4TFJ!7rX87EKCFwduw~ruA13Lph31U*3*yn4*hcQ3awl{oX}8xi<4lW1xCl
z%E?K(UDvMbubkFZ9~FHG3FHWWKpyIuVx@2`^L_l*9?t(bebZ`YK`tb&TshMi63r^E
zL~>O!$EEj6iiN7OwNBBy%=cm!{1QUR9ZKn=+I%}tomcy*lUv!M*{1EQYqFK=E+ix4
zK)9%z3`kX{j+0Kmr4gDrkz-<k>S~Xi#NGl890T6UrFH=?$GZ|9bOAL`ve|OWhzg!R
zIe!mwy~mWD+M>&Ox`p8A)+NK>lI*om?PP`u`YFmS*%tM7wv|=Qtk?6+Snd}rQS&ym
z&-Df@!JRMnJ$00ct(Ul-ocle8P(M6>t+)_77IyBKJUz9S-*Gj~l_Y+A{M_8UXc{Se
zv_d2Z4g2PO{~E@ppj%g`DEHWD3L-6(g1Gve5L30^)@=7VYWIFb5jH-%C}Em9nb!<v
zwdsRJ(C<EYwq$Yli|@-hQEWEb&$dQ;MIE+bgI_RmI){jZfgB?^`h9&KuR#&P)};X}
z41Ly?%hjN<>>1g>C!95n#!Rsf0Kj8YX&lU;6^O4lHShe0)6dJQm3(8p1~=4nLnrha
z(1+XZ{DJTB_C*7M!w<ASk@9GokNV47yB_s)IvK{znksxSHt43VYy1ucz30|%yaZTr
zqDrV0vu&tPPq0_M6k_PV47`St?l>H;1<t<KWyeCD8*X%cA)m^H@H?5O9TX};QhpXO
zWO<8g=2@}8l8Iit$g?jjK@0;6z(>^<3l(K5sniYRdf39vm=CB!N_|Y22m%KTuR+z+
zJ5xhD2)PeniJ6#~Gy*IX2v<}BE_p(Pz#D+ec_;`J>^Q@35g+6`{_9h3w$%YzAaT?*
zti6*)2tes5Z3G6vf$e3?xS;37#YOxFFpVFnGN0wR6xX!XbRI;L!)6WuVl2;Q4159n
z53>NKX#f=(4HKmQ3~%t?S@(qv0*R_Ny(ujRuKLMPNOAi+M3xdSR?)*DG|n6qs?pTP
z4qixk+55Il%%Cr6zc$iv!t$Gl#m{Qnbk|BLh*$x&I=p11{tZZH21&Ld-_bg>LX7!{
z0uYRvvit<kbwYE)wy9B?FMP20Ca^Npce0dkutk6ry%!?ML%_#_QKcC}1XprrM_H&`
z=zUX>4Sq<V|L5k(G~z{_bS32LC`K*zqWKS2)l{$9UCqMl{FKU%J*%yO2;U{hV2rCd
zMU-nfH=zc~F87tqtFK?{zEu!4J`lmuVeaZh&?D(qcTtM8-Vd@x`!n(5$`;;(r*q}c
z3YjmXE6a1+tZ9*}5u|4bf8=_T&aABveZWG#GU-*Mw?TuWIKhV@A;g&C-_3Bag`eb5
zbbr!@yo=<ejHi^`9&reY77>+vqa(6|V5Z)Ec_6{piNttQJhIhrrN;mCkAPN9d&YUQ
z-1UANJHM=KrIFEqzbh^L?o3gL6JJI10%Lc3S40^Pbm`%&tYaf>$Ee$y)8+pn`_|@`
zkhHz43Tnmy-CusrKe{k${f7sbKPAtQi=M~rXD&KO9A2<Ee#QOVb4Nqf>uc@#zn}Rz
zoNjYh@#HUaAM+N@y?0Ilp`qRFBkRZ)-Y)IKO+?>;o5YWiQt40?*_DDPWXNaG*uqWX
z#L-}Fx9?jjTk(?MrbtkfZLfpHPt5~b*+CaLyQkYH#3?WGDG6m12)jG@dv+U7CB6|X
ze!Y7l%wblbtjq1{EJ9JWrmg{uD?HXzZKRNmo2JEJuEV^)fTQd=xL4N!V5p};x)BlG
zk9fw?m>tljVd0q8F%=aR3s!-{9gWMk6bxBwsjf=ZVuWI)nzn%5BW<NB^T*&u;N21F
zZ(@L<yeDF7N51K^-d27{4*;v_v6r;t2zgbfAAkbXR7YRnj)xP4WMpJ?JX;(sxf0rJ
z0_YoSY4FCEuPZIe!1gV$(@SU!Hw}9dh#xIO5I0y0VFq(QyFd+7hj-7mm^$@OdjeIh
z7VrLBsvH;x(<F~}Dy?SewvimqOLMfm;9AH3z78&_JuZcaF~gdsE4uhH`%y@7Mo<WP
zI{tiw_Kri6luP$e?v{sk`Pu-3V6hq>!;wo4BQ8Vkr+1pY?LvHR9a+$Q-Z<@c<?~|?
zBDC-H$|+euN6`*xO6|bM2>$x+zc-)3Rbv`Z-kzlYr49BL-5tc!N-N{#?q<M!<a3%!
z5fi0`F|FtE*hJd^ligG~QOIkZOkk`Gvy;nnSp0R1W&Q_O7BO|P2EyIH(uT;|C@tzS
z%_Qph+UrCi{I4r~S}rZ^82&?1;8j>5q!LvL4W>@VZojvE#$Is|Jto6Wrpm2rA1~Fn
z%!L8ej%Lhnuj2lvf?H=9(M&`eG6c<Mrc1lr>J{3b@wwel9!4eBv4k?eF>1>$(5Fz_
z#37@lK`%8@29hBf4|_+T3*7wF+cSNbAxTch{!;tgR7C0?VYDpeyl0V`5;xjJ`EbYn
zw|;=P?hC<loDOfzv*eIt8|7r=vEbL07ips!vu%NX#@A0i$ony2x4zFqy;-huSuUp}
z7bbE|=95?HdJliv9||I|<RhpftD5|K2+i}-(*~BGq}_3)`lht&8f54p|M?r^lhE&9
zMTl5|prAa_Mxbw!#Y>r0P_(G)rxN0H7k<eyo@nL^Xy{FMBOS{HzE2brSoRdPs<*0}
zWD3->yBrT2#EU##nHdN~^u-D_-pH>stMiwfvNxyr^}8k}Bm*{QquF9L2-oN5OaZTf
zhy1z+aOLF{yC&$J9F2*_y>i(F2eMccs8xd}20MGbq=vY&{q7103CR>Y04xZ>nXrq`
zpbKuh)h~vIhTu3qX#gP!22R7!6s{thst-Y)N+U}00tQ-Uto2^qp&^I-yI<wsXh$6+
zyX4pLZhC;P*`!JfOWXy<DXgG0VO_}(lF~D7ftU03(~fcXI<p&qNRT6{x#jR2796rv
zw?(+_cx;KAvpl_F79tLiCjm|*!?$qM3HN4&VM1i7;*)~D^pIPoZwRnDb&aGnXQ5D{
zd_SR=-l|ay@xAi;MQexjCRUh7KkXkxd{rW0z4~$O92y`WvP0rKo4`KtIieZU#nshO
z(`ZkG(Gy)%FJ`Yb?>E^NlV{!M>on@eoBT?aAqjc)_s9U_0b6*4=oahz3*#Eq+AIMN
zQE<=TbZ)onpKS|c6xduMKKrOi_l8!!xuebx>)lpf{yHB=)i!tb;O3*pBbS00CXs;k
z)U!tWaTJ6$Z`t*@BCH<R``gy&qyH2q;k9exM!~e|%{-jglk)N0F2{^uapJkGZmIIQ
z?c=CohFi0!ObMat>&Qld^|d=Ghc4xmrzI;hIaIp7MeaILG~D9;XCK$*Qjf-#dy`*g
zkq0TxhKw$pk2!Gp*}n~~IZ~PB5by`FTuQzJd+f1b=%Rm)_Vv5L_*Dc>7E196VLw^(
z$cSLjZLn~K#iIM%A;vab@?EEF=lHUk+r3as9BCQRWJW875yLIumSG50GH4TRU2BLk
zX?0{YMKM0-k-7WOX$q5^&}Oi+W?0W|n7(rB!nbN|D1UZ-n}njv;w~M567FtS!+~Mp
z6t)3qiXO)kSbUu0DS6$nt{CPJBP&G;gSP?>gZIaFFVN}2%^v*6D-}UQUEBTP#SBr5
z3yS*0YEn{CGj9O=R2d+;4)O($>FNU1IE4^c)WsEBpgk7c!ZdYzdyDlGK=V=pJ8vAZ
zpezldvJsO^5=B065g-2nPR|9bm@2q=e*sXvgtCzrVE$lNRZci}Fz{-9?6NO^A<fr5
z{2QUS43#DV(jPI8GqSYO@#PDpIz^pcsYG;Hzc}BmN`wuO#>OdB)hEA^-?CAG@qJt^
z7&1-Y=@#)s8kmEqS@`dl{KQq}9&X#Ml${Q_5@7}*LfrktAb`tOGOH9%??7H%^XoGE
zWBb`;VD;T*(%00EJ0*^m-1r%eG@V@J!s|5=6^$Q0WV7F0h@~^1#(LKxq0@s$s@r|0
zQ9iPev9`V@X)atd3$=J#XD&f}wI~=8)-UC8+1sI2>fS1e?SN2E*-?38ega<5bO|a`
ze=asy`)9}dn3}cP7jsedrSUlTb#Kv~duUZPl=t>iTbenrUm=^lRhW+!9m+F*e8^Z2
zKLm4aK>7_ywNf65DusJ&pcI$Z*nVb83F~ktk=hEyUZls==Tz9nXsk<|M((z_jUm2k
zHxu)eENiV_Uv5&X$Z|0xGVq6Uo@4QtpAq5$q}y|#;1D7|ceblmRis$13P)RRPqfaL
zxx1CDwK9nK4l`s(WJ5_Q%_h|SDAUe=1*3*;hz~2nw7c4l8fi3!2}i`|Vvvv(M`Np$
zQr=>N&?}?W1}4yZA_l{@JQ{b`Fuf;nMz$v7;<og+veN0tyf{6;8h1)C9eWS@WEBa5
zeZOectW&U4X?<!>y*NLfS&A>N>Q1@umo^8ZGUeho=1g-olN0|Kh%NJ9m3`Q(y6=R+
zg)*w(OIi>zNVH;R`wb8A@#<&61wO|VpU@IbAtq7D$OFP~ou?MW7;37@=7@*!4@vU@
zGNoUtZE(dhDZ$SQx0IVIof4qGDS8hoK()%kARkJ$ys|XX>9Ts@lZ^*qo@_BfqS}58
zfIC^C|6L|G$5-$=*<Z@zuOvH`J{i7v##3fbu2Wy6-f8P}*IZu^6BrQzYd)1@0FG&D
zXuu4z1Wph3S3wXo!a$}a`)qZ&3?%BcpBwNl1E6$^0c64~fW#Qx+{^_ijFQ(~O;&T^
zc?o!Yu4S3HM%GXW2#)6=>?Eoc>a@Du{+)2!dR4ldfYA*PAn<<Y1M0jqR8q&_@HipQ
zk19hg&et$|=@`P1@gvS!9Mlce5FnzD)D~qdQW^;(l=alLV;C8QAJU`Ri_97JQG0t)
znN9Hw#k^fWdUz4{w@yXZD|TQUcHcB%h4MYcpqN41EuL7rD9JyrAcTYQ_zl&~fQ#8S
zO|LJ?YcjA<&~w4=mj-A4bbo=*XNOFzaWu20HBUyL#t&MClMZCt&V61f!wuOvTin1A
zxYiK#?3>iI2&lEOS10?&aXsZ^;D}F_vSSey^J`x$h;X8LkWRz6FlGYrH!w;uF^PAN
zyMncjbF}fB1~J-fiz;ZX6*;>wMXx81r=~t%Bf_Y%Mj-#Mr<WM9q!npRd`F;OR`c4_
znrNgWoY(R=KVtq{Rql7XDR1QCogEPg^Tc(DUhPX*Gke}ZdH;RW-ri3$(2j=HnZiF-
zT~DNtbv;UwX;tA!{Z>|Ek3qWESBW)+UVy)^VwUUzFPZx>gRFrnhT<cT6cSYk+-v(_
ztF2fY7y>GqN+JKL&7397{HF1{1at<mCa#~IZTacus9(2CrZZFj9h_40mbUfwrRQ%k
zu&tA}7qEQr4iSUAoWCtO{!6>wUY!LdFbp9DP=XhoZ%K>#zOo0O2Erlh0(4;Mh)|+>
z)hvbd{^fXHN_t*n?ema9jq}1q-)Z3X9bUdIsn645S9Rb~2?X=u{Q0t(F6D7Ps_gx+
z$G>nbPd&Di#bLBJd{$E2fRvt#;!A(e*^z{~cL3UH`$`LS{GVGj4XNT_hfSQGNA;WL
zS23hJ9WA|*qPKBZv$U08qxxq0S%_JfE)u*oi7DS_KB<8dE;MIz@V)*@l(x0SnNR&X
zT7nfh@+u{soBYXqILg09Kx?VLP-*?X{6{NyM)G}q%Lmt$hjW>%)8vuWW6e>k2>Nn(
zU!}{##mWD}S;z9}iG|PcAmw{Id*Fow&1;RI|CAGZ@PKF4uFo69luVR`f`VBEkBTnQ
zZRkZhj^lh6=~ZQqK%pv%=1}S+*<owJaf7#GVcqrK%g2m9G3p3Oq!%Rl4;x#?-@z{^
zh<+~6N?LMjQ=T8PH&QMyxAZ3=eczsTfws)c&0DRh#N;IG7f$-*JZ!ytgm}Snhgn~7
zU3}RkrjU9s!L>o93fu*ZTfd+CF~zz>uqwT}6;E(w&f#uW<<;JRWp!+CpP`wbWB&=}
zW?#!@PAMBT&);$h$Mj`wK`)rQdm-@Z*tfFC!tmn}boi@U>gJ$7)mq9dVP_d;{}7<z
z6a!0L6_^J7S7E1Ksz}t!oXe1b03P@(v((y{jH0%hntENY`iY^lvvW`>=bI%_qYyCD
zC<VCK7?pQ3LW@$d_{G)D0U!QNT0wj&Q)9p+YOnz<C?KCAkTm@Qfa2ZGOY@B4={-F~
zpC<y&3O!zw1x#&uan$41Ba|kC!l)xH&QivzcB~Tr-0@t+Yd6A@`)296#KR0;%74I$
ztT`X*d8M}Q+72(Hs>Op@qqKiDdoWVDcT5G-fKoZfB^1&OO6wb#qOSLrx9S~q4rn~!
zGiO=_U@-5Yl}C>bB%ZLcxEc)Ioe{K%zpO|TG6(V|Ql9_h5kg{JeEYcCU3k@TjB!)4
z;;D`aRJszIpFAT^NrujQTkso(-Dr()7>Q-P`@v2^Ko;H4)^J~N;%Xsn0`nNqDTKIq
zzg4+7xKtwO&qEyO@M&dRSqshZaz2Q8GsH+}m88=W(YM9Z>GNuPQ(U~;+1p#x*_=Ej
z$D5=npTFi^eT+Rxj&Q(ab;&u_zQ^Y_h1VqStyp28C&1jp#hh4vAOAq+8Ghrwd?vy0
z*$5l-!F-f%TZ#%vGAo7BLmBi?Un%`7-$aQv(Cfjjs*3g(Ih61#D@7VEvysZ>j`H=`
z-e;vakCXe>=;=LrEjXs}ed)|(P1g!JMcKms)h|ZvdF$6X&WYCh_4xu5KTuf+QmdDq
z_i(2MRU2v?(ew+xL9<Q6#?7(bjgaGO&LryMPD%%m_)4?Kg(*KLN0xIzJ%Cmqu&6AQ
zt8%pX5X@?wes9X@{#^@^WkE~#TjPa!3Q*>}7a5x(r2u-OYr3Y(&%KTWe;b(YV{`6J
zuKkF&;s^=B*5=g#BS69IDL-gyG5-@Ss+lTZ)|1ggQZPeOAjQq{mnXjm&%AwxrsY$o
z+?u`3@(mMmPLOo@!dvq+x^@$Vhs<O)shTZtbO>c6^<7V8QV`jwv?Syum9IwYF~>o*
z0rJeoII{c4|L+3Wv!30n+daj^ZxH?tll({QqWe|qmv+W}FT(qtg7anPOI!tWhv?q=
z9D)*<=EI~j&If6%Hi@12hcMJl%SXrK4vrf{Gk^V^j+;Nssj)*BE=kwi9|&mh)~06|
zi4XI8JRkPdfV%GoxltE0B5zUA!XJTU>8->}XV?e4d!P6Nyx5x+_<o4ws+7N}B@51!
z;(m)lkB1vImZd{|Cw;OgW}nlj9lBM8{3jJ^_Ashk;+iX%d_-=R>`Nh`CkpYbwY!Nm
zm=a=-weAu<SSPCh!g2`E$gQ|i&zFE5lA=ji1N(RH=Uq{)%U|f84|-IDw|4j8TsD`U
z_1k82L*BM$a78z4Z#?v$xI#Y|lh~SH`o8jiJ0584gxpiO?4Fh?`^ED#FH6lbl%ZW*
zwHrb<iS_RuO$cz#wwBUdPClYo#XdyZ%^VzRxXY?(-<|<85zw}6IALI*uP<MY?=UXL
zdi}5&fSI~K0)V#<N8g_~YCC@L{uy{~gN%kZWe7Jzmeron*{JT^qFX1@WJon;&Bcp;
zngTQI3!1UYxS6u5WG6(#^plb@<9!nOnWg-7Hb_L)pn?*aL`Wc?@LEZ5a>Qsz*zI9#
zG8z4QD|(=&rvQ|s491%hI4sQVjisb2h||v&bmO!ew@8;0U?vGMyfYe@B?LjbT{Vn>
z<94b_g)Z!OmvU<)3<>IhZIIwEbC*Op{j+LYU>G6=Z{w|fG7K|%T=Z#qSUH!i<Tv^2
zZDg#QZl4nRjwC4`j-b@C>;t=!I>ByJVX&>A(GQ3gcv#~4;0ppBy*CmlKe}kYz&AAQ
zC8#3UJmyh8nA3bb#>4a^K#CJ*OkcH8_L71j2$+87KL{&84+P%Yya$Uen5lRv>CZ(Y
zG2?yo6?7r5Ng&_sUSktr!kW1$vbO+HLk@ax_uQ;~Us$rR+tN}+lC~R9bIcYTLkLdC
z=n`Z!bt=eOa!3UEsZE`Tml#5%5GE^1@JjX<M_gFK>a}dby-AbP^JloiYu|SiaL(;c
zcWMRm!O1ok`7&>H{q=ZBv80H>zG<n!;c&!VB0m2-{pYm%MRqNXCOKO;>%%TvZkJqh
za^Gg^h@^{}^&yDf>w)osHq=a4dUlB#5$c)lw-bG{SV#e5U?<V8^W}z3KHL|63@Je*
z*b+kjFPD9+^A=wG+7YH8SqX#LhGan>v~M#h*gIC~<oK$fAx~J-6%^xqVcS(P**<lC
z^0p=nvpI`3!sZf>mpx;laD=y;$s_yuvfdxlvAp-|Suu3gSgMjD2XMx5LXKM^qD@dX
zZUW3AHc-*eM#%ex{f7V|@5%LEkS`MGx+~ru#9l2wjLBA&;)LvB8aneKwTypB^Vs(Z
zPj?>96)?lb!Te`{x!|Q{!dl`5Ir7)FmdoGh+Iy^x4KiF(1mYGsHZBSGK+2e-y*z#u
zQ{*mf&6s}v1HQ*c(_lvkwEfyk4~s@EWxTrlZ$4mvr$V_Des1MeGVFS<cu!uV@H-o1
ztu5z5F$&IQZEa2WDNc&B3}YFWH;W|<G|5+Y1MXGFy^Opn5RlBt^8l|B$)JTm84H1(
zcu39H)UjwCKPe355eNFo7~3%ind&UlEeBPFn=CNnYM*t#y^3Ehp{`)(O$OIRecUL4
zm8X<*7qC5x8T3p_<?@fy1c**(TQ68?d2qukV`JBUI~!^&75H|@3)aXy71puqWpGcX
z<z#8A#BE&$)ZIPxvFDQ+v~BFYO#ilDxv@SblACXqmb7kMi$_rG{G#ldviD-==#@g&
z#HMy?{%yQgCL);ja>1ev=US97d~Ip*XVp!y5873TZ6!vz6~x)24k!pWgkmsDV~ocZ
z7g1{>h(wcViR9P4OICU7WL-V@`uoMsg}`L*8mhQo>hF*{!{W1}z2S@1QPaQVA;&s}
z?Cjx~jKKA&fBm)mj}q~G7>r`39B>^n!H3jj+Plgd#DsmH7oy=;17yfp(4i&ZX(6Yv
z%%cG&Lo8JQ;GD}%<tUmBvd$Vjr0!Cj2t@nd2DSR@($zWI{qMCGq>9<c;0E^4l|gSU
zzW1gon*lGI4t!pd{%*Lv<lU{YrFl&<6C~u*mAbm;f^X7WQsV<8D^cm(kke}+biMiz
z=|MJZOtL{8Ij)u`y1-|%-~6GD`l5!3Iefc&#E#%|pT&E^X@z78AO`Cc^ye^o?i=Ds
zSb2=A+B11?O{#=FEkLm*h-)r_ozO~OWtyWRl3f)TJ?~8R7;ppcS4~IyM|J7Z9=S^p
zpkSNz4cP_LDe|}zi<st$N!`Ea3+b_o(Bje%P=tjD2zJ#F;FDhlP&M8l?4fr^QuN|&
zc|(Gb?LE~%e&w-7O<=cwncsLV>F3OdOvWFZ<PD&o^UMSG^0MUtD)72N*ZvRqLA?mm
z49l%JSbPJ8a<{3HEsMa<iJJ{@B8X2FgEAjfuQ_2aukj!150#ws2>81S)ceA<?@LP-
zXMQ2w!b0N?0!st>R6^44Yb#bC*GIRJR)ZJ~cg&UBW~yEOZ<0$FLV326D1>TKbzd6?
zwEsJLGE2)fnMImCr`QE1ogB^-3P^oWn2K4c$><h1`1||!ZIjINMh#S>qiv?W#&By*
zEBbmVec{A0yf3$u)2qQ7G8MYir$`JQ{Nl#uuOgI}ZFEQU8jbQ)daem^8*uD6yqWz5
zOjj~&d7s&($oXDJhVE-_!uiz$3xCDG%?&Qov8tj7w3<CYvw<$7#5VmuTezPb_d!7Z
zI#9&53V>=}x2WCv^n}b*F!|T{j+XJ8%@BWOUJtClIST1oO_o_>OYuT%W(y>J|3@Rb
z1m<56C+&{y%$V_{7^v<yAgSdI^(U)4RLr3~`0lmx4S!V$u)iIx+@z6dS9cKdh~1=C
zi}gRdmwIznj5U~^@l{*NX*)|DfJ8;FzaL4X!C`j$ObMxm1ysY^+kw5j2MJ$%_&in5
zV>An%#F0od$!lxVz^r1#?7!W(_!W#{ID`{|U)R3mlIH!(=zD#;x1Dh)n=*LHph}~(
z5Vz{<HTFvKULV{?U0v8v>m<t6M2b}r6R2rHL0R)t=sI;u;u9De=wG6vcLDQ7%dx#3
zl6P^d&L(G8Z?|5dx-7)6e<WaI=7LKF1<+93RiW*npp`z(_KDX0x-9b1g0fIN=~K@u
zS$gzEFphGdv9+o%Ee7wVQN@mo)5zdI9kW~d@6H+@DYRkOUok5U^K~$wE%Pvq2M}`T
zpR=N;#Y-If)C>E^^^WyVv(U}tVV)WArO0B9ke|J$F_bdt%6*&*SJz#Lo#wMx8C%3Q
zE_da(5p$*-&m1)ruSQOGnVAuJq3O##)Hv-7saI;VW(xW=7ypdY^iekEu$&Hrgo5HF
zbszW*-?9s+&P6ld63Ha~+bHm@uB^m;jjq`@^wD!N_BT`a-FHeC<7F61=DYzTjOrW@
ziYoyDcnDRffk`>HJ}##Ji}F;C_hn`YAwtn#Ld==<$*8dwJ+&LLjQN{IJ*ADg(?V2u
z2(SD+%U&-=M2p2w1gu<yrFDLw#+zSVA+cv>`b?#uXQ!uAT0bpG5iZD(j34&n3l{S|
zf8#<hON;Nc`pPir39jqYf^X~Mw*NW^lwu#`T6PfTfe-7in6NE8UTB-v4j}@zKKxj|
zvlYq8o<*0pxS6-BYNnunPegybKO?YPX-%6n>=0|fWh#skDxQDwdOT!36L!wxvi&Rb
z`{HTN0%b8}`iu5$y5}^UE3dT!noIJBqDFc(`R*Sr6Tg#5fxopm5g^a05i9{4rmz{W
zlDW%?uM+7ol2Tl(Mjp`nkM73P{)8b3wPQ3>jh~%Gl`l*u%fSn%4A)oUm+N+~wqgl?
zmU8_M&l)<2Di01?K6hsHiKi7UT;oh@3}taF3+VQ311I)GD^1`P#C6&Vfr_QfgZz~e
z9ZPL|GA!}8AHomQg)MJDW0SBckpcN&otnzM&S+%O;=;C~nip$ZcLlIt3Px?3?RQJz
zcL1tXC$iVaUVGjrHan$?*%l?y9Vn%6a{+EpdVHXrbFZM_w@q<P2Jj3i@n-GAW@stu
z`Nd~!!VDLqfZV+^<IY46?~<|Sifo4Fn+CESE%?83y1V}G8gE+nHY?~n-scD_B?<}}
z<5(Kpwz;hJK{0%4Ix3Q+Ik6zf_Z%TVT^gLezCJPtXTZZz0yvLV0JO@2R1T#a&RrXz
zwuy)l1WDOSJ14WIoY``C<sz4LIU=#ki1F0S%-RSrD@YWkhaVF82%&85O*xHh%&mnJ
z%S$}QW-@p8azsXknQp^R-B6vQqSYnqVyo2RSrMj~@Ba6y8Zj#<8c0N6K6waU)CYs^
zQhN)9zij=T{a3yATl<Fh<11XYu>uWoOc<{;G1(~jcRf-myo){&k2%Os0H10KY~eCU
zxM=Egs2I>jbgy#xN3rHivKdL>kKmQGb_*HaPBXz&=$?ieisn*&i#?hZtsX*k>tm>)
z|D)v7GNc=jPan&YG@k*TEQFPc@${_Am83kz0tl4C4Iz}^=e{))Senfen)b~bCou#T
z)O*A3hQ{MQOnkJQ`-E2e7rRdCGR{fIOF5aidaC8bVMl}UHPs}%cI*RjY!M?tvQ(p|
zzP#J{0Gyfqv-TrarBuX&hY}r3N!1dhqrpi{hc#JxZrgl%ZhLsXESRPWm1xPkFNQnr
z>oiJCsMn})MSc<SKeX&o@day9|LC<v{pW{lCbV^**b;)o>h`?j15z<Lx}4QYh+38T
z?Wm9=rC(SxS?zgrN;R(X$K%h>+C2(h^LI*VKI!eum+2d75A-+jR0xg#WpyJWeN*2I
zE5N^}Pm}XSLQJjs9QY2e5h@WM34781`nNbkv2++G;0o$M5OOEgLrG3}IR?x3>1=5e
zyULrTcBAv_w@f#IP|3RAsAc?lXe<<T++*J{EDNE2!;Ya`dc;>9PXu<R@QqN3u#XfS
zGtIlN6CF2R<t!^sPW7v#WYh95!DfH&5{mAt7tVX`Vdf*dlhDmvg;6I?t>saf{8iEi
zd(b<76D3kri)ZBVR1OCj&E?ES@)EOJt<5RJ;A5zu71Lyb@}@{|u6j1eEz!+htk_5R
z90vySEfVb4Z_jZE3hflh=6AfyZl9Jo`s9jPhHIwxg$CX}EeOW=)J$akGBv90%GFfc
zv-{4kQGwVKz!z_xTlKDPwc)0+m>~$g@8W0(oJyi0>xDT5?acL34PwtbV+<)=QQn9m
z6lZ1?+7EB{2zG6A6T@9d*RBT<zhk<-#R>n>Z%&NMYW;HLX|E#CO&xo@68<io8H(eu
zEc2q2WftwtY`7^ZgF#(F%y=@#eP8mk>#Ja=-$xL>I=x(>cO$yp!V!mdqtA99XWta9
z3ryy6H#zq880~e_$An%(N~+ze!jxermG(&d#avGPSJgj>fSA}$r1tv)TBrbspcm{n
z(f`!dpe2!#G6n!pjgJ4s_`I$u%>QJjpglI54n}QT0$wahz-C<x%!!h+i3k)^7-RrC
z1uo7BA`U|yfRIuK{9smq=aNvGFLG;T*w^iPcl7rDOfPuh4Y>P$(F=Hgq6Xx&$T%W?
zszaZw>}-j@wf0MtU0q#`mNR7D_otuoCjd*>NH3r^Se8w-t+G+t0sN-_ES&dQaUjaQ
zI{;iN2Yh;oW|i*fkCZzw2c*wi&Ybnor%Bk(&#*a|x6B6&m6NPsP{jdjxO3k}N{<>C
zNA2@3z|j@oKWVG6S!qlK6m8AE9gvi(qUc%cvXCRM_~+K8qbBpa$Y!cg>+#@VPB{4p
zx!5ucnc^`FceH-etMQ<z)(dL9mH*yjY6yxR?bDXut2~8ZFkvOJ>K&De8SFq~02{L%
zsx6L^X-_mIc-V40X_u<Z!+$LGU7Bh{ubB!(u&#RD@IJAAg9dyAtOOi>rR3{~aCG8D
zK}>92Ua$XJu<B0XDC%upWNpUaUbE0W`%wk7bC-4_90N+iAXqkm-ibD@JTywBSN<l0
z9ctnUqUE;N{-qDg<9`v5w+;prn10)5fc!tVkl+JMr4&1~ML3~IyOM2mtw%)f@{k^1
zw#1*?!|A)f3c(LTFzEigiQYH8<i`grWi3*iV`-v0fe`QHsEf94*UK}q{-YW{JyPP_
zrl<|YKy&CK^-4OhZ3AZ$Lc|lPjseP@9ToAIUS{EeWOl!lwdc0A<yao7DIi|!Ln$PS
zWN;jJ+;s!=Zt2<QbVnfkI=&+EN`i;PMMS}a4cn-G5~lWGoGM+Vozf{2(k|}Ko%oF~
z`bE`<jS#E96=8j{of2sq>P^9@u|#WH+NB<+>|a1vDEAqMfDS;=d|pIW2`;;_FxPEy
zX53@ei>>G5Q*MB1vcjZ?&vZYPQM%ACNPGS4dpfVf=kky3?rb)0ESX03$H7D?pa9NP
zyUB5)pe-4sz%Y(nR0}uR1Npe;AK#`*(cHs~IBp!+<}G5wJ^Pbw@Z8i&q(v?6-{>=G
zbu@}w&<I{jzb!;<KRn9d@$N^~rvJE3DLzLBrSY`<*0|+1qG47_$X^a6y&f&m<drZa
zV~RL{EYir2u|>ZYzM425N05@auH*C_k^X}^sRxB9yJ8A&?56hX_I2%59`sNb0LS3J
z=SJkU-5qh{XpD{IZ?6q)R)7Ks1u8va0U)t$T-?EFJj(eSglQkrZas(VVRWn~C(df*
z!<sm^y8<}@C7H}aZq7mEYBh)Z7XH8eKke)OL6mtE;o&MTLb(bra@Gv(A%#G`Qa}iI
zyOQ*ck!<TN84TbC4|cDyk=$xJNBt6ZSykrMMBDgAmqW2uLnuI<WV(TO+|cMC$MCy3
zl7%nfcUq8DO3_YIDgDV%Y|qK>i*m0`R`nX^Pxv#P*j1Ze-;dotiJ`_jkp22PTBZ_@
z{;K~T+s3Z^@o$G3|D-*O@QhvgfU1c${QyP}H7n0W>#gix?w;jK?~nQV51#4=*f)(r
zNI|drU@B&9adQNp=*`jSC98)ScJ+2s+T-<-e9+0S<|8R?_nPnXE1xRS@OZq4^+s{g
z?U;4Dvt>Dt#kAa+uy13YpR@SPzf#uF-k>`^3HuUKVWIfMXYZZ7a#!eA)biv=XRBju
z55jf$zh&ZcT7N+%;3}#!8^-9d)y@(2drq#btSkfKQK6(QaNl3v-+b_Sohcx`Uo2J;
z0OY#_E{6?y0C|&7rI0xd(D&$5%Ye`oI|4E?_Ex}3y}1O-I<Ong0h?PHU|i)e?g2q8
z0#$lB=bPP1pV%z@wSc0&u;=?TIc3PaGXOUqDbn%VVG~EBy-Ns~FI*{5XOpU(L(FHe
zTf~(iAY=nmN7+B$_W&m?Ee(w{4J8m<p*&9n=(LeL3p8c61Ao!cv?ut4EB;vxhlfly
zh6O$0&^t`uiAn%Yy>AN&thpD;WQcxWPtC^iNqh08OUgR8=DJ@GiOg6%%9zZ60J@GV
zUuY@|n$g&wO#Hnm(qxDL61bvnn>Iun4;mCOPQS>z1lK;{V350s?Fvu#zJULEq;aCw
z=XT_UD_JF=)TqSTg?a9Z7+QUkQ}3LO3(PPS$pqV7yfRFTQ3i41WwYVK<5aT2uHjrl
z{+h#bSNAZ$j(^Z4;3Grg!a&;N+vvk}iJXn#sZf=!3Ja+_NH$r1w6%`w`p87!&z2MR
z-p`|sZVpC@4TL--z$c<)UX1HOZ>IT}u$RwEqa6`(RhV++FbN<PG#AnVD}c4ZdBUh6
zS_uu6yc>U*PC>VnM><16-dCuo8cQlbHJy#97MGe1-1@(GbIVl9pv2?tA>M&udjcks
zzN~p=5U^1>icIm~%WFIpdieJf7Ff61E~=XjN&FLvzjd+2MlA*%|2r`K83QV^iY~&N
zRvd1o2h{cUy}1#w+D$GA4QBAyH6%YJLGI1Hv>^x6m;|mX5wy;?*ASSSVZmpnw5Zpl
zL~!-XTk{64{Tt?TlxAW&LC`vEoH8GQ&GGa$vuq{@EjDv<Zu5m8jXU00k6Y{*w&mM4
zZ!k=rdy_0=PsE+yaH_(W41M1EnHi!JiT9f{W_X;pU;9&}4paKO``hwB=}-Pv!tgN^
z-b^<v)brV(JHq?)p|G>Sx&>lDYo1bGhmK%CNoFB&Rz@M!pS;mRm>>DTajw@pvcgS5
zZ}n)Ic($D$`C>_2D>VrRiagMQ%)jjm|2kLLw=Eq5H>6?{3_7Q40y<b0v?+E!)3T}O
zV~R#sakFH!Df;U2LvP!G5*t`qADU@b_&AmAy0q(m#g=i#?`AQic<h$1u95jdCv-JM
zgE{Ctj@KP<zJDt<Dd{qqxfg1-#rsQ{O<CS4eO`)7Ds|Z4qu+d6>mwDRk~fD)g+P8X
z?o63qRtyNs%RHDWmdt6I9P~pZA8h3=gnWy#r4;Ax{=P#_OL>9+8nBv|rMIgy>tzYh
zapeOuL*=R{%Hn;vyys9a`f|;%q(<vi$3(+oSCi|?*{o)^--4cR=#^SpxlIs$GOT&_
z`prp<zn}~NA8)D{7T)!jz+!$BpllKProMig%5K9ax%zp%6!6&0()_X=h(}*;_h9yW
zeN6l>xB^5nakt*{L_@%MsJfqUadA!RaslL~Qx9nGx9tiU$j=IlLdYY+>%5D0mydL~
z*4hBN-mF_X0Xl|`0AwZb{7=Eqh_ir~cR*0N;hVzDJ5bhUxz!V->f(#)i~SPHX3`&y
zOh%N~3iPijlLihO8U6RvG%$HF)0Brd0KU@I{73;24la*G*nfJlO0Qf#gH0?WNXW3u
zH={)1XUXrJB=i8ut&z?&qzS0#QIW7^mRc$Twzy4Pk?&_vg$UFzqrf9m`%dQdndYz?
zqW3yZj}Iq$Kk>jb44DtW%TA3wY5dWM@?oI@O(lS}Cz7aI!8kVo5}%?o9x_sMyec6M
zp`tVVdA`@->FG{uhaG{UX@$-N&P;6{D{Ng|<5)jC&WWQ3=F0JL37HZ~=ur#?y7M&~
zv`(^WqNS59j$*pR^bOdCeZ)G)^A9dQ59GZ$Sie$zhj=@Q-<%8LXC4&5;h9%$UNySa
zQjD{)v>B^Tu7|e;V~Gd{WnVrzw_rd1i$R@!mb0(b@WZF;UtJ?ECJME|h;wr9v~3JB
zxah@36?$wyDnj&dT><T)E_rT~V~HA%htA#Q!GRc+Z=@l|xksIm_oCUBUQPWYN1$<>
zK@+*`zfaNj8`8iq*LkZly}b(fY#t234?$ET5wy&74L%tb`y;7e%-5^n1?Ji)C|!<@
zF~LJ&+<gM~EP;0$sa!dIgQqNWR}4IN>F;(YRxBfomNC;F;IS<2^^{DOrUWupZpih&
zO+{nKnQDu^lx5alBo+GA9I+&;N~dTzM>@+G0Rsb8yCjbZOol2mgd<hZ2zxgeBv56>
z7B=EXkFYExNkPIQq_wYUMe<j0C&z4*>L>3IqN>f>hHdWR45-BXGSpLsXeU9v@o>GC
zP{jW&cpHuZ0`V8u!UJYQD<D2Sl|@&aHYrcZsu%~#DNMQx$Pxe*&+$o#@Ye7}P)<@)
znJ^O$WEk(5Nnkv}snZ&SvXwOCk5#=Ww-!8=i$w-=<cX{0NNGnW>RdmN>fT)l={EYq
zt|KB*fu8f$6Dize=^^&b{x0h7(dy>6jQP+PYdFhJN<uh%T22GIv8ILdu<gZR>b?Az
z?HR)v50uKk-Cu|G9&`D=vGc9`#{WR(&fvUX<UBWD=okH^>|j!uXB<|>Dx$e+acvSV
z_KD073Y-ln#uN48N|X$+qfX(Q2+)Itd1t+@EXb5dM+fiZ4mO6!rB$PXXwfQD4gxYH
z^5PLF8hR5ox=m5QgrACGSCkW~EepWGmdA4|WO6`GNI_zO(58<?R7>P#Mt@M_&hwK6
zedJZUq5`EV@HQVl59(N<su{4T#ApJrTa+Lrkc>!x*TE#wj4+_wkD2Kd%VNvs3e|J-
zTsQUx6iHCBhy{Cy{^YU?WrN%j8nOk@Nb}tR0e3AD3vo##9#VM%dOe5?->x)z9k=?I
zC4d+;y0-iqVdKA9Oi@uNLNLyeF*rfaALT0|x;7F$MBssR5{kAw&tK4NV_mwb&K%q?
z|8jv;QJ`5sonO3JYE6lbVZX4))L0oF305*yD27_Yq;`#c4Wz!2D#{Q}D^o~Mn?zxn
zMF-2y*-_jiKeHC&brWCjyai7$Cd3}s&}v6aOnqPSb@DCqre>Ps-WHHq4&&VW$^gL*
zen()qy)ZNdug9vz=rbBIWMAA0fyjY3F50;PWvLO?uzpgn|K%}xTXS<Di7u0-?%!(x
zOZ7cS<jbXt%-G-ud5o57In^}IMYw0UPa2%RK$j2f5%0b*%NE3b109X_i8LN-%PB;{
zYTy^!rZFW?a#zNTFut@}tq>c;^xq^oAi%NzoeEtYl^CT7$FMMNkeyv3AyyfZw3`sN
zmQZ6*qK31E)=zZ|pE$_5IC}2fZQOFb`>)kSzS)1J0~M-WNOlIT2|xf;{68xZ;d!-s
zfwbuHLOHVAk+6JWuez@N9_`BUM#8eLlzg)B8_d(SghV^(g3sH2to9p_tPFXqIKjKi
z<gNtr*P3D`c{Nk21hcBd^O~}1sdXOUb0-iry??W0TE#AwT9CGwh@5+*fz#(hM984A
z-v%oUv?bsv$>~bb%eBQ)?b!0S;V<oDUX=fwa5g9D_domF_J+UoPQOZQa%zMMS32Bs
z|LK!A(#v~9{5B==9hZPM@V90c^6$r7Mm&;8T{IUSyPELa_1xu-?p^tK7sov$-Fhe7
zUfu44^b8KqG2gb%0M0qqH$|_CMznQiJ%-$^?vHpujlJO3pynd^`~Z}nOoX>-$G&?{
zLur@1k;(VZRmVhuXAkl><T5+Qd$ehKn!s%c+>nAO0DH-<nfEo&BZwkTx7QyvyGt5q
zL7)xn;$XS5gzq5|HSWzvweizldN^5%2P9^Akax2!-33<Q+t$L;eu`2G9UEhKHPHG*
zF>SOO&M9Q7KDxK?3oZn2`c-<i(MoMB`=?h6^6ko^=3D`;=HsJ|Y+k4c5`>Ine4f$E
z!v!^sKx27Gk@Fbadkkp%+1Bo$xX*N1Hvr9LGk)VHaPp3kfh7V%Z7MQ^agYb87Ozz_
zlLKK@JLU^H&E9yLv2nsU?vM$vRJd*5nBXnlD9Xp@Iak=rHq_vx5kVw`&s(#(Uc32Q
z$4ez^P7MvF3{=<c3f6P;l#h2;$DzZJmQj}>Qa2NYs7D)fsIQ%g+A#z@zHr;jsV0=8
zCG;jG=d!f9J(61oQ?;1OAo>Ubc5;y1wbbj?cjFqUF-&1NVO8kOaWW8tAxkaG=G=_j
znRmdU{Xh{7pI~s)dq5?2;4l7N^&Wz;66hAWTNr5G5sGK%Uf4aX=5#62RO3Z_W`10{
zKzHyiG-%UxP(j!8*@;3V0Ny5!nnfax2|{5CaAzmh^Hp_=r#+e1{kRe6GtyQ44E7z(
zVC};MF=K<YE>&yIlmcV5au8|*9HFU?2ckd{{*!Pq*?fA_rwr*rD)O!+k4$+%g;cYh
zun-(B8yL<`&PC%=?(iFv)3v4Z<rioTeC`Kb!o0j?lNU2;boRpWr}nNlilNvc>@Jt$
zZOt~KhJ>!T&QH$@y4~im=CyKC;m@b33p&YZ@9JfLlbQ6pr~oJPi@Ti?l#lDIv>KQH
z8Jz1rs0o@-)tU5fWp^Wa?XDFtSgmI%4S+&<j_m2WKFcX4&ru<N0MACpZ(35(OU7dR
zs*_$`RO>Mhwv!p9VsXUe`fdFaB6>_MpCe8bc6piFCCziYE@L>0vE%1Tjg~vW-uNmv
zC75p$FZ>;NY4<#jbQ~bXUU0f=5Vw$%swT!V3T*Fr`3+oL(%sa|4W$x;9pAOaDx#IF
z{$(gTZ~3@?BjR3lx(OTN67h%Ie*1q8ZlGCdWKcP$BkR7c%LazzM%*1nnMSV@+rn^_
zGPc1+OQX483e;;JLM)u0D5x8~Y4<@&r(Y___%HDH=#LLk=FN^`HG+dAEtCZp_X{U%
zWz2Pmj+?FgHfxbfGX$^e7@)Ip-b&kPyQd=o2vket)}n}qvUZJ+1chwlAG>V2$ExTv
zepIfCk$+>*bQsrK`#*R1HHFxd;6!$}?JSz?4_UV}nEGK5g^Tx}6V1YcdUT8bo(N(i
zM3_Vv_#6Al&HQ<bvE|aJ!&5Yk$pFKOXJozI<ELBYR!<r@^|80NH<{f+kyeHQJslSk
zf%Fz0aw}lqJNoY7p>+5&L+QC_=SqV`G^km735SLr?R61L)Q8{W9IFmyD;orXj5@F(
zWEw|>kt}@W3<e4pmxlroCbwag+g?wj8;^a<@?N2F8h$r$o3YYm3SIX7B(GgeGh;2y
z2oTb1=`EVCE^Hdxf^~wJVFsi&6lh6fR(Z3z3}lj@&LW1!tO}_H7C0dZ(uB{^8FK-B
z?;|d<jYp<Gv)nz{{(QF@Unp(7dsb_k?}p1nGwml83N4j<{c{7m?)RM3(;WCHlOSbV
zOhSSbJ<v<@U$G!?b{q=VwBn(X+TwO!2;=mkE6EM|r4ARc?7vs~Ik?!jUo8<@l$4!K
zq+#DQJ+I|zB3aC0-pTA4o-P}`qq}*c7xj`r^ue2LFkb(%hjEA`Ht%_~H}JneaIhN5
z%(4yZu<$jI?<NN)9vSi>{=>^5e{Mq&(gmG$W^5Xz^+*`%U0zD*L!5pnF2i2^7~3j_
zsvv1)Fs1B+Mi~RPrtR@w-#H373J_z5BSp}DOHxE^V>tt(0OSJ+cowI5AeYOYXYbFM
z=;v3_Y$`Xpu>t$KPwdNp%We4f`<xI<jm74_oq>(bE+6K#?!vhg(ZY`sh79%@!4FX8
zZ&z%uuP)?T7SypN6<TysP?8F0QDJrB&$B;jEldeS{|%v@D1|>)&Y(FoTI<JGB-0lH
zgP~|F7CW+{hD2~GT-y8xWuDo_JX<N;ybtFW^9(9G@*eIUPcX8<1lBk6Wz2D`41bPL
zh&L&Y$^uNK%OA6i<nd+u1q7}|(P^Mdh;9CORoV`urf8SXtrG-vxEertmBzu7_aUrx
z!#xvh`|tdtCVi{TY^{0IJ3HOK6CqL8?TmD-PrHTbh_|#qBnoLD`**V5GjB2-&zJQD
zE>>m~2mKI5jJjf5#HHSh=9&qZLt&Bj8IthEevBRJNxJ<ZJV{M!vu~5Se2$d8bZr`V
zBV6z|H4}=__|HdSwR}<|*@{7Ucu(v1b|34^G7?9*+uJ{!JUXcft`qb-g7G?V)$+N@
z@Rql6C(Nr5k{HbpeaTnD{T`T^_m(K<t!G20F%qq5w=UJ%U|yVMKgr?rPPf|f0>@dV
zn`&4G-QD%SH|>8NyjG3}`~R6bs1&}ATl4>HZ2TDTI_7l8`A_Xf!}EAwwNsz|VP@qd
z03orx33z%C<GcdrnrxPEg6tu3puP@cY+qa`0TR*w&~%P*k-q`g&&FogW}BPMZL)2<
z&E{s?wry>-+1zZ~O?A)zzMtns@8(5)X6Dy*o%22CV2ju)<nqgaCZRA~ljrbaQs^}z
z^Si(<Mf4yM-Df}lLSz0A6ch|P8(mmX(L97)g?;whABlm-2EM&9IKZAuL}6n^f?c}W
z{##A|%(1Yrkn(=IC;;jE2y(tqBae@dSm5525+Sv|v{ttkP#`fV&`PBVmKb7Sa28z?
zN!U%EfvTc@4|W*YGOVCGE*6f#JWZ&D(OiLmIeZTJXytksQ^*-~URAlR?gNU44&QW}
z!VsuKNY|xM_w*a*I|fktKhMVwDykT=@)~1<M~Up{>LsSKVyHj^Kkiq_4Y73&R*auu
z6t5;NfY95dEXyH5M)W3<kZqg$>2VpS2f-_QAWN?BJNvI{3~kSax{(kr(ec}naWt?m
zAFI~&i-y<S7t(S#ko{!A)J$UQ3*QTTI#*prtyjO+r$e0L9Z{OeNG(giZf7L=ZenIV
z2V=bh16>COwkyNqh`=z>nW`#ac=$-Q<XY5f1B6??`vz44Jq2)sc_!`gB@}mDU#d0@
z@L*5h?$tDJVEPKq_bY{Z68nbL%`YMplBfF(P7_C=uX~ab=2--9@z=hdXQShnGMH}+
zX~T(R+Y{qyo!jQ$GD)#Hrv9f}vG*oUMd!1*0!~ire<`~6X_zEkQv}PTimOOO+!?eU
z2;vaLejpg9FuMRnyTRK&H<1o7)7{}RKZxfawFkyZagFg{G*&&fSbHo?r`j9#7xtvj
zevYugzI>)I;6h<&ed#{~O8z(Kk}-iS9}I4fyzv>ZAF-<P!%tGw(0AVQmC@*Qi^srg
z-RBX{?qC~PUH0g}!UWkbmGfX{ZykdEv<S$MpwrT-?;@o+@s(mw@?AWGGc9*sj)uyU
zjlVgl2yCb2U*iMUhV3z%86E|1_D3(FNip6itd)`?7CESnrDjRnCkdsVT7wfs-@bZB
z_r#f$tF>1Z@T0CWLD&b6rn+p=6$}1%TvJl;8nY0IIc7fUI*#sw{9v7em5;oh_ER^0
zr%A+7uFuadPy2?j6zm(O5W<7=Z?aU|UiN!L{3$C=%zH%0sQph8>=FpPzF%w9XAOsf
z@-?SqNCdMeBlotme3?4iL|G44_wM*D%)bHr9%BCWj~XR}%Zphti0G$qFw}IKH#3}{
zH~IrG$en@*$cGY3&@Od&p6?!=+IxPfZFm~X^lTP>pt6(l%lwEvSl+BRHD5RpOaFQZ
ztn6obuNd@jNXH)k_dd({AWjHH*cOxkBC(*?(C%mpE*$H1n`T99F#bHQV){ep8QD(}
z(n7q`k1M`LZq75-pv*@+^7)!o9mBi+X^4%P=>gkEkTJ&$LfqknOiBPH^RBPU0%+y#
zjPxYmbJz4NcSFEOxC)Y^Oh6c<yW%OxIt$9=c3`T$BLFGdaH5EQ9*$cpRT>I}YM`2w
zkQP1~3$XHFZCC_wu^=B$yym%l5lE3H$Q@$r>-@Pu=@p^nUDAO3c4UIeUJT`8X;P0f
zm&0*N8rO$MykCcVmzVM{5>myV(sIp8eSW^X4FSTviU8s6?ruxav&UHP72`_^Ec-Qo
z{<QZr_60nuVwi*NuxTyCgtek#GgHU-JLuUF77x}Wb@@~#`KosBaH*8oC#$OYQykuq
z)q9oRkW^>7sy^Lyu_bx@Malk(qa&5CMawsdxAq<*C2%;}Jd=nMTV*)`WdCDUqtNV3
zZ8H-^+TaI$Cu-*JY=2C!sQ!DM;iILjM8vGNV+e&B_FZuQ!fOpaj9NkqZLY5yLtYS<
z2J||E(BOB`+k6m#7$*5Wq2MQ_0v@=D6qsS*?rC=eBc8^U`zMd)GDTVaAndQ7`Y<Du
zhOh!_i|Z-}119i<rt#{b><Boe$<;C_czU`fQpfMui>oj!%s7?`ZEX``?&<5GEx6)j
z?b3|Rf3$U-G?wndkM{wAq1?dRi)p3V1|b=+T5iLe3pSn9!7Q)=QwGvmG8>Yl0=GF-
zAVo+1FQD?<jQ*Yk6Zpi@1p%|7Zc>9DSmieVuYms$7Jp!Zs=QOMANkJZ2cp#Jdd<I|
zJ^TH@N4mD_fP6#e9m5W8z|{&*cLX^Mx?f^I-j_~#-zpNkFIg)6@l^hw2n_fr|A2=?
zq?4_Lo1|j<N4mkr&ymZnWDYCz@f`B3H?cp8c_nYDn#BJ)L@d}d*orks(xxFASkP9Q
zeUx!j^d{|}zb^EkJut`6Ylnj6;&%q%;jgZ)+}a5pc{U2diLM;mhe*S_Fg7-WpcF(Q
z27<aado=u+n=&@5%`K8PP0uFGwX3?SEZn^+tG=9#EVlAO=k=9ONV?xM752N!Si2kG
z2;cOpAF2k&KSfg5=U7|Qi9lrrW~*cFyq_^i7z}p#4`~aW>f?t8T*8XtvF3+{YEK<=
zG<Ebfre=DV5-TT~GeF)xPi0?~NH!ZY$ml77oO^`jM{I+n%=(kJ<Xvt-zmK@cCSz5y
zXUt3sD8KC*r4zZwOWzo#`?F^tJ66&Z;=$SU%%ToIb85WM5QT`<ru(^o*U0o~&jz|;
znzGd^g~18?D!L$=78Q<8<F0o|%uA*aV5sI99!*gZFKM`lHM<Q6#IT;q!gO@~dS0XE
zuyqHsyWFW*knf++JxXG5p49XScgp-0a%~XhZ$Gc64s4@ImZf<(zH@&W`&Y*iqab#Y
zl(?;NQ=5=QRC@<|iIwT4RQ5x~lNTCU+_uL8slNFKyewab@}PMbT0hwGK-N(#W&bw6
z=}Q)wZ3J;#-pBW7i|sorGE|Q@Jlrdo#TGM2B^ZF&*TtHC@vcFG5Anm?b3RvWf6>A}
z*S#1mBb3}W$e_PbSg)2O$($;@Ii2-kLM=vue0#pqb+4|*_9Fvn*{hjnO#qYK=;+gh
z2uW6`IS_41V|HhnEGR$j=p4CWo%ufZ#LZFiB3{c8-4^?91q$U;3~Xc4t4FUKX10Kj
z_#OO3OMvD7h{ofT71t~)Lc)A2l8l*WW7hBXNS_)2S!34sXaQ_rAm9<|gh+E+ZX5gn
zd|ir6vkP=LQQDaw5~&Rj54S{pEq!BH(%4(?a(5V%%A@+3E(|)!kdu&*1P%MS2CSo9
zr}+r2%I@sj&cJ`Lis3XB#^9Xm2*Trn23#=CPmH<T$C>XGQD%p}Yvw`%U^)||OeRk2
zyIJzv0Y1*b*}P6St7|0zB6^d(#G%b*;9X5^I;3IKKyin)ZxC<T9ZaPrV-JhbMEm+j
z?qnxc%1!!GR<c9<kE7u~@|GyLk!w;zgK#V11i{%TUFKLT1MYf#Z(UvEd@~2&>jmCQ
zA2b<vN85x$x?y$Z)5<kRoVJ|{3!95>(mz3>Cl#aAC8sOREaipQ2iL=b9^Tyu_8@9Q
zbC*pdYO?#<3%F4<`bdm~S)(>Ano+cds+>hmaQ|r2FI`2+om^L#L0`3Q$CJ|d?ir7W
zc*lzp*XN%`HE_lJa<%pqL)s<;l{iAZh**s2Gm5DmGz8~O0|;~cXz6jKn5ME3u-L?k
z(heO@x`vV?yAt<IVZDD7<|Syjjo@_!G@P{`Tu@lUTlf<Z;i5lm(HOL<bmnn(l%{!}
z;uZaOXELF=v#ruHPQ|($dNIqu4}2&N`}pM%h6;|AwxE8~AQ+Y~(?|r(9-_vcP#Gc;
z^W3?LM>5^x_cMpS&Y$vrpBT(y4A^Q2VQEwR9GWMJyu=#~vVz_I^6olV((cn4q$Zm%
zWS>egIO}%k!9@5_*#5-Ae$(?Lhj_^G=K~{!s`62^)i-$fcf|}C{tborWn_Z-P3?a5
z%wVuK9PQ8I5UQ2O<0OQ$!gS@;-g#j+=r{Y)XA*d<vi#>Fy=Qj5wCp{y{Diy|-t@ED
zEIT1zm>Y(+C=X@AOXn}NB$5!ztNiT%FAcTMJ^{IkFPHl0ylb;vj9Fia4|!wukzeyi
zm};bgc&q@wok*ZP5}TwDk*uxG-{E12s4<5~)T<#-KMNkCapjC8J#BJeZ|KUMK>?FE
z4Z*G7Sl?|c`U}swi>UvFV!qt!W7yq+6<S5qn7ulV$v^g=f=o7pit|~@i$icF-*xV#
zDECTiL7y8b$_y>`)LCwA8|w-)HF$Ay_<nqj3b_kQ4&z_tO7aj+dw&WrZH=&#uzKl!
z)IotTtc)veQyWiVoov(5%>ZJ^Xe3O>@0x>ydnR`Uu&#MEVzn@G%U$N_XJ)WHZ=(H;
zP*(?vt=Oq>Z8(JMXDTIxGSIHS{do5xR#9ce6GJwUB`UIw8wxp~>D6mUSNV+rwv@m0
zInkk#69p*{8M+cj+idj%dh(VydzR;Yp><k^WUe|VNEdt4clS5LjsR<(oWA~sKEDK1
zW~ssk23Jv=_C6NczM7}gzf5uIZDoNS)B!>IegzBBB8c-{zR1X8)R*{tcrZj3?Kv(Y
z*Wxdv{&BYg0V-X^Oqm#lWWisLMys`HBz<y9IaN~l`!I;S(B$8Z&jrY@$zm_)Z+%nm
zRfmiy<H_)JoEq5-DAge5)F5hWYoW|~Wh<dn;WeoOp=`?&76)7M%?Ov2E1?Qwp;SZt
z7Bl32QAkqUM3Z8RZZ(;Zb8q6PlMfG3#FZyd{?zusn1=2xm;pATOlT>DL4vsn&@j@k
z{nUC4Zxbjv(}VY;&VCId0pOG~_sGg{`S{pxb&P)Fho|w?=8nQmrq_zj-RL${(y~G5
z@&3x4*1thhV`D*5)OI_dk2{_$-Ef!cGe9K~DAjT2h!@o-?x5l1JC8b}VDOqmYLp%k
zuWpVT!tSd(Vb_`8SIJ8Oe&0#>>bycRvxxE)o(79Godx)_6N)t%+|H?$4;ZjlM~L^I
zq?NPs5l>BhBYT6Q$5TmVdpo?uOG|!Q1$EzPEzI{-FZv7-VE%a3tmTB#qH($t&C!Ti
zRJjWqhYX8ISfE#@gG?PkPXw+@n-Me^5ZUr<ve%DMq_DcGx#|Dnjzww-m*){a71ruY
z>MIImr@*}7rTG_3)JaxjdPHG76)&izJ2K*L{98Vq!^eOU0_r<l7R=XY+3Pgbu?n3K
zj<gK?s-;LC9Sc(}dwk7vv`_RLo|pj>tjn&BUtn_l2QUODXk^8^q=SlK-2Zu(Jg3W9
zEO}}tHFHC`WH#d1v2B|cy&rw^2^G^0@|^_5M=i1!z1{Pkic&WDukqgu@8_#E2~7|v
zR>W>G4Yw5lEt|zv1K_s+=)qx7^4;F-iUko02Yd@M083GYmqxEw!(ne!=ZZ*~T4{RL
z!nyx2h@4QD&b?^5A>%89E|Ei@Y52v8g!b^(bVT8_Rz0Sw7d_=Jx@eS7_fRygD|sf(
z(7+O1jZ)k5kc^#!P-+wG+OS4eB;?482#uSaA8r2g{Aj|L7ScrHYckf$zb-xjsCCOS
z|1Zzy#S-Z6gUu}T4$~h7(VaJO%Gj5^l71{kxsf!BDbVeoOfaT@3{Jw@)bJR^-F>5x
z=@Rkg?0@EE#Uh5&06DH=dy5FP`r4($`<iZED(J3T4l%l@{9_&ndPFIZ)ZoSSjJICh
z5u`9^|27BUTGdf8?0LWE3N5{KFq%!L=dN?v`~)?(EJUDQ{!41T+>)x|wks+sn^r#E
zlPH%c@>;BS`xW&f<bRv!N5Y#4$n6n!dQAp@X|vnX?Egk2;F&ripC#`Ogs@z@s90>2
z@s0A3lq4;=vXOg>_3i_wXIs(pqq_McSHDNB{+cgQMuI4@J7E6`mpbqwh`M2vPS3C$
z$jPk4La$xoL8VR?6xnzkD@13L{|#)2JtfC|aQnC5fYV_}FzQ2+1V6o4<;7=UxI=xw
zU_^jppx7*df~eEl46xp7+N?y>V7E5I6o=>tD@#OUb%2jc)6stq!Nyr`C*NM+gQotl
zD-|#io5eYPBIVG>g1W_`U`w!(I~Jkz9YZi6Ag(W^pb>?odseY^99H1y)4}2#$b{5J
zUc&g}e$_B49_J+Zg3tS1>%<E0l<!}lFZ@6{7!>vTzxF9rm>oEO{UbR+q3SmYqqV*(
zhJAgrg$ZRcX3GfylREvVN2x_UKgBL^lL~Qg1vk@w6Z&M$zprxwU{%W&gvXPakXx)2
zb1EB63_N`lSC+K}A*rW8uB^dAB7A{gSaSw86bi^`LGDB;eQc6cYGE5bdGRF4WbrTQ
ze=-=1b<SF;Fa}TiOIz4MV%JBSvWM|Qx!$4bzjYgsl?gxE`eZl8@eB$itAGk_$2K=}
z*4rEu93SZ8KwV<dwfa4>{k@=m)4tV^6A;_<crqt;&fKqx2s)X40|cNF^Gn}Qf=LbR
zKpzVDAc1;4Bn>}?SnaFSYkoF3uk)Am!Un}H$PT>4)^)t3W9J~1o{Pz98))s~3vh2>
zUbD{*M;lpmOr_};Azd_;YYlKbcuP#ZA<E~Xc*y6hRa|#`4!87zSD%jm-t?=EdZlx`
zFY%h40(RUJ5A<V<<Tx?_+1sLGpl8*%@_{DrdKK+Yoeg4w%Vq+_NuPDYT{``L+1wU}
zEeX_Ez9n1_Ah`Xr(yT!6Ez6)Ay(*`~4vQ5C`hBkV%yzn#<>eK$#Am4j%e>qEOkhKN
z#?(2eIn4^68s@mew*MD$_lKW_g<kY2_`fUo<1`7)0xGn#YjZh4&5tkNBQz}=BOKrx
zJCWI_&%Zh$3k!xk>7Y`uuQhq+`Cc97@(;S0$I!*nd5s--I*VKDEjH!TPbm#Q*tTl4
zSTE)mT`aQrJq;fR<^3nU3Wb%}MSFFV^@iu~9~%XHIU7xDG=hD)Ve<Z`lAjk)FDtE#
zaLb?jpPBL#EB1x?Rb?R%B(gv#Jm-lj9#hSV_w{+402uvRj&~0hyROfm(&dTD6d8QC
zL1gN^xHVQvE^MhQWK7TNJ|CQM6+00WF!wJ`ofiNv?jJ>gF)Lu;cRMbYW4R9MQsj$n
z3<OS}0FBsfl;#=*(4)Js7irh%cjB;yA0EVU>;5yW(kQyUY9{OWA(YQ!X7tC`Oei%O
zGr0pG+W3d*ub5GSn%e`3sjvdFf$UY;vu<mORUUuY-C(1QPho343S>}oMM%z&-@XOt
zLf*gFZEQjaFoI?21+KReL)p~RhLe2bDz$DffVYue^K7~J?0+_2p9{z2J5o?Rmh#7F
z>dSeFc8SC3*C6Rj0HK7~{&gbt?P=RytGK}cf2#M^-!q@z0c_1E<0g=Zswut}(Q7#3
zCR(k*o)}lOe8^!#dW#8Mh~fhzUztq_A|-?dhcfGqMuO0o35z!OdTy~7qJ0D}$gJ)&
znhOdg-`)RsY3liDl6z&1U+FxhIWBtNzS#{(Mf$%#<+Kv<Osl_hTU}K$s<yZ$IJVqB
zyvkWGSzrc|0CxlO`$ws*{3fC?2AAqZQ#hWBF1d@9eK7G;$45e-@^!;QKSL2SlRA_M
z5ng$JohgF0%dMufe_#?4{>L6*jr-6%%A+0iffzwdTwG~L7^Hj<^E{tSt5WXuc;<Sx
zNL$b!j&9&gVH=soeoXr)@XaS%>l<L}nfQ>co{0)l&#20Mwlwy&Io8`~z@PPmkx>AH
zAz#ADYepHp6L<T#I?7W{4gjkHt-d-@v2Se<rT`STi!3dv_VMz3+ZEP!KZVyL))dW7
zeE`+s=f|BI&^C4B5AQBOHDw)@`f}%~bfxryl7HWn&IDHRM)2``RX^o<*(3{E@eV0x
ztwZYqB5;|4`vS3e8GasfO${%nl30>n19NwB<b>%kkM~c5D@{?FiaMhiCE~WcCFFFx
zEX|gU9_6=4<s}&rZCKpu8bWtgd!E;#-eA@5{*(c1i8Nvw$}Z(KI`sMxtF6vu^o884
z!P8hh!b$XwNjzpq;zj83aBy?ini>Y>0gC-ILIs{{;Z1xt@Rwgh-H#0=HHSs1+{}R!
zD?SyY0Y)6f3JR54B|+skfpKvwmh0-8`R<mF>@{2MG3xfKqJk?9ioN_Dr^hfY6?*qn
zM(Hy;orzB*KLA;2%(F)#5&AvGKh9P=Go-~bN~)Gs;B>l&GA{q@28TfC>kI$_)i(SM
z4`Yee4NRs~8hIGCI%3^CuDPpb{U@Zt6>__z<;($v>hk!%iF-X>BRiWmf`|pl;<DE+
z7i~t(Y)%ouzRTvDh9@cV)TW%QB-$@}e(=$8T_!3Xa5Bb6(wWN9RUZeoF>446j@f;>
z!iN_6O-D%CK4z;{5kmEi{c2J!166welZ>H+PyStG-A)i9HE^BrkVW-4oU1`2ynBKZ
z=`F@lo7%Pd&93b&^{<-=;#G?se~4Op!QIGw{h!hQ_np;@Nty1MC@spUZqD^TYq)30
zmihl(06a3Rjh$WXs7iy2-4I;n`bmTw3EcVPQ}g`+17B0oVQbAdoT<}EjrZ1TAK;n{
zG!M4|a!-)d*d}LMR7CQ}Q;m<^o?*BBBqO$d)7<)Ja#{Xk<Y7<}&4>=FznF6XG{AF)
z&gmN^Ce@U_p+p75?D>r8#*?<GxJu@Bs<V#!ih$6I@BhBZu;_)|B}<Ft#U%I<0};9b
z&zpu$F3-^evXbaUhzGGVQ(cSR2d;0yjH@=u3)#9KH@6r|R~*8coox=fQOj36!d31U
zW%S(FN?W>G6*4iAi={@IhmS#=|4w$s5S#s;ENfJ=9`Tu~5B;r_<%$a?@8rE{4wh`_
zQfGokp3fODZ`Kjq5PUx0Qc9WinA^MlAa2tV0lC~?OdznP&PCn<*-a`6YOM*3fuQm$
zm6cjOc<hT7Ypbh&s3d}L=Wd#v?iZ^GpooIR7w{gnz!sE<sL*OMA4JgayLLIAE&<Ks
z&4C!b7A+8djJbT5!}|-}ilzxv^@#uvUw}?2J9{ve6#9RkHxMQ~2WrLsyRwpoPPM2B
z)z#6V)f{_!e_xbDtq|G!bg`!Lz~TpdS!{PQ;dR^zSf!~~X;W)OAqz~MsDpo*MB@}g
z4mAw@Ojf{{wmBN;On+|P`6I@=`y9CgR6jAUG>5N>GLq<+9d5o`jIAD`GbjMstNoEL
z9#HnZ(Dz>+_Q9lq&XFxGJ0mti4r@ljM6&I9i+*HqfL0D_u`P(oa2ItU(=x!hk40vk
zZZi9PknAXT0!OC6ispo|?~)sE5LAg(g@5U}ObYmIKaAiHI&o{#fvXibR1_~dQ&E1~
zFT<tMpZTfYLc`9flPbE#U*9Mx^A%Aew|i4-644BS(YlbCC5pFDqRQ)Z0IVAlNlu7U
z__gEnL*PJf=G<pe7ofN!w>swWS~IGC$K&-#f8Y6c`HCq0c^;csH8chJ%~UqSq~?&S
z`d0Fc_Rq}_%HTA^$?^g!M-Vrgdh3{2Q4gee<^wb7%Jzbq$(Vt|E*9<Do(-~};@*7-
z$z&*1UR!%^BD9ecE+kbj7rJSj_q@^=y9P=pXrZl&%YKRO;hB(Mw2WD|5D!Cgb}dS9
zu~y2UQmut7hG^b^=A!F~Ko)Bv8KLMmDw3D9;0gy%<Gcs6K@Y&wkEZ)8vnw4YmQbOk
zO05voRJTxwRaK!H6L#QjEktQPso<-8ms_jaiYcGf780hAE{SZdQG{iy_qFZ+rF3{_
z_K<~);U0k#$3(Iwe}%`%(@r{L@QsUVL%FPQ<40pC?e$;y_OhA}0TvG8fl<Y<z-CCK
zsv-qwy!HJH2k29np(L<9NCZ*EZ+yccgANWWv%C?nA+xmxJst_K5}NRGvXT)NRJGSr
zia$QPt-x%IG#~a=mBbfXWA%+P8O9V%`zg82o#OniP7WjjnS~Hxi&Kr8&$&Yk=qUZ_
z7Cn(uE&^onLLFdtw!Di0`S)a&_WD)z5CkL}GdF6%uW=vyYk2EHukp(Fm7XYCPy7=j
z^<D$yav`ULfb$mSBgsI1Dq&G@ZP;((JOW8!+=U0f!2F?k2n~--K8h6TfpwDvFl{f;
zcv3ws>_U`R1|M!ehacY=m~9vw5c1f9k{nD-&vIGSM-{!0^toy7`KHc|CQ3IIxi4~4
zsk8WAp#=&#tDDm>-@N1Uikbwup`xcds<<nL^5IpRe4Owfy`^~j5#EW(zX}nC5kr*z
zz`_WT=IK1UzWKb45RK<BCP`-vAUukwU@O)jP?l@uj5nKppwO*!!@gc<OGd1r#tYJR
zd`kb}`J!H)!H`$AtWXj@YqMYy{B^61U<er$(=^IcBWeId8LsygsKz95VT+6QQBY8T
zsQnmEp>IL>0MQtT6Zas<VgdA1A)m#i+T~`4FXY#)xaxkjkwK?ci7AN&Dmn2jcu0%e
zpD&S1rdA+M;P`c>dbZuyrx-9)=sOSut=jG3^j|eFC4kWDOuMq~z<uv(vkQU>*5v-_
ziHhc64z}4LwjK>^Jd;y31M#@`yW#QBoe*!l@^&yTFvoRgG-ROS5AC-9v_PRyWqVW$
z9^j2^tHG4FAnNq5tY@`SXY!aZNgq&71r-{>S5``WWa1Rum9^SCVPQ$?TC<?$HtpaU
zNWP@wk#XLsc&vPS3ZXOMZaDE|)(x&msi6reN7+YyvhdQ0AKVa1O@hwFP)&Mb+}(wR
zu6D-MUabP!KNsskK+O0K>BujKrv%Q+G&Y0#pm@Dnrja+*HY}rpEB4FGb1!EyL#AlE
z;p4}_BwM}3cFy}%eVM81mbwgxcKfgx4y35kj2c8<TO+H#iMy?JQ2S&R`n#OwVdRj^
zjMejTmy@rHulWPxP_9f#G53&YRscz;sBV^f<*vF2RTB+}VZwWiBb9d(Up}q*I58uD
zQZx(eDv>}tg0&s)jlC$CJ=qckya0Y50%_KJbvth0c!A-BT&$hdU6lE7_2)!b@FTVJ
z3FbN@;hsYRr{$P@ZnoCCZB#XD%{je|SX$XUPL>O?0)uDW0UE^Xoi@x%RZ1z#UjkAz
zxB)-<jQ%mLw|ss3Ww$-PXB5coc%<aNnfRknzoruFZ;`m!`k)=yvpyj4&LEd6)WvK@
zb_H;eREVutRlnCOz8l10wd`{%Qlcz@sDd{#&n*?7U>a&rCINeQgp4{HeM^q1)Cp&f
z_JI1_TmfBHvf2MUq#gvJFgTm)fv5Z_<B9&+i<xFHAoo=q&%eEV6Gl&ZfyBYOFK&$;
zsN7BS4$UHw_5y{G-1+L4?aHQ(J%j8Ef&H#E0@|N|)D{sk5UTx?O*24FH$|_nTmM?D
zkUCve@czU32y2G)Utja8$3zm9>T5*WAV&kT4xvfY_~HmY!)xH011iF95^I1AggA~g
zvBEHO^=`>nIPq-rTY)Tvm=Dq>dtUE{+&OdjZ~0MGc#ynZnpE1a3h%ettR9&LfW?aZ
z6H63@uTC#@?nZR*ILfW>|EF}yJ5KqkI)ldl>H?y))V@Ty*{kzNOQ%m3Z~eeG#EnEi
zEGKKFsU2rWz-U0$TCs(pCX-8dafnJ}E7Q*E-{<~fXJGAcUont-QFxjG@ANLVB&_iw
z)opn+ZDdjluq9WfJowT{ZG5?0PJ|ov$g-W(nyBn!dl@cpX411(EQ6ul<JICO-|Mt5
zSrEdkm=9*R));!WQl}stEZU}2B-!lu;kDl3LJc0gr97BP$a7!bXEX$(gq#HhiVEaU
zWpS%?xm~(yio9;u83^TY&?fS^+jlyNt)xqhv%7_+2yo$mQb7{9I5!VY+F*7jKBaId
zzd+`Q9H_qxs|@*OtI@UMi0}2Yr`CBHlf+_1<;?qgE(U-qI_y<tE;CBuWVI(vm9veW
zsPkSKcbH>5Py57}jh*VFfBY)>dK6!*D9Uph!Cx=5x5H?LADzOL)w>gW@yxig!J*I0
zf+SSwL<~p8V;--FA?|^Rfnhra`*%MQZhAig$zr4epRiwk!>>lGZ5yM%@cZ)FUZ4G5
zd3X#_L@T0}263X<*ec2HBmSBgySW-;g6N7^Z%+v}&HL64h#%<H{Ug&iTUo$wMxmcf
zBh<mlvg=dOBkwS75b*SVJgK3gFyGLZ(G4xoZzt|wKAv0Dgl<OyU2XNU{stB^HCURH
z4~;G<q43r2l4{0$Hhd5}=yx9>M<sK#wjN2ioiO<ipp;=2(lz)gzAHAFMGpoI1~NXd
zwjNy&y6xUAxxW0dqSGnFy%cmzC-vaKO?|XcD?&P5nExHI@b}v9o8H5mFok9nCd?-=
ztL++Y`LYk0xB|;+GM)8wx(Cco7XcKO&Q&gRO)2DEp@^u_fQZRuQ_1E93sF3DF43a(
zPiDR0Mvw!Q+pnM8aAUphx77~9--OxmRb1Dz{>LK^j)XUTORj0BwxgUBs*5P(2e&TP
zEEV75W+7!ZoM=*O;I{`9ph|o$qvY-;dV-ng*qa!DN#IfQm*7!;KN?)WE-n~TIez3c
z$}L*&#?L%2&x9%KI!*k8woPPsaXxCNbh1yO$neBy$msj}DKwYY-r&au2VFX&K>L@-
z$kWQ#{jtkr#iMCG#gbKLg5-yH#yIAf)Yj<qd_EJtNuXECgd$AaB8T;|@}2+9ZzjDy
z@;Bp42fhBG(f=j)K^^ChMBH+Gx#`u5?Ji$|TujiFDU6a8x|}FAB^;;Le=`lnn?@g(
zvfLL-w{g$bTG6nW<HL#hy{zNR$wJEh&t4g4`ymQL6f5~(vgZDHrVMmzhNpL<=>Z>r
zw4Tim1l48V&j`WGIvFlR9k-_Nxqp6sKD<5J{FfgF^$y7WljBJb71Qyp)M_eq<f+Pw
z^+=>xCU6t*diX8q`z8*uDT~<5#%Dn>;&M<`0f+S>O%sA|HgN!dKnH4gr{a}Ai2HV(
z+T7e!#3?e(RNjWPXr5(G_))=wcj9{=s(5nUO-#1KK}pHUvc=6p3=}#hf$xE@KN&I!
zfNLkhrHbAzoAch0e48YA@FOJ|+EAGtp0rM+|Jj)w&#R<Md!o3af5@(k#sygS;|vMA
zzAGBMz?h@ABdg5gPGL!|%%CngxklLf6Og34bJ50}J6$khC~uKGf33M-?n+k$Qk@rU
z|2b7Zwa$i%R_kvi+uA)QOZrcV!g^@92$%dc7I%2@@LW5^k!WYw7fGARjjq{aAYsV9
zlCfsHi5zAo;E|+4Bjm+z1TBACk#E(gCxU$~7{C77#dXKq2g<Ps_3I76hc+Wc&1K(<
z^rr-4SqBemSYe6X@Z3+=P>4-twJ^sh>%C&}6RwGp_%f<B#|O-;wOR-<*W<kf3xU8T
ztfAHJ1s$F^=;ogp=n@qQZLy0cJ9&4cnMi#euiAfam(CT_gPuu5aYXxzpC@zJkx-k<
zD6)}9z=2<{Dx(ptj&!@gL;7a7dpEYSSyVq@h>yc&0YZr!DCeO0WHMABUBcV!axGG(
zd_r$lz{~|RIH~=II(r{k<>eU+Q`3VBF!hLgL!RY-TU@s>f@hD2&C`FZFpc60J{&X!
zc!3LW(H9FL@7pcGzhH(Yk%Ox-{9iw-=&0M8E7h(~$hN0%>weBX+aAmaIq*7d9*TP8
zMLGln`i^J5u0Bb>-z9MTW^AY`PVnSUjGb4QR7M=vO<~+NnjHM<hhpQZY^;4j+-(ZY
z$(ZcBYV-cA%-bK?4u!nF*_jr)hysT-_h{`;!9?eVVZT#R&(pRi`HJTgJnX2$!>qRM
z&+lGSBT(o)7SVnY+KV-!AjLyk=@RR5cOMSQq7#)^sK~^^x7_)B`{caMp-@MxjH#I1
zf@$tgV5*Lp)RPBDhr(i(kmO}pUQ${eY*ejK21!a(fxf6E+=3uo_}lzyvsF_4PPm#6
zD61N})Xkat>L`i$DpSvCNY`fV0NO(G^z;l(;9nn-qLt0zZMm_4HWQ=~3qu-QCsq_o
zosB0IE4Q30G+rP_6LU*t&`P9J-BIZyCQen2LpMu)Bjzaz?VTO=C^3c{r>>?RlZ3VJ
z%LlhjKn?mVo-{ZXZznuA%8v@HW}q}2lPhTe{GPJ0(JYr!{f)I1W@y3DiHl6!qq?kA
z7x$_Mv;17rL}dn<O0lOxD}_+WdrWrO_b5v9P;iol0T4o3^Gv(KTb+zgb9ty&ej}11
zu`qp9P&<)IJt)%b%EZxlN2vs-f555V{}%xZ0lfk6jSc2{<HtyE*or1${-MwRC!Rly
z23zR+!MK+P`q#^&klkD~(!fnL(dFk7OE?#D7F_-OO4(>Mr-9!Q<%g-AkH)c2_XNHH
z0XL4v=O;bq>w6~d(0pnK-Zt;Ud4!4f=J*M*mUsWS5C2AeRaL-kUB{J#w4kfWi2hm4
zX~2KawBe0Rt{TQDLD9fI+bKyS34>Ns*7112{=LyrP)RbU3soQ~s?FP|LY`9Frx-F<
z;QswJ=#Lb4#C_i_lJv}@Ofol8tkg8Z7tnOE#gX!plFt>vW4BySJlo~97(>A4vajTq
zA($R(?<#u^<9(esPlxL^rS|Z24);q_<d8p0#{8r1e{`~2M6Ft-<;~maB9OwMERE?d
zU)!+t{<MF%PKYWshJ&C3-CL|)9dpV1kehWKv8}|jv@xyupxHcQELGa&(2dzQ9Lh#f
zEL3HVoyx!LYFhm4>opMdmVwOqk^<UJ*u)Z!d*$xn*xy?*N4mkg;8~1wdbel{R>(Gt
zv{g0+B<~c25I!7ZQ7^+>BEb!Lned_RhH(B2(;wXJs<Xrss~2`c6ykIo*G(|EE1c*S
zC5&Zh<1t$G8I~W9=V1DnjIH-)&eMMxNq0LSC3^N3li3@l;7JP|(zD?z@BMlgs^oY^
zc&v%fJ@Y!tz2Kcn_0(J5E@|q+S$q@GbUk9tF=hakUH2w@)f*R+gj4#OCUGLbnfU&K
zLG-N_@-wIZ@Ne}%r1#Q1Ewg|YvV0QCqLCYEcJ-SeLqCs;$60ha;0PXynMxgYET>0f
zSQGLAG8ASk*klu=GMq?|`=5WG2RT>fSI*<<T-6ZfTa}C)#v|M4^f@*t&q%I5o4Pq1
z-je(-rUimTnO0c=B7Z_{+TcEJ{?;lPTCo%He_tL#uvMisJH0O73(5v{Yei3HaXW;h
z2x4yKpDvUkUU&#E{=WXRKMIB9D+cvK+|eiq%7hclBuX?TKv%thBJ7c8nn`iEtfID5
zC-;zkLdhp#aNl^2r>!u{!|sqv%%-!o`XBlmiLsyZ`}>89%D@iz>jwf=poPAGO*iBL
z-Jm$TPInV+$LV*61+*bq-@N|tFHqw~S5rX4AvLIZr!yS}$rmXok>u0+$&Hx?b3vL9
z2&oKX^kevFo4liu_!IfCGyP|D&ey{xPUbSr;+>y>PC3EiaD8I;M)bY&6Ys7b(m(Ru
zg>zGJQ7A8R9_+0r8^6}+mF#+%TwzGqmy{J1#V6EFCxsJlx}!y0Yyo47j;}_9amI$D
zcYOFw^<tTl2fj#GDg-w8C0xmYfeNum9{ArxHdZ!EfGJTe)b69<G7=?dbp+|<4D@E?
zeCv%>nGMhh_mrPcactu7c951@_@O^zgJKJ_d^bT9RvY%P-ipOCoJ9TyuCzR2X{oPI
zwe3ak`(jhJxf1m;v}JBD76*cgXD4Q=UvWu$2fe=Qb_1o}DT9n>*;5JN)_E;)?wfl2
z*eT&VyraSuI+dQjBry!J5GDg+)*(V#k?;>@@6eR>fu)T&Ae5`4^$5a81pD#ATMeJp
z+yngjR%eOKuv9k*TJpN$^fQSd;GoC%IrsB}d(jv?qQk|MT5}II?*?b7N-HHSHmC3~
zi8qBl#RBme>zpeoO^V4VMab_>caC^{nOb-axYpO^*Z8|kI(N~qvqIcf8SLGEB&aVm
zcpFr)v9P;eR^g%(vxLEL6CR!Ov&Y{~0c$*Y(MYY?P*C<bHB1ta31A<Mqp(H!nf*f{
zxk9}n8=xf>M`KoSEgkiBRk8qoToV8BP&80I>o4uEKJoKZ%L-1y=!0_QR5}i*ZI;{c
z<86dq%@#`6>m~P7@I9UGhL%2J=l8Q_9mJx`f?iMxE9T2fdmLMA8tqa{c@?@#ArxxK
z2z{Q$HZ@7;UYh~*=5Zcnta)9-{cEhvwW4nwCP$A5*Sh;VR*Q~r)N<?Xcb>Kwl>5rX
zY?@&Um3RK+*uFf2mjk)cmbO=@$uCRSsmY7<OTLc|YT@b_$rX;eY9zmp(AYcLjLQ{+
z-x%yW)M$WfU5}gC)H}&(myyLZfDoC#ThWCNNq6y673VDan^T08seXZJLrV65D2Yjk
z#AmXEjpcss5Qo33|7HEAa%q0JZ4NQ;p+j-`i`BFDulMt($?7Al&LU*4`N3pgh5g|?
zhtC^2W8PxzScPr_u-W>b;N5se=v_UT&)%Tg#unP86_}k73fQ{2^mFSSg&mK?{GauD
z<v@bxtA)G;Y?Zj>m~wtTV2X-Soo+~+c$}w-$?p0kNV7(+AvdC(nX5Avs17jeQBM7g
zWniXx*nPm}xm8q`OXjlE23NQT0S#~cFdQIJ8A}T^l!knt6VUPo(xsvKk@LM^P>C5D
z&?4_6vFPXT4ku_Gwz>!DNFqOeoiAG{Kw2WYfMMS#J`ogR(sCdOiFtxt0{Fm&U03kr
z`v9ClDg3y)I%Z^ij^BRom)xR4hX;FmqJKeTMOOf!sK7&wTpANh&g=NOWJ$g9Hua5D
z2(rJ;F-$!Ri@L5WT{_)?bnVh`@oRzO@`>@@{28JG+<9R4>EN014)eE8Tw8?28^;oO
zpA-x(crbSM*M8r(j9;wYYhJ$9Z9=CybwvO&F6k2ukU7-};2A~@?En-aC;Y2+C23<0
z;B6q^nb+c!G*4{|{vtc6C?)KkX@6bQg4u@aB@n^VO2tbYDVjjtR{pASx7=iY4GBCo
zWkNJrT+dB#=r=v(K~7PT?KAn^mMK|G_psC@j0XQZ$6SIZ!yL4kFih3Z*wcV?R*KJO
z2zeo9;f9NNw^Ov$qXjN>z6y!o!X2P|7x#m3Gl=f^C*bvUmo3;|_?XcDuTcINja;M$
z8S0ldHbAVm`ggipNKXANEYz?+s=&EN@lep`it_0<)sSNHTiiTByVDa12@Fd&m&bdC
zm~C^5xOb1D9lo#1kyxwVRCMOnBb_<MKr7`!y{Qk&_oxp0>*P4_r;!PA>dnVKX2KO&
zM&NP@B<J;FQX>`<{dX`B&LO@;muhnUN{ck0A}%G_QmI}{27T9f|L@0zha0$>g_-*k
zLk6XKbHJ{B=I4Bjw%anz8OEPDg@ut%t8dO_hWZMFQnZ=Vmxgn^#(S=3&t)U{@R9FI
z!5qwDGX~wodhaNw(R$VdE<Ap#!e{diES{;AWQlsZdvnQ3Vo*?Vs?&J2vYRWhoiBM4
z)Me`|8<JgxR)1V6j;5wz{NLF)*6EGZ9IRslTm-(7kN&`9`|=XdvQo?5)Q4~EWFyYf
z)*ytplya_YDQ{)VjC_+(Ik&)&n7EakVscaM+XJc;%yMf8HQp2b0wl*z$Z!$&S47wm
z3Bbn(eFr^8ET>8=INF^ho64AL1^pDhbUpA#3?Y14^}PXUc-o^{hvCO5H5AJvv3<aA
z2yU8%+lU9qtCeQWLsv_+@GWQTk<MyXq{}2V=Nd(oB~kM?cyi^gyWL1XF?rkZFAlIH
zuOkrE=N2h#Dzq~SlNnV`tmZq&-<9vTCE@tJIO4wx3Jt;j!m+a<0HV?CBn00?(0EPH
zhIGYcQ}KLoO3m|ya8hed-bR=EOnu<;TI||s&<gI_36Y(7dU+v8yo2Q4(0%QSDbUqa
zaiLV6*5zn&P?Tia#%8gcH2E=LC(xE6C`qS(zW?<SgmO##;OR*Lox&x8+?&iZF}}lN
z-0_`8<At<7Ru~Qhg~TbV_WF1_u^GPH?5-mt@|>gqz(YbpW<i}at|Lqi1dlt#8iUuP
ztJDg)49w_<pnvqJP$bkT>g4L}3Z`b>CSkz|<@jQ15X@gIoZRAxv%@Stlhs8c9M*9p
z9c<!M)y*Pmt2oAf>h?ps=Y-2k!WlWCL}#*#jlT)>j(?3b4+ZW8yV87mCnMX4SIiBE
z2)a-Z%j%ZYwE#1ogOIVn;#BD4t}4Xk)`zlDD{@2M96E*+=Q35GnSfjDh34wueUQZW
zS)Q{{TP*Q+7<ccBKcCRR!iT1R#+Lxy=QBq4e*0=*y~Xm-16HXU76`#VuJi;Ucgd+W
z5$(_7)9UF7qaH<0J;E^_<ea!8C*b@0(ikqlm71#L{EU9#=J|^ofzweNGAmT4#khO3
zqbK8?;yS4BQ8Qb_-TcPaf+LZ$(kb(ijIw{{wB5?D%vjaBA#i3bni^&DzB!qf^!&#T
zc|9-Nt!Q&4_0M4P@`rlFbIDLwN&b^|wW43A`!0=Mk5g2EJDW0@{!K6h1PR>6$33jO
z<;0-YCG9oi!}k}oK8De<gE3LYY$Y1R6(M&+zRg<`#GfOrkAE{_(Al5N#1&;%Fq^%0
zf9~>}^YcCZGHE9ghL=&TRduqSYbQ4Qg({zP=V-T8&a_maFY`cjOQ^L;YZL?p$BRY#
z%b?H164EO0lm^0M+>wWePA5BuPv+p$Bni_`5G=_gvR-p@2*4}?g5ufL7D?{~*`Atf
z4u0Q9vc9r6qXF9<Hof9<d%h&~FMjvRRW!pm{d_dr0kQFXY$!d{@u<H`b$k?fAZc0a
zJWzarcT>Az1pW8;;W?Aoru#4vjd$+YqieydZ^Sp!rP6oow+ELRbR3U`fH|iYPMdTI
zpG4;n_k2KZk(3}CAyRd`9c9C(W9%Vpj)^9bFI^@#_vU(AWbNh?&MaR9y5&{)JLpT-
z{$9q-IIAv`NJC1b!zZQrFnWyqeBhP(aD{Z~Zaw7F>HS68bA%Dru~sNlNVHMd6s|Ef
z|D=VG%$S*%xm5d8i#?Nja=<D$t{1kX67UOj4r|?xm%Vjz-PP-Gru>N$g+(t34h|0W
zHwUycjDE8(B0Z7LO0L)GDx(8Ce$`|#gAnxz)vYg|%iqC;3f_OE+wGA6+DalPJQyq_
zL9rl@bTkH2e`>SeME7-faarI(4)q7|Q8YSSQ<B|W`%?MZE!H73F$CN+R`bOnNr-bX
zpc^urAb_rW3yTd+&^d_7+@%Z%qMyWPONzfhrivn2jyevx)+&f!&nj>P^F~7Q5$jBs
zd2~_y8qXNZBZYKoRNSL>SlHSHP%eN54zJ8bWF(a={}SM*(yX?*FAKbeesWEwWs;)k
zty$HZGiEPw3R3Poe~+kN3sm?vfz~^|p_Sk;R-swZTcn@jh3d9=E8*4$B415sgfXpz
zMif|0x%d*;%YRlg#ACPx$Dm<8mcls-q8G<iIPZ?}jXDaC;iVaPS#9+Tf`N&E!O}hz
zgNmM%vkD6kxoy_hCuYS)fasb-qq<r&^eG@vkBQ|<#)4`m<>>ig7B9$~YIBOgvoxfS
zSs`g~CxzZyowoAjG4C-8e*D$LiF5l0t$t?1yGZ|8^5EUbjQ&GGz&mQWv|o&<MlN+|
zK|i;v_0HQAD7kO}?F_^3Tf*|7LLo!JLXUC&xhPW@nM|J?oR0}FnuhUm`STk!jF=}6
zNOP4T6#>uWmnm1~=<X(R(r3k<WW>Dj*oD|SjW!u*9(snpyi8oF#&`nr!y8KjxZfj}
z?Sl2}pA;{4@3}KmeLeBw8x_Hzm>kZmPYNw7Ft3G6xKe5wL8=d`JA>{*E+&6)?+%yM
z&ScVG!p>x0-ox(t=`B+0Q&Ymm!>H|+o<3AleZl^GG<?SIQ6hR~n7Ut`$s-u%iG}jJ
zWHMOLEIiwT<#c<*^!k3;)BG!#2f;{z+oUBgJjssLCrDTOz}NfpMMrn{m52#)kQ)PH
z*n>4{bvug7_g1jz)fGL}r$?j8v=~4l=yP!WX@ER}@1SHzU!M>bt+Kd~9JT#l_^ZeB
zmD88oeLw@%3SM*y+GGxT>PS#riXETW?N20V2I{2FbP`#Akm$2o+ll*hpOhWL>ry&j
zt-{@?UmD`AOgB=7^mDER)`9l7FYb{7j^7Xu&CPvgMV^vE>`Bsd%zvPfRi!sdR@vM&
z&z9G!5wyA4sOp$DElW$g5j^m@@2h!h3^`3vg)~sqr~Bh^^tw5`02aMI5ehqc^s1+{
z%13;2iz8pT)RF!nj$Qvs@2MWevMBvUtrf!v7RO`16`g^^QK8b*0Dk8{IAc%ZWM>Y;
z^>m!mDN&f1d&O%Ps6TjWm(4T6y&Zkqe(ayO?65}4g`iy4{CQHu+WRD}fgrZc_>0bK
z5c@V(9NmvAzxx1cQmUkZ{t!q+8r>nCH14J=CzY-77w6?U%5Pf4=AYLriklyQZ);cm
zx`F$TKh-bbCB@f6FzPxUfzGbVw)$&#x2NZ*DBQZ5zNh{MtlZx%(n-AGAI7TUDB9GK
z9r&h$3P`qYf5-}C6VeRJz6X~v&6W_jOAFc8=7ruJ<crZ;C};C5e^G&#S`J-_sW#?Q
zm&h6s#Wr^M(dICN6o4Flr9G%95aj(DDhoyp3l`i2nvZcp5EdfVo_8z<?Y+A0{ek}k
zX!cRT{zSWZZLo^x{Y8TrwVRsFtww8r9u#*UIZ)xPPa|f$+wYD_|F=Ll9R2I0M*~UX
zW%{E1u7QjgV7Cv3P8M_;vRy^enNH!hG`wn}iT@E>nbr_)JHWnJJgv`I?o$*BMK~b3
ztu6v-d<=fw85D3CMs6g#uu<BHjhDrT-}aMh`HB4L#4j%8JbxqF0T@H{HNHOr!2Vlv
z8-2@_Dp=b3QQdWDCt8*tau{4s;p~j)+o3y15C;8ztbkP4pK4<!hl$KgQ1s|*J|P<3
zT2oSI&^O<pa(n`=)Acdj{#rYhG*`pt=eqM3em5Ft+7nr6GQD6F3HdNfw#uLie|I0L
zI62g>fNXfxXkKRb=AY6CBkg6dzHb3~24~M(TFSeF;8&GE=mCwdS|^F)flmAH%WeAq
z-swsUt<9p5m{lv_>?ZgiTICO|LH8-hFb5FwzBM_Wvd)nm)$L;%nv$s@j81X9J;>)X
zP8yDfoAsNJ{^*EuW>`w_SKtE9_eo@~-2=f;vPOl3%~TSSbA@;o@QzM=KXt7Ip%~~g
zPR>H}H5J=|&(jH<$cb?g!)to^9~<nmpI8f?AGa~F2JNCCG1tuJlQ@}73%&D_ga-5D
z1`Vv8P|S4Vtw{fnEhLUT9yA0jt3|~ZJbAGsMKb~FCZqt=h~Cz698P8=f!6rwmJixt
zprPS?{D8Ha-JU8mZ@bYe>4Tf8*HDCD`xr|ybJr@WIO;qjuzEjds}u^&10NzT3!^ed
zc^;kD2Ngw96HS3&dXt;qVg|ybGDAdh$m0kWhBen?3!F_>6Kb0htqUBsGooHDDc@tI
z4vA&Z4Gk_)!Y^%WXw(~|p143ogR^>{5KZ!xsPse5?9B5|Hb*t+CA{qr&o!5_2IJhj
z44yGUVl`j4W|JIKVzT;Ij3$_AnQnn^c0Ke8%VAPiMl)9g4mSH`YXwAat3Dld^;au#
zHi*1NU|hX!{9rpF=nXX!mz~cA%bR{>q$X7Gl8&PPI?u0@OAPVG^}5kN1Whqtb^DB*
z$p5)r7v=G}mT`0-s(@Hqt9yWh((^}n+@kjlTmqN#yeqHQUG?u35cW=Fc<!5$oFKdz
z7Gsr`SDUDcH@$ld0d2_Ns@icY-ZTaWuz=V`2<sq}j==jw!>{Z3cg?<JHM{Boj_5+@
zr0>vx@kk}-dpf(R_&XcKlF#<p8YXcyc%JucRJVttLfc*iI11oA#Afg2Dq8EaUUfd=
z&tCm=_~wc`xDnv<*xRI~w=V%-y?g`b^<2ZvukkpfqE=%f4$%i>&0R;ul@j~X#w`6)
zuz*7=^J(4u{IbU4d4vvXCy*jaZxg>>Yx|kk0HUd{cR9?b*?9f7r2d?{;!9gfff6pU
zWPu=@Ol2aK@ib28^P8U6?H)8}=~>@!4N`|>w@e{VdqdCgtqqhD26bA|*1wkh%n$|D
z)0IA7Z>!AY3RHl)<)EZmrD6%uoOZT{ybMbIrhx&b#d+184x&hZqN?Zmi`%K|mKE?_
zH4sxi^gwjt@<eX>U1k7?guUGi+>;`I@`{yw@=6W4Z55rzZP`pr{=gRJJl30k1&*0j
zm!Nfhlplv>ty`Bt>PHE38F<C~Vy%-Fs%0#PQH2)|_%D6~1^1(qBhVtLDT$izgKxPj
zzJ4~mDRrs(z+pJ@#hVmFUZ>WZMsI7LE4j~M1t~+R$BI>TB$3>o^w6rZGT-9WUlMOx
zg!ejoKrC*L5Wo$&zb_cOF+IPj&;TrofED5-yNf8rBxx_a=V}%2(hRBDYH$coUH);?
zu0AW00}ZFln<t=U(Zz;e?D*`(BsxoZC6SMr3-72$Mx?bbTbKy%lX&wB^IA!?SbW`~
z&R-iP^n+-1*A<Yeb>>)efC;Kc_}d%8e`<~=C_}Op-VNqcnV2_sh<g6Ee*Z()TZU!T
zcHP2oiIjAMgmjlQNOwwugp_nQC@J0DC7n_t-6`GOT@n(~@SWVx{k-qD_i=3hdR&XO
z&X_UgoMW_-S)5j$e&3Hte_UT3cnFU|p*${X@rBv`wnbPsyVJye^s6iQK-s^>@q8by
z5wjEVf!Eo}W)yvw;{_jS(~zRwi`u(aVfQM{G4Jex+A257FT2L8iruZ#1u6Nmo@6%v
z!iDZGCVTpNbl<#I`;|Fm;yxee8+#t=<#Ifek*McpGkx2Ux%uLxgeOiZ){S-E{K&J@
zt;M!A?^1wNm9nrm7Jqf!rPO8|!MkrExH3tn^)1YJb=atT`DdC5f919-fVsw|TpW#s
zN0!YvOWpn325n^RU596gZp?MeBlm)8D4*xaszAe5BpV$7e*@*)s)aeQB+c#_bY#TT
zwU>&#F$_#i92MRC#W1K%$YUXRsD#|nfM!)Z+7w9M?zQ{AcxFTcT&sC@<=QQuoeg?y
zZKURq+HYTsIPVUnkYoK-TmhpAn%o-!lZ&^P9BJ`(PGR+pR9w_*gJ;-xOk&V2A;uU#
z%Hs>QMR*^-D9!Xx6a<xpeQ<tc{i)?8{|(*4U{%fGpnf4~WXW6pduaXmw~}L#sujB`
zRL|?P3L2M*r=LDgR!TI;(phKojDFIXjmf2>X4mCkYOX=k&g^(k9*!r{bkAPzwx?<r
z0Rqk67NwU25|9C-P=9j~`Ct@`Yg$AbHoa}474H#21V0YTevcfc$jXZ6W@d8^1x#OU
zKG4lfzqx`*to|)w=;^G>HJ&lQ924J)ct{YPVW2yV$n)=Vyx6V=q)}2%2<ztR?=l~+
z`!N|~3Qvd@$2eh_=btldXA+k1uXuEQ<c0gW9;KXC25{Q;=D)`t!Ne^CoyPxkBNyhp
zJQqi&oj1jkSq`iEApB>yh?U)9K}u`2b6Fq*8E)48=Tg8#?ec+kPVH-Ly9(>g4isUI
z)x9oPy~#tT?6lZ*v(ELBZ~;OQ(ZDDy$zAKCbn+2;JMmqPeql^hSfz#!0*1i|Jk}Cj
zRunvn92^xuva4{#cyfGbvxEI<%E&&{=##Un3q^tZ-+4N=FVh+hAIBxRQ5p2AFot=Y
z6K><>jVf%kdu9W;))VW7#J_8-<P11ggf=EPQsh5kvn$LR{Hc8!p|4UhMKBUQ6r`1J
z)as<J9vBil50^<~lJ&_^{2G7TqLITK+53X|g6J_>Nc{ZsSS#i(!|{q~;AL8?Qwz-r
z#^^l)X&Anw-&dLCD49PYuw->+!9YsI>1<0Hh&5L$_U0<zU;ilvLK!0cwj-MQE6wf~
z2h;RtmzVkezef>aNZyo8$OU)WJYhW$)7(s0)}c!>ZI*VNXyTqr_3bGgjPprwbCX;L
z7H|!SN4zq-=WM>cjsQKcgRqoY_y;V$Lb7|h0^}Ubzp2!ITqEiZHDjWR-R$LsRcQ*T
zW2xFY)B9sb=9GvC5{jsl`X{tY3T#|uDHlu|%uCA1Afsw@?jm#{VWoPkln78nz7CU^
zv77C91<OQ8qaYXY^$|zAH$tB~Bf<J?%<Ke;=n<jeF|+(eT&pEO!|5!?0VeWR;O$Nm
z#zkaQdI=UCbduYjy|b9*txTod8U**puRDyo(ssX6_#|t`zDkHB3?$~5pC^u;B1{$u
zA6e8G9#Pen-$l_SxuOt$Anyu|{zQ4wgF?hRF%j^}Y$t?zs<g(41|GAj4V_yac`HsL
zYwoiI3ALw8OPNzFWApGhXcFBC<raI~a3sADD5rn8wKVOy-3&h03SeybgJ8J&Hp~5^
z{RY|Js9juG8=H@CssmEGGWpmzkHp28(YbAC^F|*7(hEj&L~fdk05zE^yNY4YF}^k)
zc7JkmIVyt)mmASnWNw}nOX(|$DVB=Tzy1+bE)#YwYcZ5aCaw(Y64%_kH>b4ml)HMT
z(rmi)4%)X5`{&f^0%m0|tk*HI4;6Hq6Crax`kTL5)oCx*Pycq`x{FX?YyAxm&8b>U
ze%$_oUVl_O(HDi3U9d7+8>*yDGT^2-Q<g0-KtV#>-o#Zn7KFLrustviIx71Y7qzWm
z+cDbQ{%kC8PWhZ+Tt8+!jGnr7r~|6!*ZbOOivw(QGZiV+l<O~a<vK!4odSb7wzEG5
z(27>Okv;STRn%Mzk;`Pzpcuc$&$7Ek!Fe=2Zp-#_{iux8d(k7^P4$xTi|bdW-X?9*
zj;7<UgS^(I=!X=i2{F`bw(lwLq=eM=7A4EaBP}N=q_2Y}dBbmt81jCH6GcRcRd(yy
zAG0Qm#ENnE-F1)bfzCwn?_qh_8^3jRDI;Lk*0??>FK+u9<yVqv(5gZzA&F(Dhc$ab
zD{xW4L-?DLnp?s3i1f#@SWHrJCd0TSs!VD%ED5w9#ZHx;@ROhjBYGlU&$>a=JJ{*j
zE*{APs3Uvpk5`U*0EarQdZl9C9*_W!f9J(uq4g#qQLkVtRs1LCx}-iBLuUW(wI^HZ
z)mn93X!|sAJ4m|qw854ygQqQ&oB<+ZTS2?|i0L1uw*~a1HEZ|ujcbnKry0tjgzg{N
z9^F1nkd#uMBkrEPobPW=&zWpCT*80*=^L_ktrFlQ*%zv7j!)y_BJ<-sZP9+~Xn0g0
z!Vq{gc()0Y-7J`NJ@tmu{tW+DT+$ED=BpXmzx}6L+VX<s!@=L6(y{Awci&rbS(CfE
zZY#W_T)#Z!dOT8=)_|CLdf;z*@whW*$RKPUX39)67Wc--w`-^qb_&;hhaoNz=~br<
zoC$iNUG@sh1ITh%OrhCvg8~WmOCMj`Nd0y1{Rzs2pO(1TG>Sexg@UffIt$J2KX5Gi
z2vIp!KO?O1MKh!rM>7`GZ>sh@d4-yn1gDxxKv-COu?66x-vw!!qY`LSr@2)GIO;Yt
z{cgEQmob)>z$I?Z&X^U2QSSVuwSznl!*-cPhH;%_sGK9#cmfG!QBQ$K5F=qgb?!W(
z^L4_1J2B~cFp}09ZBVt~G4D)J|3ExFvMCdmwu>XaHig4p>0n~7_GC_R>VHC7=d^E)
z14m52jf0SkpU^3@7PZ<qQ6%iy4@wZ1M4W+4dgsX6Qq0B|@hMKXos<V-2c}v-%y52M
z&m@z2Fu~J%x8zu;Qu(6(X2H8|1E{I26%dYKBrHTeN6Ep)bs`dzHva9-7%!0B9=p^v
zx0o31y4$R^$0|D@YuxJbL;}X#(l>f<r8OVmn(rw##M!6}!wk3M_m#iwlpBF5wTj`Z
zx&3o|`<i7Pcgyo>_lVT<&^xHa_%ALeIlN)E6FqR~cRCOc<tn+5z2sHPMnM1fgiLc*
z=W?3I{rk6;H1A0Ay{o8zvb{h)T3puXtv%)9+rC3WTO{>l$AHA|I`Rp0YHu}`O73gS
zCnq<fan+|o4g!4LSO2$Zjk*MNHoF&u>91tcR{lD3tmzQ*f8J6wa@-lAP)JV>M1C!C
zI9pK&ey`l<v~PCz_sCEx(Tc^8$?j|eFkU$Q*$_Xwy24HC1DqHRCRN-XtE8?&N#H2`
zHhE7OKrt3(;0&wY&wxbr+#+TfweqJ)6zM>T)335xZ{jU##))3~fQh_Bn8M3RHZ{}y
z=q*O0^KjIVb^M%IPl|7;l?(rvglt2~{kdwRDY}J-rE8xG)=00g?O9sr&1ojfePJ~`
z6+D=sx}Qmz$_o_s#XP%yYyK!W_Q*JT`D#cVFC^o}W0(YQWW-Ywlt1mgJlie&f$IsM
zxJ`%tQqW7JweDCkV1m-$n7{1MQkChwINfx1#urp&Xy!@g`83zhDqzUU{Bg=&Bt7DA
zCht6{Vr*e}vG_|6XGw5`6x7>+!)@bnT6l8seT6a!p*Bm+;?wjR6m1$E>_yGm{=%t4
z##YXj@QLl~+0QMV=ggVBYL;6H*_ctm@M|ViM4lpW#P?!=tT42B2>-B-x1yIK*W<I;
z2>oZh776)<8t2zP&Bw!G6J^C25ernk6k)kd9Gwx)`QIwarBU3pj#1HI;i;(1B7PJs
zS7bApFFSWv*hCS`1{x-4go<?X0f6kMZC61f^}EXXqRSdLzuSFyWX2_6KK3TpjPMIo
z(3O%L(a8yU(Zq9kP1tb0TfUlYa%2@0M6X(e_N@B-`C@PVSC?~1D>D9D|4<x8x-CS}
zJizI#9Ds=Je}e7!!WSA!;co(Mu6tjHG>hhKg-*kSjvB`*EJC_~XR1s*3lfJZ^5Qcm
zxX+t(&B0Id96!tCAW^Fx>v2saMNTt$9%n2e*Qx_CUrR<Br0#56z6oHnO|DsdsNtxr
zg`Rn2ZPN(<{uHCk${r<yPVW24qvQ)kMTBZ~6ihrvTh)6KwQ_|wz5t^RA(uaeFUJ@^
zztSu2U!9#RVTUH#QxYm23bL{jE7MN7P!+V5p&R0^y%w^6iKOj6Z^x)L;TpI&|K&X3
z!;+M2fJ%jz=R%ERMSf4NM?+9lrs8*#qy6TnUIa4O->Eu|R#=*mJA2!uvs(ux^YO6*
z1=1O@nQdchJ~HM9Y8sGwxWk#!!n7hqq`A^Lmh0GTn1u9AS-{(wf2F5d1Q*aPOQM-L
z_@N+wcjeoad02H>#e!N2^QCr%hBU_C28o$wLw3b_LVovD&X?3V3xoG&R}r|d87=ds
zTjNReGq@`rIq+}w99IKSNCb^>jRhzGH-|B=`#NHayD2dzdE-#9X(Yo`p(q=j3VU1S
z7<01Z58CuD{ul0_W|#LunBIm2+qZ6Fe(f9iSO9bv=9hGU7)=k=oax_`g4iPrA7{@(
zg$)0yJ^-?spRRRu4iFcHIquG+?{H4b><p#!NDnjXwVH*!goVYxdZ3$^gt%!yx|NO=
zep=hS0ogVdvjJISCz>l=z<Wp(zD66R=+X7BPs#6mcP@@Kymp>Ed6Tsmnq+JRJZCwf
z@Q9;dzXqcptQhU)2%*@sIS~WF`$|)m;y=ftK06Z7l7BT8dX{tZLw$0?2_+6g(BB|w
zy$i+R9L(ROLBq5QB@${iGFE(UU`f)%P0Z1${luPUVUpp_+xKx{Ij+fHdX^Tehdix9
zfV`sl2CJ*u8Z-oM6SAfDA{~g9(>h^~E9u0;25J0XZ6h>$)FbKC1L~!!3Ob9<C$jU^
zp9-=)yLdC`wX5;0=W4@D7_QSS)V@oB-Zs<xU*-o(wE-P}r^GPcerz)4TW;p*x7mP-
z9}S(h?1?!9VnSni_+M!Cdbi>SIdJ&hZ!)VJdod%wjX_^C_625V)1T!xpZd#B3ug-S
zFTPDE)Kefvwbd-4Zp%T99?Xb-zm@ooblLqT!vmtHJWZEi-q*SgbJ!hB0r*NL5_jNP
z9Svypyu|q>Q#CI7y1M)nGMc2c>umDzRj3qKi3C*1zQ_#3Qot^I-3Pv#p!miusPOYw
zTWJwJ$!kE=QqVcGwVx!ela=SSS4du-WlCpb4v1j0E8_#{_g?XJHTAo}@=Q`tJ#jtd
zvU!nuC=bL1$`Q+g5x*^gN17tq@;z_;z1dCqr2tD;#V3Vf?CqnnA;d4+WF8a{it)~$
zi^cL?1gg2|y&<MSg1suWa{K0h7C@Qx(jpfov20bG!yS{=-C=RI{d7^K*6PFU-AP-q
zHrJ(CK!l1Y6BE<c;h`)jB^SQJ1<v;Yom@f@kh{X$z5B*#y`Tuj1Yfd8d40*?bH%y7
zOAZ60$04`e!sB(edJj1kI;02;+O^@JMTBZCCSL!>a+bKpZkaGOch^)3VP)J`5;`$W
zdxM*Od(E?VcyvVZ+p5{jIGo`gwj%{@q*GL7rpd3!rcULju_N4?@QfXtq3GbZRU_M+
z+E-DFxI$oh(1_l9@qtVXE+xn`JZz-p+lWU+=_f&iB1F%-e$EAZVty}->C%rW2kwm+
zEZVsQ`)yDwJ4rpBWo1tz#2SW&$hZpILUAatf#_X{ksnpreUBpbx>#H`V7L)7JuYaq
zrz#Y!uzbq*=Ca9=iJowWv))RuwvG(IU!7bVyeO7!>)85G5{efXCi^b+twpfUnsDbz
zSKr?DAs1{Ax99JyGQ2yIh-B`J+Xo)?_+rI^y!H()O=0;wL29{N&hLxxaJ7GN;IKvx
zIc>7Rc;^_Pgx~LtczyAA2w!F>a+U&a{ORF}q$e2N&#vde=O%t*FY6t{Hu-XFW%2E=
zNosfPTlcfS7lZxh3<!gVvIfPOSbf+q4HWtf?q^k3y$udZ#mf2ck%MG$Q|{ACg~>0>
zSI`o76o?bg^502jYFzvj_JiFx|2@wJCdskD21k93B+RBj!iDQQe&ZpTZ%@h;kk2|W
zDHrg)f08+@9^4DAq(VJ)f)@uV18nLeijCcH?$_=1*@N4I@%bL^C7skMiRZ4oWQ-a?
z4?O9ixb!kd1k>N*`>kE2N|=pIqx#l(AU*o<Fd=`lxZk%gXh`EDyz{;ceAH^pjLhOL
zMiSkD1hDCs0N0VP#!KJ(j^HhUfCo1yHAPh-z;_1rgVEKWv8R9k79=q{M|9@NaepA;
zb7Am><NP2LLBK)QeqEH;46{=8o^$CbHtPkY*51N*KKZn_Ax!*kgEA(=RFN1Sxo=V5
ztx}4(NnJ8bc@GlN@jWz4xr7k;NlXpVt+_@+U2p!x%^KUGHx=_Du7hF!QDx$I(-UM1
zNiD>=<5mU3B=rRMkqR))xouwpIk0NphmFUTmW)MZp4%iTb&0gOJOr}qUkRs9zKPRa
z2u3O}$LPyn8_(}z=P$CL;vG!9WV&7OVB*P<!1;rq(l6Cnt8f#(g=T#>?Up$%qmamI
zH!524v56R)xx8vhx#UhX5GCK^=4=8m>LDbbjAx{BmB>1mXyovDL5pDVR5z-0!XVTb
zez=}><Mr66;bcP1bCMS>#T}s_JmEiaj99|6p|;#9I_FP{QVTgP_Y5+->LB0RbQDK$
z!8w;|KmYL4LhVmgeXQ2lI_0;K(ysSP^?bp@ySGT_a9<rHDR+~EYOUsrEN5+|9-~5P
zJRdcy?APyem2F^(@mo)FPlSmA9VIVwbCFPuB2g+*jmxEKM*DW0`a0r@$VD5sHipgf
zX@EfV6c|;y-A>Xo11cki=W?_w<1wh;qECIHVHLLb`LuPL-Kr{^xu{-)9@5SyfHlXn
z8c9s==dlXUCr3tHg6@*g-I~}!j;iuBPD`fCz3G&{kEuZzJu1@<y#<Jk$$)V!px-8$
znL}-GTgEFq1a;tmK(DXJ3EpTAwPZtKmZXhZBak!jNhCz(ZKz0>M=^~fA}($Jp7WPo
zh{X96QWpkPv+)%pG`n-J)z>Jrgx}QA5g$??u1nK%Q($(6^rT3uR8)?uQZp1yUwi$!
zvkgzVpCdcj{Sn2V$0@!{<`HZM<tN~^83+C5!Oh?GS1;_-1xlX2L%#bf+{ItCR?ZC3
zt|u#^!+e){qJ8l{lo1zWHw)6=x%#S?E#@7)!Mc8V{y8*CCnXBfMg1raCd9@PIk19X
zb1@|_scTB6Jnc`KrnXsOFF#Z%**)x7qu0Tttt!4A$#nUB!E<oBD6~*zR7fVdoS{o>
z@uY@OgrP`@woJidabYGC%|J<cF_9h|X0zGn_8y4V)7NS344O@q<xk~gURK{;+=Tg@
zG~iw<2G3pD_yl|35r?J`E`M6JX?F{aqv6rEi&F8}^M1I*Z2^K-8$Dr#+$VeOr=U-$
z#^c5zBO4MZulM}cfykJZe7pgdWR=HoNa1PPdv`RaH@m}Wl3<_|{ki(av@z~1UrznD
z6pu+b^uJsHxjr19Z!c)&v&x%)oC6UV<*>hP$|K{{`{!KKsH9ZX{tswOf_J~jbdd&L
zV3XN4iftIdlR~IFqM0>za)}irU|MeFZg@tuFdA`>b0e7~aOx8EROz1*3{YP)Hc3Zl
z)fVC~s+xWLGFnkGR#)eri~lrNWi*Lid+K?2!-u?ty{DQx8vl+se=}pzT0^wEt6OIm
zfJIoehz~xvSz;t%)i0Lg5sc!zmUP;PXxlcdNVFWzpxsHh%L4Ee8@=7nwzBCp$`y}r
zh|KOXA7p;GntAmgVt>;JPiS+#5(v9>w|R2O-m5FBEE_Wg!ZLumzQ@S8P2)J-Y4oWz
zSoBX8uki)%K)uPNONnY*gW^tCsywK*JadQ@VFY_&KoPdjSTD|}cpzF#yN3H*geoZ0
zq>lQ1Y;~7iL`|<Wn=9lTDufUdm^R?3T`TQ@45-H7iZt{!HRpDJX27>O$>vYO?`0Cm
zQaH-eTNxP6bzjZqv>@r#cy&oMr6WQ6udcLwd0B>CmT1sY7Hf>rPWwq{&uYH9CkGLi
z`C}f)4760$Kf*_b0knahQy1{D?g2E4>?6oE;RA@{uKMx2>g#^I{j>-g3Cwp#GGK^;
z_|1M_hfL)ecv>{y<Fq_scyf`EE)IoTXEB+-TNv=V{5Drdy=^4^SgVJ>!IF4+py(}t
z?jwLR9PiKg^t{ZO4=zqCl|!TN?zyKe7E-AcG7t8aBQPHBLcGhSsO{sEW%>wUDt8E`
zip5#=ZYQ@=_q$oC`wp@f(tgE5*&X8OewLjTpvvdJLlqCs9|gnB_!ad6&QJLqmxn7M
zhP_jc9%v~Lr9~HjhKS!*xXYULcIMZZMsaPM5?TS`faiPjP3FyKStu4~Ms4=O=rtiN
z?gA2HxisF-uf&L2bSX~P##Ij3t>`Fa6B0RX7RoL&jFNhjY40!DZI(1E^{PLG2)}at
z^7H$d8l}E1!aMgfvG(6<s$IZ*{Wfu;BI%13eqnx8Esmu9Tj54kvcX}z$bdSg{iozf
z(cRA7RHZSWS_+A;qsc>}DvXPQHPG&WDl2P&LseN{_+HfcPA}DiJ2|uU$B9DexeWd7
z#XbJ&3zaAsm79iEEI~cSg;51f@`E|1$UGqI7O9}ys^Bi8mhU}di{o}WE_r((bwmoy
zt*6sq@X9@*SOH|0d4RsGWWeJKyw-NHj8K9P9W-@3l>Pyy)somx0`~lg;m>0vakz^5
zr<l})Aq&Qm@j__yXK?Rzr6^Nfau1wePb7X{{9Qg2MCkL>VmZsSen<D*{EL6s6+|#c
zE*T%UK&=Kzlv*3lm1K{9s968>LP7HPN(-Mz{LS8LhRZ(rEJ-5Hm&z62ujaE$4=9#m
zc~l0)T46lqC52C{uZo8kpa^)36HA7o)@YR&FUU35#;Y2@XhG^b1H$hgex$x_prWXl
z`a^@_NuH!BsPsGeRK5I<W@Hzc-N(le)~#qlJ`*JLK75P_jq|5J;|lVcbfqD2Qb~h2
z+n1<WK3oHX{bx0m_w9X(o;=<Ul^oj<2^{e@Zp{XFR@S=3&D|vSLrD^k*&4&#qPuYR
z@i!yfPb6)BE(_<Jj`mFv;oaYZOf47MFHjNWVI0$B>rWkiaQhX-D*N62n5xxg1_K%V
z^!&&O;~fD>y)@46L$w{nmlfv&MuOt0{v;;F71E{<6MtMUHZrrCNDki);Cdk6^zU^Y
z=VpsNkM5jM6F{LIjxC+CH6O{p)(ymRkj)6SS-H%e++|h{G3+U}eR@;{nPzeDXEktS
z|Jge=2rq`){WYunon=Y;m_pp!m)K`R!u=SO`eyN2G?B#VFA}w%{_x6tR_|GRiI-S8
zaW3U0PX0^MeV=YxG!rYaTq<sF|2)aR)h9~;pDyaHOn%yiG2N)<UBQ8rgD^f1{-&{Y
zW4B&o4sr<i2HhCyfc&@hGLB5!>xi)?uvMd0Vm|I(569%Xt@cr0v)<*-fV3b>z>^kq
zyybX^J$*S{gO=asxC9ggw6ci|BIt-+OnASg3Jvn_us6FyBy@zOqR`0$abjf<Wi{es
zIw!ErW0$sW5(^Gzg>@7hR?eLK<+o{Lm`QTI7I&@}m2sP`i*03?Zauc`Wjh^Klr<X%
z?yQ4x2Yk1#@CT&LpC0d+J9eQI(sdf0LhkSHVRl?sTRmx23gmKh#GYOZbTsogUbDq{
z-982&nn=l+Z=9d6wVhxI*6uj*EbaX<+U#Q!#*`4uchX^X+BYRK-gc+4q(i<ryY20C
z5P%<)-gxdabMR75y|<uSRV8*yw!pn9I7k}^D{A!7l|iE}xc)v)64vb9+j+*TJg4}9
z_Pqd8G2xlegr|Rz)6<$eO!)oN1D0Q=*Zs4kA?D{yrsdDdSDE~3aP??9FmWK7)c2L#
z-LbPeT551~I$Gdp^nPk>&xzX{UpIJ-h9(4tcM6`9ERX&5aJaVu6J&}{9pO$q{UjI4
zjoq@Ln-=D&q+Y!TgExmk<10V}u2?FD%)}lE`Epn@E%4TqW*F_g=yo1^aq~Nr`Azct
zb;M4!RFatXr`8`TZ}oW^o0?u(2IYu{NhIg{O~I*XaR2@4N5>u6`1z8~crZ~uT%Lfu
z7d5sIBnHNB`lX*j8v`dO=1KL)iqo-@$^!?;Mi<VCjF2lpZmRgxVj4a&;Id7xzzyHB
zOO?p7?<*&iQKj1-LyN`Q@0D0_&*vT68YC{>I~z1XsPLPT?jb&sh*BB9=YL?bNXJoz
zedKP5C7db+ye|KEM@?pEni!ub<R<|imv&xhFD>-LS7_yDrr4R%{~YQM`}0(~`JaEq
zJs+YscuRzWUOL<>>=)^?%oWAh1Mfp+H7o~nGpl%&i3vZqx}OKi#M6*+7}Z8*S4laH
zzr8_cnQ8sGkKw&fcS+(IGm%GsmRC82Qa+)jGSH>=5yv1M5hjv9y{Pm3`)rW_vTiax
zr|u!avYS7dL6_{D6+(OTQn4>OCp;P7Cmp-r%*p3VJG?+GM(Iq<`_EA({^uyC3<}z<
zK^YGv$2e@U90O@AoQK-sKXv&`ZZB3>7Z=mkrC&!6q8j15lgLFZt7dw&ZT;t*&nfbY
zn$w`h9t$q1(7<(KUom(^O%}dqUQw}6)owHu&z5@JJv^l6B>3v**K?u|N|4T8)GV;(
zR(mDv0y>l*NOfZ)oH&vKJ={IrHAh58;0$>7lj3&!Z~7<mD~ml;+>30BmN_hE6hYw}
z{)|oS`EaKVa3z?^>oyy0Gz2{wimKicP3@A}z9&wM(nB0P&##Q9-lZQEZFC}84@Euj
zIVF@yX32I#S*HjqEG(4LcoP4I0Cnq^uC>~Gh~miexW`O|-g|oWvTpzqI7kTQ9WPNU
zRc`ZofY$(Sn7agKX=$mO3`0XQAHWP_fMjIvR}8rc*Dpe6B#XBH`e~sAAu(B~P!F18
zaJ+Gu`qI#+SAibWWVUEfcX42_g^IOgg3km7JR)Ka<Rv1S0-I3rUeU8=>en+jKv@{M
zApTRY1g0zzs?~8cOVD-iB$J)SiUs_S>;?A7cVLppjf*wZAh48Z77k`C+VZND`rZio
zSNMsGnp^Fk1?lL|N#)lRHq5dB>nq_02=`ZuUm|@mw@3kAhcn-cK_I1N$s&1>wSfEI
zkWP_u%?u+t{)RgQ?vPiKN~jL%s|bkqMi_jbrCKrbBg=+@JgUkCoDbsX_8d*zOFcCx
z-$+BvY;lLI@2&C|w^a@Vk(+1^f_N-V!0Vpo-`4{)p~#`)^5HZHL#Z5Lz`RPfJ^X#W
z-b2PKq_Gw}0R*ndWx6f+N?<xEp-mO{jP#$?twFU_G3Bi>{79Lq^8Y%J5*Q@e_SIEA
z%^jm*-E5JUT>*qN9Dx1zSz#OQAZ7tu4KgDQRXCQ!@Ihp`(fLQYAw_)Pgd-q>nQCw}
zRg1wxIMYZ|{Rr`yed$rghW@75R|}e^2=K`ijS88jHcRb2XUQ<xrScg3KnmAlri_-5
z2HLjjnXO8Kp(MboCdHzVh~^Y;{hZ8Zsa}B<mJt0I%mU33c2}AFPki=|v(rUL=Wz_I
zHW`*_`EuogXe{u#g6_h3S22Po3kC`qK@aY9avKRoY7S6Uj-Re~Wn?#5kf<Rf$NR9+
zU9i5<+%qCa^TGuqOz%laNvZmZO(gUHSoE;RcLf@{2+4tjQH9TZM^!7shiWk`J+fQ_
zwQLf6WsSue2?+PfE`uayzWu}F-N{BCDh|J0<M-JATBd&@$QUSn7dsDcuube%b1a4E
zVTcWz>?-*#Q^6pUOd$_f2>-<Q>Fi>E(j1IQZ*6XFMn9<V<;%1yBXF0BkZk&4YI)$U
zzrjtj0$@;$)0{<v8HKOrxLMhOBKZf1BF5kS$)LZ-%d*e7BB1gTbBH1<bk}@cjCUJD
z{^h6IAjKMk9p7!PZ-M}hpwAy(k88V{MTV)QAOem5-Drm)4^yA{n9*d5Z04ju0vApM
zoq9W85H?aN<nyVl>gwv615QE61P1lK@8R?3nSzz0zo2yAtdWbQkoH%8TI=vra9of)
zX#4!L0!H<f=iPwD_M<2SK!X3=Hmm@RijNVs10KOI5=Tq?m@iA$e_`0E2MByZ44(iu
z6fwUwD=UjU?88yo|E^tSDa_-|E=ghM&@67CMRo-<I>)gbMl%1GTFoSgK-YYc^<u3^
z2)E;ooatfXf}R>g3qkB<fB@_Z?>NQZNsQWw<$7)9L|FtNir3mJz4FnegkrGyeY)u<
zE)V8jjutZ6STsmo(+v`HTuB7IALZ)bXCwg7ORUPU2P4@J4u$X=I8?lcic?;t|6~Hs
zC(YMRS|21gQw8#`c^$Tb=eAIA<D|-AUU{aUs!1pK{E%LoF4d^<ymLwAv`+outh=}h
zlv-856N>=($}@;z-3Z=ijRC{$d<U8Gedf}KM$77~w)?YJ260V5@*X09dVNb*{7g@B
zz^~7R@Xqa2s96dMvI%nScv4Z=rk~q#;c5Y`Y{Tjozv(|_X7rGjSK)unJ>s_C2o^_;
zzin{b<y2Rd|Nm{Sk_^OW-h8&4Xgx3W$6ZNh{|pd;MuclnvWT7P5Hm+$(5a7pW{ByJ
zrF2}d->Az4GkwqZcj1RA(3TcGa=S=?;+&~uB~&Ytj-!faF%VhDghQ!Y=TpevV}#!2
zd%|wn9W<8!=K3=oC`PH3sz>--=F^>v7NQ7to~*W^$}5TjlPCunL24k!1>pjt`Po)K
z%#i-2TaM0K3(e<5<Jl(Wl)!8y-;9(G^$`|x-zr@yL^N)8B3yzGQjQ|#Pxp#`xV^+2
z;Rg5U@WMz54()*gP3?{A7Mn({1oA??JvoQXQnDS!`{$RM77_eB+P0Ak$N3J283<^T
z+S=OEX;yyQu;%6yofcW0H)T1v5-g#)U{!f0umvc9O1?~j8HfnuK>?*m(DwlTKd=c%
z|FmR;%7<Jz;p%&-<ka?F$AzaJ7h(9sGFmUxC~ffLp(TRoH5>ROL{WZI0sFP0tFQ&;
zJ83Ay2XLboD}0}RxMH6w5$fKr`>)~X_yILttit%(y&!D*F-XsHi()9>ORaPfdskOT
z2cwxtv_5p}&1hBs%J#8CwQGaU$|kOh3+PU;ak&~!yCN&vd2Rx!wO-8YIx1EvR6gGs
zCNH91vJxY2C%8o)hUT((0-iA*2mzm+o_+)ATz!vhIQiU*MN=!(DfC|s8Cs-1HzDN5
z$0FRB+?F#8B_CUOJA}C~$=l&sAmLEtYc}xpxhy2_4EySjIaQXu-+ABvfpgxU=&^>?
z0`y+MLZ`5|ZF2f;_=+LGEGdcV84#8kfUqpcF4P1-{xa-(xp<JWY?2<l<KT6#+l!Du
zR1&6_CU%+U!(6M}Irw$Om|vCCKj|C7=wL&;PK;@=*vwUuK112y(fyJBSc=yJ|Go(h
ze0?bk$&dlEXI*lRk_lRCYpZgVHMqAn96R7DDFJ>{bFg1AGyS#Ntp0FD=#XQEF*qo2
zS_*(e8;ILG<(LHe4Pkcca>ss9mU~U8%;cXa>}u5t<nVc2e&b~bdS?dQ;Mn<E?`cC}
zn1(2V@A-x*WUu1}X5CbQJi0j!gNX1>FFv${54t`K9yBtVovDGq8h9aODg=#nc9?G&
z;O~D4CE}}{TyT;Qz4q!cp9SPd<4vw7>Ejf29q@AzK25JKpr13sd_`~I71{09gqm14
z5y@~N3}aC`_YTD?A}s3FpQO!m3I3k7G&k#%c$Y8W*JhuTq_2C?f)x-V*(}v<;$+lq
zR7Jc(CF0FPN~r+*^v`w)<%za02i)kxTkc6Vn&TVu-R}&)zKNVKR-vxz{${h>s9LEn
z{M_TH9T^3;?FBFP(pR7!Hj8L%YPH(<6`043X!W?+@EY1e#~1f`c`ISCF69Uh0|SF*
zr3{3=Ql5PZfd9FYGA5DG^pQPZ2wl%?qxtM0b7n3jSWFcOKcj#hNTZJrcgxPn@0k8w
zCN7CJ&}?wV^LJG6DY}WfoujuLpo2e(WsKVG(g#T^0Zdq;Sw)urK5J$Bwm>bY;C&Vo
z$0f6FW1d=q4l4N?3yl5_-e`1@5HLq^R8=FH{IWo7x&VY^k^u&iH>}k{f&=fN521WE
zBLE2AkDGu=t)K{GhssKQ95G*AxMS4wLhJzOr9h~0QlR@#s`ZoZJ5b&w;Id)hqB<%G
zf5rBE(4QAjL2PYjI8C)gm6m|bBwYK|IW+A`+KU%2GJ=`7df?{*q5PMRw~%J>5J9~p
z_7G(Vu?nY>)t`GL0qykas!%Khqd_@|Z~-^DXtC9kkKgM)8H_&5kqCI$nSIf5?<Q}D
zJOu9w=j}n`_&rZr0<)KoY+<(B)YLwz6dpp%-&j-n>n|Kp0L!c;GvwIHK{5iFod6g&
z*=HaD)XhrbxQG687uw<LDIhFARKN|?s<X+cnLdimn5t9(KM`eVpbH}P!OQ}E3D5a|
z(9d-`?5;?!a6DEsaO%n@D39UD(O`X}NCahkYx%NxQQ6GU*VosD?FKwS4*4~Of#!g<
zff<}eSG^Sr&gy6QG6aJXkbi3Mwg#Bc$9ck6CboKB+ne{1`M`NMyZs?t3=W(6Sgymf
zP-)<I@v_=_krQ_x+tu%%D*}!He3;LIiI*QNa;J%17oJP#VW6CI7ikB6d6VTXy;T?v
zBf)S6Us50vKA1>&r<ijy0AG&<QR?w(WgRo>38j3WQz)Ah38Mer!liKEEh>BPW|$JZ
z7I<^+z%Vy^2>Vhgf`Ef?S~b2TiD*-T)<v?hLjoJwGl13ZxCM}&-Tvt#F$ZQ5n4<r=
zxk<u8j6<%bX*YpBKLT4x$vjRh;ID6Pq}g!6$#D2t&Qm7H<G2$)!j(Hf71-so6-pE+
zq>8imLe;8yI7=`K<RlW<bgJJ#TRSh3WjXf-n!c}Bx53E-qoDhl|G`wTC@6vEbp@ef
zM$WLB2SL5M_}LX}IYI1OL^THY746{ZYSnwH)>@l(X_i>6Jm$YERf+6dL^lMd#V}T9
zJz__@OUm-)ua)}*W}$Qr#&dPc!^0!*rlN@5hS+jT=uOG1iw;WoOHlro;U3ojHVowd
zUR(OaR}+RXn%`_$Rh6hWfx2{&kS|2FNRbr8Ch*Um_5f)`n%p>!PXUseR)yeOL^B4b
z2$c8|L_E1~zgl}z{d;L8-sGSM4|ICyk5ajkeTH?o+d=cmI0lO%hVRH{!C=@IDG$~(
z7pV13fFOBVn$QG>j1U@UFQ{67Z!G&cSfHJH_Wx#(AToXkEBOPw8&oPO6zNAWy1x$y
zMDx8LZ-NFA8B0Gb>%<5E4^a^Xx7vUMQBwPc%7n$Fqe+}ru9OSw2|Nafn+~*w)>r=V
zNA2)p_%Jz*EF|G3$PYAZ%jDZa*-(&kHelr=;~(nkCZlmZfNPloW#wDI&~@}Za*z#@
zO4<(b^rZQ6u}7L!$Wn;hW)FJwCP&*{V=$m+#-ie_CR-;>M>F%u#6uN@grLAxiww9~
z5G~Unx7<7I=-^@sYzIkFZn%n$(b%T{#7G8TzV%q!<L#g(PNWfV1pllVVJOtAbLUf$
zZ%@GJ1{n?~ut93cSd4m+inDDj-Cb$8Y?ipS+Pqrc&7YubuCvy|L419EfdH2B6<)YR
zu8;Y=`DB3-NG|t5>_IR+FfB&$>^EcfnuHcXbu1eMTIj^x*Y@6^)XOwRfAzZ)fb;tx
zqBzTi;p_cjGMqyxn<!PW2`2OMfB|mWZFg6VL4o$*?gJB`6PMb9D4^Jq1YEMHT8V1*
z+M2$_e06@}w4?5SorjM!jAk+kA~MXaBDW!M$H=~6xXcqri*@;)78-|XY<F=qN@Ta2
zQG$7+{4}t=;r(P(rW>Qy_>v$W%wpk6xj=8<cce+$_%sYIYY|ARZt<^~lu6r5t@8|%
zYBM~qPe~uHH=|${y}^;qLBfA)wg@$g^R*pxsBIhKgZ^<^!fuJsRFGoyLOeO%5JoD(
zeEs@0rzWr!;svG-;a}SWeaUC5uL82?1d7(T|FzmE!fPPccxE34Ootr8>*OVvi(^p-
zZ0*i{U}nkbV6KW>JZeM&l>kW70dteXYScUSP1sK-$$aStjUQCn66^2*s3rHiA~jpC
zxoQJPKbgYQlaIbu7$y6o&pt!@he<>YcElzp=$|t(Xx;pl$YF$A12Xans=90z67x#)
z-fsAMM2ONnFeRwfrlU*&w}7%L8yrajs&oGTMpEDKA!9?ffIDs~6+8t+9TgR|*u#I)
zK`z{<^bM3F@tDogShioO6K&@mD@b5cd7bHWn_L(>?>dR}pD(l*N_*lXsC#brr-Fq@
z9UUrW3VV)X7%T}VS<OO0^)TbKqtgA{#GGrDmC^otAiR)S8|W=(aS<}8NOu)S9^beN
z76WemcJIrVFM|=Ih}d*e*&-lyxnw&f`tMunQ$wsw`q+I~(x9L%sz3G1nQmZ7w1>ch
zlt9s|EnhBWBIJ9ci^~2)elAGTl);kZzm)^>e>zQFOMG<_;PLA}a0S@G+c3%xZQzGt
zO3KB&dcI^j_tm9!b{^2CQLKjDZ+Z~Qb+H``O3`ROHgkhOn@%o;kjLQ^P6khvn~V${
zm{ZLQHsb+`fLq|qDF~}f4F7x0YDA#48~|#V_iR9p<_Vw^hL7=}19Hibpwnvuhp&M}
zlF%g4`MJ)d+c*XSFEhY=UA`cGtbrvAUM`W9fXW0!O%$FM3dt-T)gtMG#=v;D4pw&B
z6Wv_wV@BYyem$5jr6yYIwDJew)F#1Jsyx|PsRlFuZ-Dr8MFz{~^dO)I0Gl2`j+*`W
z(Llu;B*JH-ZcLHJ|9wXW$u(f&DCd3{8?>4at>Ik()Tw~Y4v~CK45rq0RblQEKs@3x
zwIc}SJF?oF-Or2k+Pp463)io4^W%JNGGMI+=TmtcO};*5zI=r<297r)^DiFqX9FwA
z{fcnW7swr>{WdyUZPT;Z?qZ@==_Vb^5r4IhrLhmKAUjo|r`NDePril$Un#W8{(9{l
zp$R&e2NK`xi{i?NQTf+k;rZI3wJW*OEG;!UhjE+_>-RuB+1Bi0yU=~~73JAXhV+j-
zINrdngP=<K2OrD%U^l!N76=ULP;S`GC#13;fhm#%AP>1{Thz>SZ$n0a5PL@v^N2<G
zqCFD?^d0BxdEJN~M|XMgEmk=)Jr-!)tp}07M@kBS%(}Aa*TSoU=nXeh($l?h@_p?W
zoEGbO_^4;#gkutp3%@7^sr<-+OrAMhaz}yJbBfsR107Sax}@~4hnf}gvEE7BeIXZc
zdwnf6eE*Cj1J_GyGL(FAf8%K0ly3cy+83qrQxpnv0Duyy;c1XfD}j@{4=m>y|A!{^
zrBbToWzb9$#jf%J5F7*Y#*~`L7MCL}z(`#B1vyc269XQk+dir6<|3!t(PTi`6L?Y2
zyQ4Z7VdstRkYW&0Ph8Z<I?V*|{Q;93owl4H^;0$?K`fFh4+BR;H;Ie$3aKdu(5vSF
zYbw)W4`w{9%fnt-M74)tQs?2|)O_P}+7kk2x$tj<T|~GKHavq~tIp^8zQH;XH@D3$
ze1){x@7j_cjgrDt7oo%T;HoGAV{i5=n)EGnWFMfWeE)EFm6RN~vxD|;5}+#9K05jy
z5$jESAg4fZz|D?nCxCzRCtk_<cRJDd7c2TJ!52nd9W53voh6{#(Lfa4eEX?sT43Pd
zRJI2awxx%GSK+H!{`HYZ;f=7c*;>1=a{nnHf}Lod1K~0G-B|q-Q!rF!sT*ez!cN3<
zkcO=-89_kFF*?0Z$QcQ>rFwdZ$^7vh?fY<$3Jt*rtb^!fvf6miO>~mmNg?Vv`UuiV
zgZ<+WmJ|m1m=4u~leFREp6kb;b*H=2gWpYO12&b&E;0-*ob*X2$8RZwCJ2fhr?Q|3
z;R<a>RJIMA<^u{8Z*izYh8dkXeqwgB=y%pH`b2=;$p_WMb0zUi+J7CJKdSya*{YjY
zk}&{qGKQ;PA9`h`0f_Mv-az?R`)Vae#rfGa_jki{zw-U|?ae$^JSPhY=t+<U&|5CB
z1cR=|rkCW^KVPh>L~FDwLsj}21cS&<vev-Pn;ZdhY7Tu9ntI9@)(|9bV0Os%ArNZi
z1L$`XI}L&6fj-r)x64+Q2w2Sm8>Ov|zU#mZocf7fFwrE?ka&`_vw6kFJ68U%lV-bh
zd<6iV?Wpe(5d@Lub8_FgAz>Ngz&TC=KeA=d%K$+`%E=1tCuxo+C)Vg*UoT&x*P#5^
zkn^F7y}NjF+91z2{%s@Q!&AhzB#O!(sUP_QM$Q|?u*`mAnm$xDtb-amx7p6A4fgo5
zDs>eo^zI3zpMT2;P!6A#C`j_?=L{h)8GQtk0YW3CV5bIhoq|m#V12<WVJbuwAvpf{
zIacQDNKZ8huYcqx0k0s*4@%b-w!&erL<j}u^$EBX44$oD`!c!@3-k^)c2GHRJ<sUl
zFBM(x^`1`z(A>lUH!y@x8kp=a?L|<iu#)khCCN|!Y)aBnu>buIdo<nZc~^}oO)iHA
znz<tI*#dpCw9MhZc0Okyz2W!~@f$P}?E)|$xS%p-IM8}y)V|McmdN%Y2x`k2^s~l6
zJd+2%?g+r2uu4V||15tya)btB$eX8LK9vm*00~LN^8^ws!tWaaxFVFl6Q6;$?4#G}
zk-CkvanzIjDuqQJ^KawXzHo*<iksbe-M-1I-avzgxfSstIU|D+KH4>lo;%p+5y=M2
zPO3I5E14#fS|%G0oq{+h_9shOqG7hl%UYCKVX5elU5-s7tHB}AJYN=z@_sx?Ka{GE
z<a0Q8o;>D*XrqBAsqv>A6uy@F3a)~Il-EgB?9TiLm9ZKP1IIpBhT;F&lO!fY8EC-)
zC8UNM^2b8StCI;|&oBAW4Q4T9Q@U1GRVoHX>9qnVwQB|ivL7@QU(^sbx*S<}8Y@-8
z)WDPKD+*|RL!JO=2XGi24eg>P&zmD84E1_rz!BPh?6^nr`9jYlltW@e=waB<k_iwm
z>@h&pOC_{vHjM654Z9LFZ9+Q^zeD9AhfMrV^PF=zTcigoe_d05<Q(`ThA}Rfz!MH(
zD6MP7bXA4vZ_37-@^P>^nv{y|-TNX|!9ucn*&quV1Gcsp*j*VcJ#zTxIN;5iwKeCP
zH$oOPn<Z$(;oN^+!Bh$@jM@f~XHU#}WGAeLkdNTQOcO4LjpF--B^ReA9YJy^>nP+j
z&dGXdl^)TBxSilhy6bkMY9e0{mB-c9)vb&hl+lAk`A5qHjUvYD;xT2P6rv$nj*h>~
zg!-FBWdEXBLQ^78wY3I@=%WatUke@qHEG@`eHVKxs`zsejNNH(49)ljygDh(@7W4H
zGb$8T2#Q!cP`yCJWnvh1p&1q$YzB6t=JTaRq1PAsBgsH|k8VKrOFtw5e;Ya3#<JSc
z4R00!m9DmVhh4ad)p^C})h9PYZnrl^+PTxHwf?^NWkCz+oN1^g%v_*X_lB+Wq2T?g
zUk!CJQ!pwld>oZRdQqf{86MvuFvGc2$cN)8MsUR+;v^V`Txm9R8o(W@GoL7e&fu;|
zYPFs52iU3*ck1vbJ_vAX<Ubvm;d6n4y7`v?rM#_^wkHJu>Y|Vj^_^z_d^)UXlR&3%
zOa9j;TEMKaD=?}m7Lp>xVKIDS5fgtRmVB80d!F45!<Gvo(jRE>NJ#NJN(p}}i(#Mz
zu{iBgll!?Lf)5v%;I7GL7Id7b%U5!?1G5Q%_89Dh3lm}S9eCZ2NHC)dXk5iv9IfP%
zqf(<=jR$cK>9&_r0r=rLOzNx?V3+q2F8pAVNyZY5@N$2$kGAR4sV1tyV@`v&@^0i&
zJkx0cdN3LcZ>^=Mr3PbMf$QnUFYtvRp5mwd_`6Ce7+H+!93BZsh^d!plKvDvoiLN)
zzNvIRRPQotHT_QnW@iNR#{U>75oD?}2Zjwha61R2pjg3*#zptDLqH;x0P!HagA-0h
z{4PW#(f$FxeJfH#k!B#B=Z8myXmbBMNWN@R0Z40A0NYN!ksaj~F!%_sDZ#psqxtZB
z+)jax8@+Bdt|o>qb7tAq1#n$wvB>p|3g%YS=^?(Mx^~7dgTm*WvryPtPZT76&)D1k
z{0q^DF(fHa`NZl}E7$qZ`e|-9_GSWf;l~hnMv*MZa+GRtWTmL!LsOF~Hhx0jtuf_O
z`>v(7VC<y)!)8`)vo`{PgeLU-O&O!yI{8?Qxr%wJ8GwdVYs{3)OO?hK3>kq4%_iYL
zdtCn-!qSsn`w)q|N+BIP1^~GHb0;$-y4QQhoG#z{{Nij_?7!~PS?*d!FwNo;L(uj_
z^<P38fQaAh>G9!{oG*Fv=NyaAAfVa@K3Ta4QG4gjeI}_20BQ++p<uwE=|&n(Ve^}C
zlWJt&XV`7^;5ncfRcNl$H_{~)6cp^dVg_O9Gv?K`vXPig`$`gPph}lP#trSWySoxd
z9Lzzo>yY$%p+C~2quu5+=nyo&45d`fvl9-x8^y<mHhrU5Dde^A(-~8k2Kl=g)gk(?
zEWtJ%?aE%Kla=LL#2mK};GZ!+y4l4)hea5&W^8Lx8z4E)<>r0o@lTYPg8~7i&a|nK
zzArp0vj)s0k&}O{jMu+9pF??Btc5|X3b^$d8N#H8K(Y7R^Ab~Au34#Ak94Ix1P@sY
zS0ci#1GGEno(KRLTD()?Lq;2|O-v9F%qPiI&cPgy=LQ|Ruwz^zQE%#-APFF@y?#-M
z{0FXwBY96Ptv5`p%OR{>>j(>%Z|b>S4$PkQ<Es1~0FzmoUwtzy^Zc^mH$zEc9gswc
z)F+e_iF#XjvvBr{x}uR5;hA*@1xgmxgAzEWG%hSbknME5#1n$~TPKNDCGc5korp1~
z`A1QNo&IJ58EpiyK=%jvz{Qb>v^ulL&Yla|sACaJGV}Tya#2_f5=^gABPzK9=hvJF
zlD3bkC1R|pEG88B(s2tk&8@SxW6Yj+lOZaF+o_6WQU<B-ypW7XcNouCF*~o&bNePo
zmcR69#ReaEs|3Mb9+L0!y)9e4W`6j2K(f&E1eHLqqBS>UD--d0hHC+b3=Z|Rv@7xT
z_(cAo+)hX^%)8UIFB_oZ>~yiGGyqcjUwFR7)i^Y5-rK~UYC#E3-P#zO8pP&-{<t<)
zD;<PBq{WLWR|q##R|g&iod@?an*F6qn}`qr@b4Lx-M`1P{6@~jY@7=wR3j?@C>>)m
zVYh^5gbxBhVub<j@%XncvR%JSOWsW&NoUNm$S>C#2wZ}NEw1Ek!%#n?yQ=@HlTs0g
z_vbk;_b1c0S$53dO{w1T|ND+TuS_<Wb*QjaW0wtEeOK}toUXGj+`=V02Nv*`K9Ao0
zx_%dC*1FMx&>9;bhrAcXs(darj4LWodY*z<I!JFoS>X^xADm^%+a?^d&abha@hQE-
z8_^s6`_>&8mDuJp1HA9sJA@NcH*x_9X9tb(%`yb#Z!|ImwJNTni8@G`ejjw{RDMOU
zQw*}NSo_3)Y;U2wtJX(eRzAy;b^9@6R8=!W4*`U3Uvz}ahcxDe>7OP)GktS^$>(U0
zM>|Kb)73&s$}E@j6{*d1`l$F&?BoCa>F!M5Y-a{s1wnJbb{uz=|75VDK%gzUcxA=f
z8#v}il~cuIpn`Uwq=&o+oP$-ejV;O4_Ox?pmkjYGwkPz?uOQ@2;5P?MWx=2iJT)1b
zE%FD3F~C+&IgB>)WHOqrP?hVMBVx`^tJchX@z;}u@+??*yjyS#_a5^^a}Tb3TzriU
z@^#99lIYJtrGWtBgi;V`@tt9E;Fm5Z#=D|&Z6fsrMLpkA?=eY4c*KIdKZdcUOLG@{
z*6h5FPaL3Q)NjG{Ku)*miLkoZ(27B;Ter<i7je?&COv6~qNYeCjQ0&pxl$dpZj&1*
z{`Z%OC!l`}DvdJlPX5Dga!--U?Sf@I6rCv%jxYH>OQ2BJG%{eEx63e#Qx+xk1V78`
z_!bjrxJFAvWC(%8Fqiz!rC)`5yJQM&7M0y$@6G31!0&&3$z`#4AeqW*Pnf~y?TeVi
zAlY`m6hWs>Bbml&o6U2M+h5%ihRr%a&u!BmPo9O9x%(@o%Z%7m!tmU&dG2(s(fKv@
zP;$HjMVNHdy<VKF18OL`Zk>%`gz5}r`>J9ow*th2k#4e`I|Jmy(Nos=1;vYOLxl=I
zD9V^B{1Y4o_{-m{IbQZ_%cL;~+`L0A_~CkA{P}0e#bP_De~_M7G=~SW$6X9$^`%N<
zxBaDCaX99KMz4)I-i=&>WGZR#IRSVeXbt{99;mTZO3LXm3JMI_@R~o4*RQOv7bWEf
zczr5k|DV6KedjIvAr>+1hoar>ZMY<fj{UMx>p|;>4ib(w?2lw05$N}y-xWOicQWSV
z=FMj2d0&i&C+>YnX7jBSer)&j-7sw%v+)958{5+B0$TU@Qr-%TLY--2<n!W5288EJ
zm7fkS)i2$dmy#ndU-GRP(}lCY)b8@A=K>>DUF^=U18=6x6aA;R6DpMz8v3xeLLG(a
zZ;m5OZF{(EtU`$}2bOueu?d}AZ|_CZTlQsI4obf^@G1~-%GK4O9)L?YZ9o}yi+IEI
zX{Wvwqxyn-1166DOg9YX7GMs!EBs9=0YiU^b;l{<`4S2VUZY5&K?I`w%pQxO7Mvd_
zBhuIt0Cu=tnJIO$)^TDUk8pIq`jjkgIcuA@y!e>F_uYFaRd!Il3@g&nR4MZJ{9b2|
z>{Pa_tJYg@6yt$pUsz-!>8CQ?_srepstBaX0lh9Gg!2K#@o-ZYL5UZoWA#~V-lW)Q
zO{(O9D0$M8CS$42ulKU&rkT}pxQCP9#@rM07mzI0<XX1Q(t5WxvwIG}&}lcu6T`Bo
zd@B3B=z8*Leudw=Kmz%o|E=bf#k8f|ljUSAC_jlUy8O;T+#Z-%dKs3NF{$_2i5$L&
zT0TPoR8=riH{?Z5%~o=dvC*jXK%<THqRVlP2!Plw%Nox2R~Z-cpuS91c943(jj5k_
z`I*gTdT#6)=vcg|nFvDqO${@L+ZDdQ^L?@Arp@~QW9u)2>Ik<sP%v1q;1(Q$ySuwX
za0?LJ-E{{K4#C~sgS!QHm*BE-cY@sJ+;e8WnwtM~Rd4#0wH{l)Pds|eEo#n%JbpG-
z7!dxBt23N6!{c*gt^C&d6>gdB!0`LYz~njFxIueDXQ^I1!tSIweZ+8c5Y3Frzs5?7
zndWKM$(}x`647~`daLCxRx>63`Yri`ieFP@=MQ$beAP<6L!{pAXFSd|3^KVK4fzfC
zwIZd@h#0;R`M|t9%Mt6;c;E_jdS`}wB35AAAItBix-X=Z`uY`>LN+z}>2)Sh5`kY~
zI(94nK<@%4joC1ulpZ-IbbS0QlW+R2GQCvyFk!A<*eRS?Y9lqWk8i68sonD}Oe5P{
zayEv)aAe+U=0?a|?(d(8&y3gej?_}*283L+RjFt{pW!ZA_qmBAcUuR2T$MU`?bMz)
zuAip|#YXw`aq7iqU5~=0J68fA-v(?&#(%S-3E*gTTH|JCWR7A#+BI|DfKKNFC2w9*
zvAJZF;*W~}*to96s;DDF|I$MAw7&l2?<aOMVWy+$$_V~W!=j(#GU=Rlm;v^)W{Ru@
zN>Z@Mgh>J3?2ucpml(wslbuk%#|hi+V<*)OMt<?cV=gIXvc#`6+KCuTAs6zwpDXtt
zHNc}F!L2Bt@sW~rx&Diu>SPeV#%9c&-l0^9IPc1zApAQ!JZoNJyqHass3}|at+u_Q
zOL6=tL6Uoog<REi8bzpUu>PhxkyN75=C6oUOm<FyXt(eAy04{**=hleTAQS+C2WI4
z`rTZmxjn5xWBm;&;Gx4LjoqLd&BjXGh5Cq+Pba8Z8paG;=Y8&1s*P;>PeX{4$|B16
z8X|~{ljE2puWPX)KjijkqzhivuNXPA4WDrK1r~)O>lAb(+1=uh8{l&CAhXi$+E#(`
z#-QC902Cc|8YwYu?&oV#8?RI0tY(vB4JSf=pORig(rUY6KX!hsLHn`e#_XH_wXs6m
z^kQa@PF8IQ3-58MVmO`3U3%4ZYZDi`n3j1k&^pXWLTska(_601eG`;8f_0dgg29~c
zd^LxA#P_|1e?+~~YA^PA0CK5dnHp>(js}E?G9+Ygukwk8w(aI}_IGHdRFkst_YX+Z
z`5*R7w9vfv?Wh5ERv_|az6NGToy8z(Cr{l*{8GLA)mb3&2j7?O<1FJAJmq7e>v@64
zqb23wg^6n_Ohl<54?PFyGBg8o@yq$7xNG}W>^XWSj5ue1VAk~Y@G_5W+<6-2fmrk6
zOYt4M<p}%A>OlHcky)jWIosUKor4i1KX1J!*8C@_DND=DBOla~bxcJ<L~Q2DVWb}3
z3?U+ZA2q95F=k}^Mc5!p9*d3FK(z)Z2}@@}(Yq|8YeV4;gWw`}Yn``A1jFBZsi?A{
z2#fd&Zy`Pt*?DFPc+;T2jjG|h!-vJlZ}E-uIFsoT%k6b4EA^HohQ66@zUdkayJr42
z^0(QgziSkNTbs^T8G;9?A=59`wd<t)TU7d^$MG#rN(^Tt`Ta2TcpEXOASh>gdDsdE
z9$FA_5-U_Je81mw(sYO95OMy|pw|}e3;k5!=j)gI;|C73_oZ$S2@&MuGDtoj-dJkn
zrU%L|e~<1h2^<+21^;f`h-rLj%j&dU{hHx4L~T2o?`VU)$qq4K965ECu<@1y9Er*e
zE4PQO+!7z0IZ^EO67k_!1ma^$yGAvAZpv<6Xd{*4zIaNvCogG1R2wo5#yahdkt&>a
z<D$cXopN#wT8v;uGU<vBMv-LG*u(hea}!`tgeS9vijRy0#hi($zRD()i?f+e=Cc;c
z9@$b;@6BmuEO}gP`mg>}9^@nY1Y6>;u0b(wj;F!E&>v;$l<F$!zS@fPov+yn??`PZ
z@M?F0f%EX0fKwk_-j{=<TFdr`SL7HCh*JDfp*_|R{d>0{%&WY|G(+py3Lwm{I0}?u
zVLe{CKYv%hoyZUx)5N+!M&^yLZu}IyAHYzutTB-KzH%BbslUi-lWjy<mIZd*orr`s
zR^{sk%1V1n7hO5FTa#>}(S5A6+fqmkImd#Wn_yS4=(d&slvrQj97t=pBpjY$dXBfq
zq)LBIk%U^NZFt@P0$Y(pu7fU^qlg7Es^Fvynax#meU@X(NHRjj{#uLnHEKjzv35Q-
zDToHOe^wf|4?^XW(l663FB)TW+pUSc0z-mF{O)ay)3|TlS6Y3k=CSc!-{6s6vetM*
z)sg{1GC&wF2nptE)RFoghe!al%EPN{%DQri$)!;*krumWI;#Xi7;gQcNNIcf?(KZ}
zqRZ0>5AHfe{XGi2Ge^!}_cIj-hZ^og2i<l2*L^?+5%q^MRCGLz9n(w_EPd(<cTGBH
zzffQMi@@o0y_{>@pkB>R@WX5&B3MP)@koS}n_Pz<)aZl{PNQ_C?Pvzz=TYw)b%heC
z9%;hrAp7-Jh0M150>ZYtshII}nQLijZ(y=BLDMf3%{ds^EP7ldVveL}0zSplmHNCg
zb=ql-o_q<pWVs%{#cbXopMh5>VwJ+HUy$p7SQvM7JC@smErCfdfRbxSb{wJ5MH>H`
zr53awbw+@Z_`fd%Y{hZNzN+{<l^Bf)dn{v`qqQ@v#rcnArUosH05XfZ>B?lpeRSX&
z^LUiy<GN4+EC@u)7dH>Rslh(IGlFqVU5D+BW?(rO4B);mhr8?HR1PiL4URJ2R$wyh
z%FU#T&aIe6GAg+1-$UwzfQNzzv0!jSE{A7dCH*hcpIBtPdc;~8jg<xz9fQ{MzHkw>
z`8Y)C)fy632gKYyS{5xhv3JYFaLz1KKB40ecwY!twn@Z9YPZpJEunP&BB%$FxX<#I
zsn&h8U8{{x^V|59#%`ZWp5Bp0hnRl0T<3{#GTFzDEI;f+6|{~(fhp=O#)@6M%OWpF
zQoj&|CXeV}b9ni5i~k6_1x%e3Lghq=j4L+hw@j0i5fd_0B7(fS#eYj&IeWYUEB_J$
ze4_S=qHje}c&N~uR)&JCKo+_z1Ygj}z$-eS6;i>0r0KE(_Yp*kF6H#306VTw{7`T5
zCsL#eRvFwiFPCNCQKAzR1immybplO6DarThDQ6IfW)9YV0d7{I<}zmX)91eD<?RCh
zsS|_RBVvlSVLC9TVlsC`&9Jh;1G$ARb?$7?<3fe|!mcZ}nNH(%=8`b(SF#8#cdMa|
zBn?+Kg9g!=MlXVX&A(oKg`fRP{6AECqm7zNRRSCPez&|%vj2KeTN>X!bTRm1Pk12C
z`BX~^xDAy5a~sst?F_YaU+JAb?2e_;$DN0u9l#<`75aL;$b;sTlrD|Z7lz!x!gh~c
zg)7dCQxaWYbF)kDDu=DhkX@exyYFv~L-5s%Ml8!-_j%Il_p_J_jdYqj2DyIdFm#Cq
zQ)U#LV;dB$ZLg;WWoD!XxOcs*sKQM)Sn>g>BPtj8>)cjqm7@2-sqtMJuVPpuNr?eV
z<TB{9KUsDP`~Cu>?Bqh5OJIvUF$3T~!#&lRO{eN8ZzmoP)nZU<RHU%mFMMh}_SR18
z_K(+EGC+yxrTdpTW%Z^xcN<e<j>5(eF6rlSyAq$pqe`jW?wzctB8)`J5p+7`5Nhy2
zPWtfJYD>#wHmTAJAMNPML6XmJ4DN6vN4R!It;mQ&r`aTRwOv@yrkVUs*8Xa8|A@oY
zb!JX^wb}8oePj-m;3i<s?J7}JDGRQ+kfSTNy_aBQc?Y(%3e}^WCcB%UfK#vjty|<!
z8rRJ0xt~BWjz6^(t`Hk+^F5l1|0;4H*J*^Ty<MjOP%6Y2b$W^u%IOS<F|5>^b<|N-
zQ}{afH`F>?Zk2Dou%)pL-jrg<4HQlE`9wXKXOTUWfmuQUk&-Ymg#P`iGtf)5QW@NW
zI-=YD(>jH|@7o4bip`eItoQx-)b&v_!SLYz>l-8gqupo!S77X$a$yBLbxh6Z87(mJ
z{oBeEW3`0uZ%TWQ9dF|2)RoG1Ss>V}2(5ulIUT*!a$zf!-Q?IdUxMi;Qm7kUwX)Ns
z7JEZ3m#M;;NafpAZeJ5*6xdyjJcCAUGV5oNLTZIJMv}#W7#(quGsy{qK~*HXkZSHX
z?L%sul}8ebxc#v>y(Q&rD%%zz(3uht-IoVwkWuA(Kj$BdI&73HfoKv!D4Q@v5yu@Z
z9=$#OBuXF?3z(_l|NW+o8I+1=k<6r9J5um<RV(`>to%*EHsI@XlfT%(O=?}xco?=V
z-xT&1-3Pp<ttc5wd|#=qX^`rYKQI7xRpE(jsnfILhHt9W8bvmpqL|$u3h9m%%8%{Q
z+(Hr-?<Fd@Z&}+JXR6DH3?kaW`yQVh=7KpBx;I25uxPw6=|-P4WAHF?B$0L~Dzq!e
ztm6uns&fMlLP!zf=4e)F0waS{tKB9zQc4e4$!O7`_tSHSY<oW1H6EodClT>DNwE~y
z3%|U2M+0ZE`(RNDdJ;3J#-y?<w9hMs@z*`{JZ%ZoXkeo7ID^)^7E?nJ?CHTXsnu22
zW3#`i!5$Nqk-}<HKNbb<zul))pABo8S8DQuR-Ip&4c34wx&o0OXVsmt#)s<$XKREt
za!Nl0DK-B52>yojIqhYicrI6Krlx_L8w`xvVK@3~OMCwh_sMOVLdW@x=B4=tLC^C!
zB&+jP+Sq-)(3@#}*7nU^P1EAo{`#_*5|K&S|B_eX;oldef(dK;WE6Oj=f_Zci*FtD
zY64O!R{Gq9?1i=4HyB>lf1W3Mjr|hH_2tv%y3=VrsEU+bSyg8|YAuGt2JhVZkFHK5
zC<|Na-s*?cqURISsXr*ZS|!rSi1;nfP?%PQ?~M*yM_KXyy{_r&GI=GiG^3R2>`h?x
z#ilvXYjmip_pDmV(X66=c3#Xb@JtP#w^GeZI1CC)9cOpj{pxA;;qEE-aidj_>hQA$
zXVKH-H*J$XFr<8IjsZi^GH&pc;DB*-EYYbokt;~~qh--Zu~Kq0_ISDk(1a8inbN9s
z42&0C;D7${Y0#vMiCgajt!7cZa%eZ}<uRxvU$!v(=KiYRWq*a@h(U4dxyhDlt<yW%
z%kA9g3I!iBgx&Z^{E(!@FaDewG_7*8d#I|@X65-aCU`X2*J2j@wwn@7C`1~8Ld5w?
zr{1#qe&6{t?MAG|aa^|5>y#IZUOxm6bWhgPb5+v=Wzu39KOm>dCngz5IhK4U<eD#w
zH8T5?H1Ja!R_?uSqfPFZ%ntV-Pwwidg<8(!WGYi7v}Ax#HqjFyYx&=zk<=Pj24MaR
ztp9UPWHvZlK08i6_Pw{7E58{pzBx?3UjQ!#fqWzGx8zCtli9`G(}K#;370ogXF+YR
z>Gsp@`7tGVs6oX)iPQb`-81jLfkYrgdTu1)CSz9pWSBN!X8*=yA4Jv^Z@BfVPb;C6
zrP9$5Vg}rc=-LhImtEr>b5b^x3iUlVWp(8*vE(Ul)#o-&)lo*!Jzb=p`K<|mO>Vh7
zAxs=9sY9yzV~$;n=#vFa)1=tuReX^XRGUA2I9tqNIs5gyZQ0CT#J=i9`iB59?XOJ}
zgomnAh?YVIg6*?c|H}dZ=sE|HR<}#`Odhw`@95yem-IPl3*W%@I<vUPQN(eZ;P1z;
zBQ5bR7L&-lt|mfjoo;a}O|}g2xrfoMMC+}R8*`<FnD8xWbmb-ir`{<6029=JP60E*
z<vS6&noO@Oe}{NWbMw4qi?`ZU-QMAI2d(g#PetaJnO&7WTt5-sUN^+uhDZ(U+r@%9
z?{9mYOhGT<XcRLS;Sl2^stm+`Yj=s3M{nDRww@P-Vr`s92M&D>*nBzX)h;_DJT17W
zb`XJ|YCs^5bH%F;m!yeIZu)T&Lt}B$%^$S`5jZiFQTPmf&0NoP0DeQSQluus0~<}e
zO8408I9`T2*z7RSZgeQ><#hh>^zUQ#<}L0@ttnJbf~tqfVFpq)k=zVdm7ap@?}1wK
z@_A1W+j>0KvP^Zjo5jk;&KwbVFkKv-z;*q0IYjzn5+FELdxs{n{@=E;?CFq~edu~n
zU?1hZ%d{Fb>UQoC*oPc-zXc(kQEx<*5^bllnU^^5?x(MKK0QHI@?HmdsYy$9&=w`@
zjeLRLkH25nbXQ_pRo#7Q_|0g~9tXJv@?bxn62q%~iJrDu@8zH9w1c^wu|x3pGm1$O
zoyB7~fHR{Leib2hMJszV%x@Le@qK2FKGkffHLbQ9$@{+La(OUUs;X8XjTUj0J`#k4
zDpuT3bftrb^6w!0!{1-5s1u?kc9CC`lX3rE7mf64AJ<|O7tmzu>7FivITpB_9Z@}S
zn9MS`wxZy{cAK9$8<_c^F}NF16;Xi^&3}{fe#<9~rP?rBYcT_Yp`a)tRp@y6k&i!0
z`9m&K>QdS!a#X3Jm1nPV@j*Hb!`At>(%6bP+1Tg&cDjb@G~n`y3Jem*VDl^uANAUl
zKwvL^{{d8IXDw$e?}zqBA%oDhtN*hbQUA9aeII|l??&h+K#@dV-JET!4Rt~bHST8U
z7&>AqL!hGcSP?dka3g1``1u!??>4t&FcK+l=f^UW&^r))pM_^OiCrY~<#`a&%XcW)
zW|mHbdqdg5*mYP??xxZcr1&mZ(qc^o*;&W7-#Z7_Rvsd}nz}OK=DyU60A|t=Jq{Yb
zLlc9IGr}9{%|pwov>lx24RbKlVZT>81d;Ue?FJBbXK<L_Ngtm*3HXVYqI+2DF<h=#
z;1`sl|0$B{C9UiD<3r(8`2NEi-1QwiCA%}xHeDwL76riN-dFS5)@YN&_wOAhP@pm#
zLr%inXCTsV%)fCo=lf_X<-A>)l1Nbb`qv9WN4I!C3TkN(2q~7p<d|f6=^@TKv}@&h
z2Cq(h8A<v=v^A+BOPcZb2{B$sq<2@TeTmm!Nd7}Vrl^<2IzIYU-$YkhEV1IJF9AUf
zBwn}=y=O&O7!q!pyG8)C&Kq`NF70rWCY_QflgwCY_-kE&U_`&&B^c<_8BZUNDR>{i
z!yW0!cBEID=Re~qioSXhJ%n$UEs(1|Rz&SLOzZq{c@!{`qI&CIY{b$?H0u1t?!8Yt
z$zS<UX2b#ulxW=_{?E~i<kyMZw)c*1|F=XV{XZp|d&2pV%(&sgwrb>8k(SpBLic3G
zxQV-)OTKEj=yPH7^S?}yEGH}(3(;$?lgv8neTo-J(T}$?PQF-w47@PK`&r3jEv3%e
z3o3?pROrdLN#9yRtNSejnN-Tr7I)kYl#jPdZ;o6S6zHXIV3bZN+~=pbg5~1D3<cwF
zr5J;$BVTrWt2c7c_h@DQo-L#WvS_eCPyLq<oW9<FRg+Em4Df*kS+*eKJ=Db(ZzY|_
zNu8_vHVjI+@85G=IwMr!oXuW|D1G<OI@FpBRdw!xV7zk42Ij*!j^5O3P7!_-!*wa|
z0^rO+Q|*cHa6C2mmr^axcsn-xvzl{)mT>?M@CJFTIZytt^UVA6e-4}0{~b1K2A$<p
zQ!&(aw9CN@{6V{QC_)9hz2u5Er1Zqk8@9R&U+PT5i<E@B)a1?Cm4116eI4STcP7sj
z@5P*)s%)|}dZBk2d=q%_y`yQMD|OED+F91ct35?-pi4{}F_RHjKq$&#(0sPMiRPQj
z)1uY!c}9b2f00i79h`>Fy3!0q*05RHW_6Gr;M@NTCV{(lqO`h{|7R%FMxb}&dz9q3
zX&txi?h^d}FOLMULj?ejBqP;QMEhiZ>Xjy7PwMI6@p*ivc68EZ2h0#~tu5w7FW=Ap
zua?92`9axP2tLG?=ihqj^}>v*=z>wVb$-LoVDMr9mChX6O8yk;H3uf5wOI6aSi{Ta
zbL5Y<FPkMBpR2taCqFZI<Vg??NOt>hgoK+q0)Jszshd2i?7R19Oh`P}(y$qZ?OO7$
zASksO0e!)Mus{C7YrDkFlQAf?{m=z?IGBFQ1x2Blc|*b2*gl+8>9$}A^@jv>KW5r4
zh;v-AV?nPNxx7mGnVY3V@vNX%`Ak)S4I50^kULLk?0H=nru0ykz#o*F*)OzPfIaF{
zax`}RTT9(Aw9B0$lWo1*y{CRh=fnJXynNC8P-=o7V|J3C(8Uik!n^!tmi*b8wA9k?
zI-`H<>+O@kvjN2^>=VBw><*hxhGDmEfAsWO0{^T8Mq3@LwJX@e?`|hUh6zmeL}{Em
zO0*jFaIR-d(P>Q<Vy2@BMYHA;Gtq0_)Y>xzbJdefH?}(=H$WzX_T+i|psGGn&=Urw
zQhqk?vr=Qk<vX88t5zxD@Z)cRGE}};r|UmXk?#iiI#?^93x^bbXE0OV4waHBK)m!7
z7~0i88^F?Pw^Q4$b+XBOnZy9+;kOF{4oWG8%y!O|36}pqpTHS_S6<K@=IkRE1^(_w
zWAjhvAsdr@@5B@9|IICR`RDgS<>+2X=hbs<G|JJc&V_gc@0_nn@0cbG{Bd>kwv2~>
zqgoq3POsRyODpiOD7HmwXpw5lc~NaGSAT5rw7B?BOwdql&&}QDmm}q#X6&|(sB~Hs
zso#U=otSs00@V5ypL_LvT`bFrrUT?$8bFRTn_TJ(xR(bTdxQ4WuDwgWrFJI{={_{j
zFN?u<es+d}LuhUdeW%@vpV<Bkh}Iiz%$7_W(=!;cXUD{#%vT%4RAmSlvJ3o{5dF|;
zap*+D54t;PVOFn`Op$vOZELjqHR0%tMU7|V1iJ+=Pw<qkR%+F>K`(4HB%3fW{H=0a
z_G?7}{i7-DE%{HD!Tv*Er%v|of6kbW%UEkL=6?k6`{$`FrFb4bq_Yjc`*VU*glr?9
zP^}Xm9x)yY$WN!I(2Ngmox5UvaG3Op0K{bmeEaq_`!$eSGn%0OEBCaz#mOHUcF>m5
zN{tuA9-YcB&MIvQSF8E2zsO!VK+`kzMunesJAIIe<z~b$55MlOCJc$$$IP9%dD_+^
z^)~gM%lHrPAGl8O{iS`-34c7&c|Y5&77I3eBY&im0jRt8JAC?%vLumZ37zqP{?nQ1
zF2H0Vp7FfXG|!8hidT99hZxYY*<ySD2Bgk^Xq{|EE_-UGtsLr7bWD{|;I;2z@5!Cw
zZuR<n!g(3B4>kk!HnWsE@Mezt!giZFCT{L0Pi8Yr*wf~-abzmoij;f64dXCsOG5p-
zjSFy^OJ_*u0sz7V6AVwM{XBucT%)Q8Q)bl~ZGw88PU<N=VaC753M^+!mAXVEXLA`q
zhhvT%>>K!8yT0L=!(!VV?JDp2@`s0Y=A@2C-QA>Q2mKMky~6=QW|Qhv$^if6OQne8
z@@#+)P%)AEcy{kVWS?YoT>9tBo#SVXpkd5mvJ4)3Aw&=^DPKO`_5e@@eGkCd3<GL_
zhEF2q|EVWU^S8f63eSlH{p;uh(elFoTL=7BUe1o#k1_px)tb{6)@4AS&J@z}<w_tK
zv=g$*>_@!pw?n_dF(Kvt>%Ec9<%k5KiI7>D7j6Bun{A3<QvEa?(|fo*do!^7C6CT-
zwrCD=@a$*ijkXO!$}DG7omVk-*U_;Y5FV3PmPi*;uUjcgQ3v_rWGWk+LNJoRl23oX
zx{*CoH2tUoS1OHFqL<I>Stwr`4v$0$Bv0Ictq<1P?pZMj2r81X+05i+<M`Evs>yrB
zC)6BuHoL2AC2jZLp5+WuU3Y70(h;}~FMLPelqc;nOI?Q1UGB3SrKV+d<8F_}Jj-?Y
z2m>*JR&6I-eewb`^0wjNs0IYXvCeuu&&>!nYS=Ds;pC1&`*{R+aXr>x_Cw>pOiqO|
z(5zw4BsKIM2bgk8G=9?okUPi!Aa?|slI1NqJ9Jjgf|pHIcPW9{D*yD<8?uCnFY$&z
z>hropL79&Yw%0R2Yet<^;~S-iVLg0riRy{ac5jKEx{%S>I5CuEz`8ael7sACn@jxp
zK_xFv*rJopME-kK!4}>mnf{RN%|X)m=r<l`)K~h#1Otw1`?XF0r%O!iuerR~?nmr}
zhY$-mB!>G3fF;!NXKNr~zyqGhd2jd+)U#cfAgIS%rreY(-{rC#YO>Z#`R1;GFGu>C
z^#W`1dQa?<4&Q}DAli0+mM^m-iO-2cbDqg#PuOB-Rm+)XycxPV>9XgD>PdA^qO{fZ
zM^73t%R)JnhCj8sD3*ozTskYo9#Fvnc;^o?ka7FjGG}Q0`@NQOByc^C$qL>f9LBh2
z(9J#?dOuOKh^cd7l<G_R)Xx9E&s#6YonC!o1};7JN{>8sfg*1!ehI|RIv+0j#t>lF
z`um9!4u5r@6<eBWN^GsTD>TAIpDN+iwy6}2cRCOLVKW!aSAYM6c;)@*&q@^~r#D(y
z=;i5}wh(F$Gqc-qKP&||y~&m1)c9xaw-&7>2%kU2eJ(~L!%tL6A5<V}l%MZgI;SRR
zZB+g*q_v*8g=|0gW07|S(GMsDld&9U3!jS%<--p+3Isx)ve!vw&pe?YkqHg!*9n$8
z4ww7|yet8>3;{#X@$fk~g7oqq3ehh^V`TR(+nJvT;7Q_N_q&H|F6XblrH<L03OYx#
z<x<jkkWFW+xqrgYhi`m5osSE<yk5c;;RaiHew(wkdgf><leap8+g;~aBsf;@Dx!|O
zKK0!LlJ-*`91GX<RW(}GFMTG(>jvKF;-g9exeUqKh#-pEf<DKJD2D_)BuiJ5#78Tn
zqoV6HZ*VFj7N-PVXGmqL@qa|cz|Mz(x2=-0gH<W+@;%g=?B#r%pRoeaORGL5SB5Uz
zRvCOo!F;-x@bY{H1nYjy2)tPmj7ef7VNCUm_WSUGlV&;1{(p8M<c=j8RRwdE?YEtX
z4QW6RVf%m(JO|%)K0RR#aU(tChqo)Im^oThC@=ca1j({r(|Sei>gwCq*eUQ4?W7+K
zO4x_^BP-0)IK$>}A5B~on{&-n&uG8W1V-`3i2938*L0>9d8Q3tm9jX^hsw$%vM~K`
z9$W*vZC;SR(*MXtb`?D`W8@E^I5+ZK_p_DjL}?00UeD{4O6rzuemD@}t#2@OeCg}6
zGR@OG7n%l2&fwWKlBYULw>0lRF(I~XVuW(VMd-K~GhNA9s_9)-vWB{(bf;INYw15%
zemN_L>`f%9B2-Z!;@?k&$6!z)-#)0DXGxASCNpYhgnW7!w@3sW&PW*Nhlesz_;dLH
zVEJ_~9E;kyC{pyFU+f88&`ZucUi8UG%qX_zZKG%}#>@5jb&%c?2Q)U6%i3v^`pfGO
z7KNDfyh59ry~9PCm*DD^$Go#*HZx;8<kJ*9#M;VXP{?XCe7kg`u4keO<xd6fy#4#D
zj+@GDVoZ6ca`SqtU2ZAkW^Y?{7y~vo16vV=t}&=uLzA=A;_ScAnHoME8wka$kjnT#
z_fx<1cFD(7cwx77DRpx)F12M2n}>HSQw~)5RswfmSsU~Xb2%l!M$7n*w^sd{&c1W3
zQTs737VBQGNofqN@)EOy#Vi{pxoxlK1pym;xo}HuP9MwhrmO(qvoc`K#B9i~(Euq2
zSrneFocoero!ahB2c)Ny&8$5n9RR$EifgknP&5S(<$bDpJQ*A_`VH$33jmy*_)Hc%
z{rfI2%en!OO`==A$DP<$>n~i8rNyY5#lU<A8p+gn?WV2JfQL*xqRAPI1G%s+B&LkN
zKF^lHQ)T?5Zke@zSDR!%mK{{iQg=NcAdRDm9tCni(WIKw)6xXZy_A{7G6^^J2Ue7$
zx-?OPC61Z+^YI}w@Oa`?BJ%XvrMdfWJ+Cx)2QpWt(tjb^j%$Y!q1Ve`AlNGW2-7s)
zJyoMrfMY)D{&I<qq~H#^{Gvzh3yDV_Bx;FqJa(DGV9{%T4^I-KHQkRa$67o4E6YgB
zLPcUN+VFE4Fn~88+A+Yv>ige^c$I6+$baO}nxneAc^<v|sBkMK%$^WBzGpOB9wVE0
zsk^?9^kEj|-8jV13kWwoN`}w+3&xWmCx7?&D2S9&?OGFvhSR0s&X-&)3h7v3gc%$L
zJw+l%0do1P&?_uv*U!o$9Zu?lz6@WOf)TOEpS8u<iWCcux$IS5C^v?U0KaPP>04Ft
zXBL-hlff#1Z8Z@)>#}<4Y9mFN79BNebQ8#!IUgsJWLN9rTZW}pH%1)dIPX8as=TVA
zI&McS{o!btezch(g5mM(b;yrlQN7=-QiLRi@aC)@v1%!)a*aj19d_^JS`<-AQ)PVA
zKD)2m+MTj5njfRVF*t&ItXnUi32&>D;;ay&D5g4({p)YTyXGN%E4-Xu;v1G;iLd9M
z;tfMAc9!kWQq5B~Ry%-2v-q)?1yY$1SiUC)_spRedV%|C9o1Vp@0=QN+^nx7c}V$k
zuE%b)O7(juG8AAL*joNgV&w;&WD6}p2VD56K-tUQs}ngY56R?=!73=UkLt?mj1z;)
zUwncc?B`R9*^9N^%HsJ_O-L))gu7AdifWB{v#XOP$Xo@p5bq{oHV@M#{VF~$#I>Bi
zlu&O7oZsu%ZHtL*#P<GM=r?18W3{%fME=|uyivLpQMgAJX=rP34z=-7?)2IuizS#u
zr~@;)>2*|!fCi58Q}auiwTlki6Cb>xsx4`Oe~v{9s<6KUKRi=gmtC_+X^@6lmReT*
z<`S7_Bm_#sIYM!PSTFzh!<xIPAAJg<#1lxgd$z1O8S^e?$xq6L^57^wweuBrOUO-b
z_NN|dN=46~a=v@W4)hnu+hvb<#SI+Wh#mAoR;mJASmLRbYo)HFg{~nO#UJJ;W!dq&
zDJ%<;k79(W>*)A6Ia<9;AEMPyjy1j_Tq|!VlYZl)e;tAE5jY6$X3Jl~RHlyhfct}C
z_R}Z(h%x~Ee5{<+&b;xkq;b=QQ+TOx<F_9Nj9`lrM9i7Wagxvw*lB-exxMQhlg;E-
z93bU4JQ<2TGZ~BZ7li4^!J99Wju(JSrYWpOBL1~JA}qQ^e}A!_x$AvMpULN#a`vQO
zl)>{mbk6DDio{C&-Fd9HYYxC0-aNHr^U5;$09q0q?t7&U?!AFgx<UfgO1$|5>Zaz_
zr}hT((?5#L3nCS|iFFjYIVjmB$2tZr;KZ?nA6La;N^HpdIZ&c02)u6lSV)-Z1s>3@
z49OwuKc6??VRx1EePa)g-o`33!A})cV(@hQ%x>ye)Qdi)0yJS#^0g4pv5Kp{GUbzk
zg-}K>)tiL1W<L^MwY@k(l|3fK-FK}$Ci&$Y_j~|!akCY`YjKje@{i4O-*~~vRCawu
zlF^^R)>?wT&o`;OZ#!~Phy^*n4Z7BP!P}ony04``L?2?Bx7!2@zpALgop;y^`Xo_$
zC30(au^Qa+xX4kH`KutN&el{mm9{+)@!ek_x#ED6D$&5li+DU|9*jYIs%oISwtOF)
zZ&itOsXoQJPsUZu@+jxJ_eEM{+kYY@XYV$OiGvHVhdP;_I^b1fD$%^~o>cPHi6HCS
zc`fn*jevGavW!fB@ST87NYt9ACS16nipnID3O7>J3#8~V0W^{_T;8iDtE)7b09{eO
zKGv+9f6M*Q&Tb+JHUkFUNlv+_^UD=k<{j@@E_G*)C$gh>PvA=8vncA1@9{aFXiVgC
zy>`?e%H#P;ml)>ZEz1!^yjnIUXRxO^-IQVw{gunNOHFtZ#%YftEXDbVdl_u-K9h0B
z1v%&|IDipPeS#UoqJpxHs}LdlT*8zADhm#Ty1Gvc618Cb#a}sfMn}}r!Ef1x+S{`#
zeiY<rQEdN%zbt)7thlgp&U#qgimENltb<~wfgFYX2B^^_9gPNa0dQ3o>5QoBm9pV5
zxAGhzfx?+m^Wfus4+6A{buku<xCX<*gnFeDq7cF5pOtjsG_V!J0=q@?j>TehIn}OY
z(Qoaa@j`WJX7SFQMcnY}Rx7x<%0Q;X?zClm&iz(p<^w8A!#Y)D`8{_WRau2L-{lM<
zQX%QKX>-rB9J)OJb3%i%q}Bs5GPac?Z{Xuo+{xbiGD6g*t@}^GwA&7ljbcE__HR*u
zVL$S{9Z|>cjQ;DYL|chek|g2-3Q?!7OSlKxk(s-b@xklC*T;BWYDQ5p;<TqF)Bt9e
zYg5oL9`%Zr*ndyGj%#&CjBu42NeKNao_qO~xvX{WbPk)Sud3hpaZa)kHCXsE##UYj
zl+*VLf+Kq~ciXl|an_p{p`XqJyHFJcgu$OEoKj5Q+SG|?X(|}B#2zwb{hP~Tv!Erd
zlyxkfid#_=#6G-LYW<-#Q4miIYNQTgtqJ^;HcgKILdMQuXns+|ZTCSzIb}^b-sld3
zP#%w|sD$ih2vPoY6<Z1)OE#T48WX%j^eAVpNL~+&?u$SF*q*Ei`@Ifbrxam}9AEx)
zvjnRCE%y{8YQwBHktNPs=&w;usy>{h#cR$UIp`bEToil)M2`(N0gz+@vhqU1IKERm
zvgr<@Gv6$}X?!s>iJCQqBFD+8&R1S7i~0+6tZW~hq(11Mbn=j<(l#AY6c@7>(jQl%
zkJxRuJq)w`NYpFUmD1{HwA(*3=rL23_4V#FK>^j7r5?;(RVIA@6O^GKb?Op1G>tXY
z0(S~};CF*um2H1gV#qIFe1A0?DTIlQ9FkHfd011~bfxN^Cwe(FxSh6>xRr(4cPRwV
z_7`jPqA_w>JAQh^sg3B+sK{Q3)q33hSj%A=xo6sWl(l1JQ?XaBvM|S{B-6B8hGoED
z8F9@Fsr*hT*6ZHar-=RMM*R2KDn+fm)nW4TcJ<Spes<$W(uE#fr_M|gwx`GIk`U<o
zIu6*ct-3=I<9;=Na67?f$h#$R<MF>s!}nRoeFb-<vM(4U4tMs15W38q|H*k`fqP&+
zyt?rj?8&$!gk#gQriNPy4v4NU8F%`vDe0fIs@#cB6j0RZdW}E_-32el$=m`($L^4!
z=X7Lg89ZQx8;bT=l_f0Hu81dhj~Z}z`!*^Eg2#Va8uw34j}y&*@tEwLA@cRi#X_NF
z8l&v4>SQJ+4nc?8icjP3T1nCzq%nkho`;T_K{47=n~kV8b47(!#YEshYujpRcY$Q4
zZ+?Vn<_q#=TY7l&eS=C{VW#6?NKqMPtVT_t5oPOuX5+G`I}I<tVi)CdTrG>Za6-=I
zKX?|)6i~>}Dq&J&+e8UyI_%{SIiDpNq2k1HdsP!YTNjJys%M5}x83lah6N-nyqxe%
z_-7^t9a_`9_3ne%UjObG`opmzavIL%pDwnbGP@jPQf7xGFLy%M#GT9xR$CaxwD%;I
zu?8*AS1iXq>OON9?f>3dc~njj?~^M|>(%8OzvU^Nl2wjxu-<p8wT{A&o5k}~#jf>6
zLcm7(WD_ac95cQ;Bn)N_h)KH03OgHEyvXug5Huj`BiIxKYI-E#+pos&jA{CIil)TO
zuJ$#p#lfTWOYtZ&aAttw{Akld%J!!o-=>+cT+cmT53Y#<2!`Zy4&c~&wSx=<EDo7?
zE|=`uMxGdW?xG*dV(hP~H-x@o_q&^~YE75O^kAywpqOe{Lq9!nw7n8e!YuB`F$mNr
zS6j@k?Y4m(KAr30P<mbzqY~_BjHPL%Wod4ow4^1yYH86$Bd_A<gFqO_-!EZvU2cUY
z<@g>Qp72PB)O7jwlJCxx_v|)suorCj@d*z7utl>D490i5Uwxtzo_g*2Kl)yNf;Y|n
ziUIu=AM_ng(c%m~>aBl!YgwY*rvWuB2erFN3oLg1Cn7CZ7?eBpN=z56iI-ooEz&Nr
zG{{cAq!DEQW1R2`EqbvOKAM&Dn}=et9|JmiB|jX=SQ;7olMtMUu4XOT2QNlNTw&*1
zzS?7a<)F-t$=>Cnwz$yH*LKH+hqKiVDT)#GJiroA2x#tk%eC`QPu4&`LSpcwaDYI>
zRUT&*1w*N~;yYU!$L%RqR!Kz30O>L9F+EKR^Xg3Q%mUQt&~34~m4?}G*=>nR#cE8|
z3ssr^Mr7EdUPg(BMAg+)9gqP5WKqH!1d`Dnl&!o5ecs?h1-wqcmhCSc25^&xT`*rL
z$dGu)edhPAJKBe7V!nP<J!pSe@TSH@4mXn~Imrv1KU2E`hG0z+zdS<K^`U}MDzf6x
z5&u5s5C;kD{Zw12GqnS%hsU(n@<8LEzD*SLAZF@d|KIVEF8OaBj0cl2?uM0hk9fGi
z_w>MDZ?X$j*QCjJNDvTCk<d6ei*SeT>BKIeni{P7@~SO1ErDx6+x;wNyG6e#vI$EQ
zJCVgoBm#vAy`^kLnxgIaRO{gOAj@rYlQ3aiX(=`Q?`n+-ul1C(0f%#66Y-zPjAREh
zRBjRGc41lMl+B<R2HxXCIwkuBm%fiAQm`cd%o}>N3#ssvHe4==cz|ctm%~0Qxt-=A
zrrv=r7Fi<)oe!Sht<2Anwy4!pAX}fvrC(`G>k!iaH3YxBE{DDp^vT}r&An){>vH@5
zkrZ3lb+mj#%46w;6>+?Df{R+reOpU<>6a^LAa96=0Re7yW4P5eP_~ud>`z%tI$2f9
z$iieM1LbPLmxJd=#Cm|cH<sX@fQD>sm=(_pvj9G;^@LFVQLmfd$1NA{o9H&DHPv(M
z{B+=sXIl4TZCS@ho?U6cC|Uv&`-(#arPgQviL2M_{e~J%*=OkGMyEt=I3ItUZCG5<
zZr*vMSYK*bmsGtaa6RW5in9=2z17fn2we4&|GVmA{)0UFb&5i1DqQ#TjST;*z7?FA
zKLqyzSN%79oubR<TF`3&29<&)0$Uq**s|OZHa6UB*T0~#@1~=x8g;VefV89`yIMWy
z(2V;}Cixh3KiX@zAbkI<5Q;#oQ|34QT>KwN%-a9C9t?^8)9Z-IpSCPoF&;dGuh3JO
zk5Yfr15S3QB%k96{)bp6#8h@4Jki?cqxdHFX33HtLhYOhE*GkBJPEp<^B4zzjf??x
zI6!cw%pTUBg**p7e&!yiq{MDg66M1dO^XH@#77yT3MDJijgNL~I5`L6kQJK$&Nr|(
z$xeiJR3NoEvcHCp4ai>Vg|HRannE+P6+X&UC`nFSiCBxBDEBqR%a}o~PE2(Q7yR*k
z)@HS;ZAK0C%&@7&_>dj$2u*%-b5bRMQ-G88yB{3L>d9bOjlIH*atn^C1#?Br0cC{V
z&8CR!E{QYoJGRAJl(1Fn!Ebi1I7($3OY4o<Zu+=ZeO1HK%#&Wck)o&5p@*n9yWGc!
zq^VpoDH=JjJiE~lNttqGxXl%S@v-mS3t23#`z8Q#WdMw42MnD-+XGQ@&N%dWwtQ|f
zQC>mDK;l-{&4DDbkY9m}YO$Q*!~J;L6m=Z2TD2pavb{R5!v*Ro1}K%yv;e?dRZG$P
zO<T7>?;3NMh~$~s;z{u=BX&RtIT9<_F=kdZ8FH)T++{r}_ymOB*e=(k@e8;z5dgL&
z%$29)%F>YQ|J19lUzh6Kwm;(&HG<*r6OZV%Kzv~hpQ(O0&|0Wvjv3X>#@mJM<dB%M
zEv~M>Gkj4!?l3f{(Br#gKGEY5sN{bMvKsFsDG%&;O;z}fK>*efTasPacKLYCapRE_
z(AD6>@FlAfoaHIok@O8f$|-Cei6lwN-aQX5$5~!y)RP)}W$zxfdB8t2pH!W6G4YKE
zU}>t57-s)deas;22@KIEAs8Z$afz5gZm5~ok^fFp|6e<f3M#t5eT=V}`hkH%Laj7z
zn-pk-pq<D8H1(C1s>7qsJ+Cb}-^0qU#x$#r*yVb4#$gBzfHKu2CbE}suWcY!^_rvI
zehpAd&gvRwRqsFHlOmi8VamwZh8MGHG%Rz!uhp9?af~=p79e3VFDEeKPgSjAY)LX(
z27V@d=)D@XZ2+qGr7T)O5153=B*(m3D4cN2qPMhS-7nzR+U~997QytsBdV=cw&t1D
zcP{qnMKbFFgdTR3j1f@`V!W)7(;h=|`5h7kmecPQL{fix(2xnf_Exz{Zsp84^xUI7
zQAa!ak#H;5Bo>`_G4ZRg7qb`-F6#u-@9*r00ii4TbmmG+>Pq4`e!w{P!qK#dPV1GF
zaBLcRx;K4=cYn^Hd$pXu+gU(=LHw%X>6=vb1{`c={=iHS3t1`9AG3QbN#k+YeJNDT
z4EKAk!mZr6VbbezR86ve>~r6}KVYf+TNBo3udk#|t6o4}$_hvNCim{o!rT_q0v+EG
zUUhhc=y&)j0G3F2mKHY46_=w~dsJW5;-ppP6u{Wl;d3jyja0`{^pruZyo4h>KgsiH
z^(cy!ToBNYzCu7GjPsRa`Vn!vq|h#w$mxNOgjK0lxXgRmi6GE`y;Ounr>O!&RuZgx
zZ9`|dG+SrIlGQJ;A4E2=?H}ySXnU4#>-XC2^R<Y5@;)n~ZhyZ$73M7!_!afHhJ%ox
z+o^TsnDdiB+u8_T#e7#*ww@G1*Xkm@um9=1b*93rlj{3D0p)c#aDP_ydshCJ-O|qA
ze#|GF=@G33=Xp#9FU#p>(p|TH&mcy7e1YtI5=6|?>J8O3GfY}ubx*onKlTx_x_mJz
z^K-s@#^(x+zMidrimLNj^lV22y6MDN88ObNThiP_CkInZ0F7RdnM+KI{AP#)2O1b+
z7HpaN-4sfaqQ-tYJ`ctBZ(WqcB-~-QG6Sh?yXdHnk_aCT%!&4EAc7+3<!;MF0hA}Y
zZp-H`C)Hf3`zmZImz)RQ;4Fd=3z0MQ0US9r)=6V3#$6V1qUh0qqTL;SJ6$R{)thz3
z*B?g`*<cfC^nT&7+bRN<oBV0AkQqAC9~<?LLcCP<&!khZJ&$LyhOHm38$>#IcQe~a
z7{0Y;q_I~p&0Z1{>_+49#o<LXcO^3=UAkH?CP2P57zKWCd);bZ)XGMJc^VlImfkB9
z23YpPnbe)K&v_=pzec(Mzs_zVY5ClOl=oWOb3t;9Jwe0|wnarB?%lfY@y!4EWDUY{
zwIZbwdT8pQJdyQ&l#(AiL+-~Ss0AI{0B9tK5VHNuGPICx{*R9<g`Ie3=f(5%K5oLp
zcQ?}C7}x6Aft%$Glx{cHKil2;alSB}kKL}O`a~gI=}{-+ICQU6rwVVZnoQ9mL3;Fu
zMQXS@2`Iz*+Xv+k;N7WP&*vYh)2WUh&0B2R`%<gZGQs!$4!*Xyyy}9(yIS<A#$aC<
zO=iG=M1?*;1qaqZd?-`yC~K#bj{N~7=SWot5qbsFCNR(iKQ>P$u#okfOxzy#$^-y|
zI$9Ng^uL?kVb!O{WI<9IAQLUedeO)7dzYZuKLQnDXsg8%MSYsvQ*YYCFmy__0K+PH
z^gqRgi2E}vjyi^0zxHOnP4&+X_U`Wd9E$j$2-%M%<(;~U2mz?DF5{*802l8Tz+&q^
ze3lp>7+V|~;#fRnn<QS`3QiI*?6e9q+pUV&7u^MCm#zZ~5tB7(kWW45O*|#llQv^=
zrOcP7-i&fo3R`hvr+4<5Z>yWrE%IQdV%%~o0~>JrRa|0ESnDNDz+;FolS20%LITq(
zt5)na4{)H*VBKsB9}$yL>u=j`pA(;uKg&1yFn%3NVWBqn|MEM1zFe*7sJBXRY^{VQ
z7j*^yS1mV%(iP@#LQcEi!$lmRkLH3vCmH)da|0MTD5M!W-=eR8NMkN|`CA@Or<bSN
z<s2LCAT<+8hoc*9ODk9^^fV4j^0#fk6U2f+Ig6EDt}$;~gsV970b-+AWEYn&<Xg5)
z>6t@JZzCq6>PNGYIco}ng|c(P7B|69K1^dUhbJ$6(>w$Zl5T*=i<x2H&nN2S`q#fn
zIBxX!e(1j<e8Lv&l4J1yN8zRZ=Ei_@{a=Bt90;Q5dYNv+wM>{eYi$b}EV^M-B(+<R
zU}s&8VWZkZ%0Pg{Zr~I=gNXkIxONJ%das`qwy2$!7XG^2_256zNAm1g5O8At_QSgC
z^P3)sm^_SIDxl8fL7N{*O~Bh-w6yJuer(V8s~6>V_xbV+gNbhXH(6F;ua2_O!i8Gs
zr5kF+X*nBm{iAB!&84Not$c#)!>1IwBLN?}SfZ{cX)f5r*iJb5vII7?jM4Q@=VVD_
zeEK(%dd6?R3C8uBQ}}~xHa3KQC=#tp1tVa`TG*~N_A$;?>E9yJpeQ^8$ktaN2Ss!n
zb?6sOd^;1J{6mfbNy-*3r_(3EO22NC{qOdS;+pPl{J|L+iGWwMWxfqObfz$fd@>yo
zh~l4EzBtu!%nK_hEl-*O5x7a8JnV{k@Fl|=XEaur`}a6{Ze}Yf^wCiRx}x*ky%t}S
z{$+8)e)i1PU#3#=kf1qfZis3!f^{8H@Ioh9UZ>^1vsolV@3VW!_1@_+yDA-(>T+<H
zywBPWzHmgeP~J6);XMbDD|KcXtvTYc@W|(dS0gge6sYO&VJ#U`$|g?}{rcmi&cjbf
zsZPAUA3N_9W~i`<&OK(<miK1Mh3&0(zKyI}v!-8GFk+k-tJYPP(d4!K+TjbA{@2kx
z!Fm9PeB|HVPsJlV?v1n5t1*A=mRkR5>wd5~E^KMsp;i<$*QN8Zv|u~>nu_)E_V<M~
z^UfyJ3WzLr8<C<+qt}okLqo@@m#3`qe&MCnwS{c<+t-1?Nk16)*~KC94PTOkp`w!x
zCD=bt>C|?g%1C#c9ORKz`xatrJq^!vaa^-O4{1*?-gNS8<+}Tgx}D|bfe<U}Rxurh
zIxclT8BJ$r2JUi^*%s@E$$}q=g?x*=t{Hz##y8F2zjPnr!ZZC8ict741!(!%W#mJi
zE%^c6T!qnejH2Gc!wh(I(v+8zA;hjj466fUmOKubO~m%KRo1*S3o<r!?3v4T11gOc
zT5|(Skp|i$^)UL)+lu+^O9#e!oH_f1q;XfRd<oSfjmJoV7)h&x%52+)u+DEEm=5e{
z%k%X%rr>=~Ct##ZFDA;tz_{+rGgAE=3L%fIeZcxxhnh3-pepRqH-knaEZe2}RK{=Z
z;%nZw*bx#g0n5Z3o}PsTfKVP$Rfx`F{;0v_F&w4gbSiZ~DNB+$-neS=$P1NG?>9FR
z?nU}^7h@*xXR~5SAV8@v)Q=r-82k4NJb91vrd|CbHX<X1+8eTal(J3<&;K4|cN)nH
zrMVz$`e#HSWBJv7PrZeQw=$Qj$3QkzeUV-J9|;g|+7ZjGj`f-rLZFkOZv>!cVK~q0
zU)vHFsawqH(`X;UqfQq<k3-|8rpPI6jk>Oh(P65NLfDsi8)arW`9{t~m%VZ~k{J&_
z;X2cf>Hfz*iFUyNV1lIBA2ad<7ky~$dq#^hJ;5*UB?jjto2jn{AhdRU`iJL>8O<yN
zn(mi`Hb{Ksej0g4WTA4Ib|FwHU(_g(UUX;#wqA}x3pEc;2Y-4J1MdgCP?E-W#~rxN
zon2>0e<ev;0)Xc*Mp9!Ko-A8dyzG^WkytOkIqaK*9yTUE&XUfor~Nwb#7WX#fyeDn
z{Z}2#>P~5;`h{bh_c=BME*MD;?#j^lJP6$Bs{B?cB@%=b3>NJye~qBPTsuiTKe9a7
z5eWo=52P-qhA9cgk<nI;5Db$c^?l6&N#W931{w6qlBXi2;H#XOy!6W-*!Hb6O7@pK
zJ&*`u*JP=h*qg4Sx&=3R5pnpDO_~8JA8x+eYGJujK*0wHowFqwM|Ke?7UA~hYKv69
zKmO!Fz)sxAi`edI`+E<gj=6FdkLNL3CG70Bms-4g%`I(}S0#9_W_5#76(>`aj8o*;
z`1CL<$^sTW>`k*}A($K!Ln$$>Rh!c}K?p=g*nmL7`IeWt2w7rBG|H~0jQouGMmj`}
zTB4qhH;-E6vQqyh^V$QgbZwuGnlU9xMtFCX{<5TFlqh7Nk2iI%Y3cVd{R49))K)8x
zs296>7Ubjgj61{1kU}63lQMNfgF%xG7{p%&6tt0uGIDs>K*f~fx>NTIe@DEPl$72h
zME1txb<;Xn=FL;JZw<CYVc<bYD#P3T2ZP{oi|$RsE0jIJ;E?<7IIYk+iCtD9h*N5E
zOJnrMx${f6Q$fk?(dTGWEXSn^N@HrF=Sb|r37BXwTI*hMRIbQnmeUZjcWNm$*D9y5
zrn=e;LUoIXC{Pj({8MD2;|MuCT`zefWH)wTG(IWqfUK@#LG<E!#A165?{Kz4^^tZ_
zh!1wmXCC6G!U=gVO5oL4ISFTZt$)YvA7f9^ae}exps;2PcY|rn%XS-;ukK}EgnhVQ
zDC;pVjN<hy{^`nkq*?Z8%KhE(0f2~iff(W91}rF!*t?nSIi%J^-{Ovu32)ZuJu;-(
z3^-%A_-alX!Ms@58!p~|nkpDDm$S_OL%cS<Y)lV(SHpz|*vb_pVR^w73R8Cq4;cuA
z?y(|Qf>Ey32Q}L#KO&UFdADJbE0MZG6k?i5^r1x@B1F9D*><2DreHy2A@bI&UN6UB
z^|4F9Jf4vnZb7rIBb{EUD3jtt=%KUX49PtqYBz*(CmKhOjU&o+y4u4ZGsT`iKd4XG
zRxvnRK4gX`i+0e1DH6Q2UAgpoPey0rV+tF&jn)xz=g=Lq1!cY=VCx{9_KO|XduX`=
z%(m6OEN#pI6AmPySnJz%2j;iZ9j1qlB;tpkfs#JHxrt}@SApCs*^*W*V%N@t^B-K}
zVr#wy{n?Dv6Jyo>EKME#{SyX0Hjk=TP(m{WQ7?%_sHnA{sBdNxKlBS5ieLf|J&fW-
zN(6@+AA~UZ|M2ydQBihX*b)*VA|N1=64D^8G|~+sAT81@ox^}A9U>(l-Q79Bprmv+
z4AMRHzzhT5?eo6h`o15J?^^c{*1+70v*S8@?`xm4SMf0PTHo^aA&Q4@Y6mS=B^{!$
zL6@AuuM#*@;=2>29UNq_rH`mHo{KUdEgjQ72tZth*#8_v@pMKdXDmaN2FQov&4cgw
zja&@z9ZXLVxNpu%KasGY7jN=O3?bar=)@t)N|V@<N#)efc?`PeOZ+|q7k`9qx!?yH
z6|#P%JDZO{Kp2NE7oopp(&zWEFCL}>Wz-MCK(T{wB%i-`XEcMx{kDf}cf7c*tk)=0
z=1C3jop$aFNyEA=8C*~2T#hYu5~OD9TVg7E_wx6y!QFD|Men$NM2da7xOYH870d!}
zEAY2nRS`W~nlYw$R_Ffn!}r}C4xu}=VIJdOfC9KRjjtkB*QWWCYF%^EM`>jqTPE~k
z4pDK{g4*HyNiZYoX{pO+lu#7+58A~(8n3)fgS*a4^EMS_92KBVWMo~@&l##$=Ea|8
zzw40>PF=k{QvFEpptptjh9&k;=Dc~<!odBS*M~WbIT35p!}S|*jK+vAsHo=fht>j^
z=5zqM_|rL3uxo_>6y{BQ^-tK4f&~tPw?i=CHo)bs%5M^&@`)ZlGOkE3)vbpGw8u3m
zF{nWwD-7uw!@`t+8J)f=-%dYW<r@ENwWvIw>Gqmb^O=9kTj=lr)z`F~MipCJSHwq=
zGha)09RuVq*F8^rxj0p`ITCdG-oozb9~v{BkTb`+3JUXZUQM^L&4+SsQdGCsz4iRu
z18-5`QR}lx5aMbB@Y?WoQCt-%^!2{QKJ@C<A(IU2-E2>D;@gup^)E|6=L`E087Fn^
zt;Xz|8`w;`jmo?f?d+FW?oTd;BJ1&@`468@efP*9Cm5QrdwF(n@w3>Z`d6WvZd;)W
zMrqxh+WasUY{G0F_&`@k_gB)`B8SaSzUTBejbP_~wXZ;blp@v4wyYD6Mvz9<Y?QY~
zF=<bB`s(i^RXko2pKq`$X?;~8O}-J$>5i-1!r338^h^wRvs5Bs(P05LQ_rxsFZty|
zjz`NHL&ZMD>vuncHv7{152r9fesSJf!^Ku|*;^$1>A$Ee1&;c`)ifjUJ_Z)yMMMx*
zbFg0=mgc-yxO=6kWh#1)y|suQPdSc3vGQrc=qo(tr&5Zq?!jKxF0LP1*<T!P9v<Ev
z89JqjdTqXpZ-!V~4ZG-Cg$g(OWD7y4zG_s3*Xi3rtRCvCGnd+19iO0UNutsS5pE5%
zh2*XC#o{R9TM)|S$zW-$L|gV}S28!bG;gw8bBxOVG2PuJZK7X(484disKVvZ*GqhX
zsDVfp$E@Z>qDYu)TVL$1-tBCmq)r_VhEuZG`b`8;^D&7@vpYub^%pwG32_{m>Z%GX
z@mn?AO=nYc%-+>k2s&Y2_Ceq`k2Dy#<6(y7k^DTT{}m=9$D*vs+N|I_J2S|4Kll#b
zw=}7uGrwgU3^;J~xF6{kAq?j)P8_Vt=%{#M73DMl9r{VFC(u>#Ng2ALh9X{zdz@h9
zETN80?Jh>|wEi?@ieynt)p{jY$+x52`kYf$BzslVQDfkByoLnb+*h1aY#nQoFmzij
z@<j_dyW<Zn^wLYZBFCqX_95>zlkwwy!B^4VuBl;7izo8y&Ye@v#<o;^#N4S+VMVZ{
zi|svhe?!(4!b5Yb)hD2O>YG~}-Asl5HeyBJx856n++XOyukq*lo;*PZb;_!=W8VW7
zKfD=QP41{Ey-2S~bv$%3WBwq;oI!Jn6Ki7gb*QIg>3d;$m@AWF4r4K!N`dlIRFfM;
za2vKla}t(lj5XFLlP>X5Lp`BvX6i?_4{)nMvFKgN;hQ5`xL$09*=@&?&ll+@5d!p-
z&9<6T<2GkLvxkvjmN@^Rr)yW(WG~(ZH$6X*O{nf-;!hA~^Ln0FZo-w3!w{+GjV+U|
ziQ9|K2`XIMbCokms%cS5@;C0PmFoJOg#Dhlxz$6k#E|*p^3~p*gAIMtv=@&X!wT1k
zQT(nxb&XPl^zF}u^*5g09{b6c`?8C;1QdcJOQ^#J0-&$M_1Jyb$8q~hg40@_716Jd
zyig4&^Y6yWSuGt-c|4Ce4vu<GE>lS!{fkdVKd3PIc>qhd!h*9X(jAGe2JMz$(BiY+
zp0u{PV-Q{R1tvil)cd1GOCcjNA`kK8Ft-t{Ga}2@F_cN?h)2%L)-{-ZgRx3>=eer7
zRgDqjE<@#&*z>6S=#O&X<+@DeII?Zxo8?;}@Jw=GKVP!~PLu2|#Eh{jq2mOqb%5(z
zf0UqqQS;M#lg4o>;*ASF+c5#m)W0KN8Wx3aYcFqt#5SyOv-R!_&XRNbCen6RfBFd0
zyJ>ekC_ORM?V5nq#wd}zM3#$|fdoCDn#P9IRX1rQ|5_zs(%WZhbQ^r&C3p5qGNk;g
zA>~<O+Fgp_pB8V{?6^no3(Z6Oa|zxj`0kD)*|m?wr$enjMx2>Z5Rg2Y?%^)4c!}f*
zhR>+tq|Dvz$F03gnU=k#(TS*HuW4aoq#+6GQzXuu5v~T?r+|)`QCBm%cOQND+%=Lu
zSXSPN^Wqp$8Kp<(q!~>0ew2dZ2>~2lByYSvV5*TZCm9#KsJxG%mqU55LEs!KpEP-*
zs$VcOj%C{ET9}%?Uo1Cz()Z;S?6>-$<Pnt$aCIyf8G*~5`KRw99uYVCZSs~=Spni?
zt$N`7ft7k-2H)rfjneKM%`5_nee`*j<=u|xf$79>e=M{`=jY1|`4HqdD-XFNTjBkv
z2qLu|MyA6)5H;tXKCP;|US0d?C#RjHb~yD9w7I<~H<p7QV>3!jVJDB)(OzysR3@#2
zN$<D>8()Ioeu6g{2PA`%2xuWL<<HWKEU-U%cY=|&5k_KNkeDuHxgcbbKmXvtb66xS
zc*#3I#Gl)dRw3BCd+8(14lU)`1}ccyx1NhJM*0qR1b!Bb!%AyjymdtREC0YhRs=ou
z`kpXYOa1VMiO|{{Hg}4(wC0na+ZE%NvQ*!~5p*FT1ARAECVZga>ivxnnZDOua>2k&
z-PPc-l)T{mKF^*isNIorDj;j-)OX{T8lTT=Hy*b5vUNyF7mgdV)jr%fac<e8?ylb(
zF;VNDhIeD7R%Z34nVMB_Wmx(!)VN!Mp~0N^ii-n`sM%ela@lG;lmT(Zzo_3bTFA!v
zwLW0HZ?#NH44NOl<g}&nV(k5G6sVz_H*N?93Ll;!^_3NeUjt)KK0DI#(2QCw*$(VE
z){=Rza+r{&Y-C9~bx3T-ejm<74C7`by7N!dX$8d8`PG$6?Th;4dcJ)xx^;LlIY?;U
z;<?5faOfe{X|hW%?YHr{G(2iu>}tR%hDbBI$9P#WJt>alIEW=$)`DF@HHpWE$i#sc
zC$f~9lK!#n+TknahAWLH3D}8?J#k+IuR@o|!|P5)vnrS12fESwb5YnnX@zresLz@4
zWNGhb_Aie5NgfL0+)w<cKG;oGJ&mUI;5U{y;wxRd6+=BzOS@rIyV)k4Z{dn7(JPhj
z3^h@zY`?|~htPR3OG;*nc<w3jv>kC8gAYD<EUZG4$1a6o*?H-Zgh~*87M0@r22YNS
zxOO)Yg!!tBntbO5L2h3mwW5*kX;`Jpj=4oa-vo5N@SqGn`QFRjI(LVkdPPpw>Q;3n
z<@9_XyE}2^gR`rolZJ)uk<kU+et6%Vf`g9SvM(gi+lsCzP-ZFLz`5pV>3tl>bsaap
zCe8!xhSHa#<)VjKOb>>KU=D@1OLOW&!xUDjBeP;@`nRrxn3(kM0x+{j;Xb){TSusG
zdpP#sY|E`925!J`t+u^Kg3)D*nDZUSA5&{0UoF5T<icOsuslNEb)MJoG{HbyADTRw
zuXrNvr7ttZ$3&m2=;kD2m8dyt54wp?=yJJl5mMj9PZ|f~j?v}a^B^k%!0W-`VdyKv
z56Pd~+=CX$KQ6VGsyQt73+bj<urUZt7fdlK8EtwqRs`%lD@gy{E}3$;1t;RuWOWTY
zn7Qk4FZo;OazvuE>5Zas^I{xsM7H5I2lzsR9TiBLOqlf1roypHN6k@a)gQ_oeX~J_
zrrGUEcw524zu0}rh=IXv4YP3O<n%0O9C<u^$Yw$L6Vl3gfE*UwCd!x+$-v^2=u)|G
zl(RaJT%Pfrl*xQ%vK4CrULL!NXiIUAfaaZV-ihT4*5p}bG>E$57p7r$(Hjd4#<*?S
z_~y#*lFY3hI9Vtvoa9YQ=zE1|ix?aAObQ~>e_tJj#?!C_&ovM#%14D%-R?P0SB(S&
zwBL7V<pXGM2ZIkrAfN9CkrIx{7B5ub+o@0HI#t06-q)@HV{}lhVC1>&(ERzX&;0TA
zUGR8$BTOy%qV1Ce%DaZI9s1%>utd8NAX_h_lPwp(D|gV<M?A>mE&YVP+G96YPG0JD
zdC0q{-~ab60N#}WmJ+*n!-EKGNc~wlx68x_+UxIw$I@1{19A9zhE~-B#;ZN5?}PdX
zq~&9Jk72ESwlOLUJ=$$A_C)h}l2S%jFH}v8PKfCop&#g&M;gddDA$YNQKEE30~fT`
zy(F5u6c@ONoqidCEIDc@tHv0gC$GS+zpNMjKDy^4+MNP-FY!i=C*%V3y?XcdjEJ>c
z;><2cC`X%+Ecmxa04{94JSoC5J}zAWJ$vvyjK9ru(_5teY9Qj0Iv^UWf4~pS*QiA7
zVz?FFhbhB&w)=}ejvV_{^c|!)Lg26r)`=QCc5bb?$ygmim61Y6)^o#E+b?%v0#D|c
zr9oR`!;o6SIs_&1ZlrFTt)v!NCu=wqS5h)`jo{=HrOS^iTCnhwsq$yxqO=-aMHg6Y
z21T%+Q0?``Ik#33<b#GS&Mh+GoJ>DI1DqqwawsB?%hq;`T!3V_F=+ZfdvP;11ms2E
z`>5>a;ivA6J&2xgHfX20e=W)i>QSJ#wH{`Ko=(K*JQusXo#gS`4qccZCB8X4;zpqP
zluqwmSgnMINjM%rbe)O3k}Pb99)U+!vDA%T+VB<We0QaAf!~izixF~A_5DV8M6z;`
zC^=g+)Z>Wn&cP7$WJ4wHaT=3%&UIC?7MGgSp+UR__6EpPhmnq$_gZomk|%9yH41ZU
z@0>bBSV3ZpWy2KF4wmF7T~bMfRpHy7%0>#~q`)MN3@@OQy$jyXnN4b{<}o|e6FMWH
z_XD(nWBZ&$RkPw4()Otp7z>0y%;#&6)AFq63!Qwu7C)m3lc(#?<*)Zyegy;Z10j{|
zU}x+Y!HAfgo{@y&6c-0aB#!Y*qak<b-o?T;gwATTaUvk=oJv@$@WH`O$X-SMg2K}4
zkDUn`61jK5H*&RBsvN-8{C$QtM;#dM2%6lfmrAFw%J%GBg;=dv0?62cb8E@{BRSzq
z8*2=zwKM@IS%S@BZr`7#K(7Ex=FJRO@XfpjmCf#!sisMWrH7K%v)SXYz<jtnD6)x_
zm=S_4AJ<Wzx;RlQv?RLr$oTR`<d88OmzQ+%UGZQB1sIiNGVyg@2gY{jbePc|+;d&0
zz4UMhqu#W+wb~fLx=O1mA^)qC+FCY*j)}Zu;Pj%{f%)5|HZ`lUtS*OAmMLXj9ZSeL
z=zH_-?Y?+du-j<q-D&y&=r4#r33NbmbIPYC!eMcxM=pfW+BcDhcEof#P_N-2TV+--
zJ*+Fi$LE90MnLZgHsU6^!Sa*IlHZfpK~gp(QqZl8<@9sEuUJ6130_Ow_$N|m;F4*t
z47V{Hi1WL}{ZF7-x!jQ1J1Yq1I7Pf3n9(b}-#xSm#u0L_9NT{VCnnG%n9PVS9i2e$
zaWF0rZDElSnL*<H5oGvvH_NX%Ms5@gKg8-~)!&UY$?yHZ^}_)cy{a@}y0jSKgcgd!
zQJu?scYo#AoQ7*A=c;lgdtC7C;#E0n%!6wdGIWyz`}rQ4(Z~C)zN7U)4Y!NI+0Nuf
z%p-aJrsSrD%}<ws@v$Zr4Ob%$M@f`OzEw>`Yh<V>ciGTDa)0rd^$=Lok~TcUO~da(
z?wiWk^_s-+v!)ErJ@8)1vombmZ)9u9O$DPT6Ot1Ys8y`7L^@fQ=%?~tF13uc+x1iU
zuW3^fZx71Lb-LDN_U*Mk;_PkAVG!z6y`wB{WZ_f3<$&Om#LW48_9S$<III<CL2~LX
zLztPq7TQtZteYE3n9?@dFNAm=EX%+^l5Kril3q8M2F-SYkTIW7-D*?h!eMK21vTF{
z6YvHzN#`YqNa7V5W&HZ*L#Q(>zySK=B(yToX9*cS+2Fz9@xp91^lY;BHoLT<>AMrH
zPaEGCM3coGN7xb1pPTF<`8K6D4<lpBygt>ljBl-)OhWWxa}4j@yJv&Qu!xARcXE@%
zckm1`1(E&Q8qI!F0xPQ@n(B}^vxbnt+z#!Wo+4-(#`8^>4bBxQ=6h6X^2}!z{SUV8
zG>;>Ow!TrPHtmX1ouE&>H)FBahAn0!_;O(rDU$QGVoMQlDIc?6U#FtQNYXnvh|YM}
zuYPixl2exp@s+{XKJMkt`WX;$iJz6k(F?LNR?~FNn5L!j_On(Rd*tis_q?pK6l+w9
z%<OnJl@o!}R4=wT*AM$f?CY_muZV@2t$YV<qBJzpUJ442uRsDp?Vs3U0M3@9m|~$h
zk4R~M+GiJ>;L;8IaXw>o@Q=y9s+I>3nZ&fpxmMk-pxqy6m{Rs4kguoPu76u;KG$2{
z^fdwAe6pb8ya3o}dz0=VJew=Qg6uv#z>>&iqW#g}h4ENLm&?M5yKqCJMe)>HYBFW>
z=B~5RxXt@Y<0fP|$MM?@#)fq226zNl*R*zI6128_dt$oXUQou=N$6*7gum8=ipK+i
zRYd{1@t6t%8*vAP9c8zG^PBko<jPINkwOf{XuNf&=hm339s|~cRsQPY_)jFfb0(7(
zF3nE|_Kd>06qF<-dDE%8nX8gB`PZu3&Mt<PXg`16UuO%<1EzVnF9~+eLa#O4FZT8C
z5fK$;THnOZR}r_n6~+kIUsMW2frf5h9M}uKb2+HW76@J%z83P5h;Bx3H`=rN#DY(a
zTRbnP1lAyxlOR6D(GnTG)L$cyL3=~GKDIsJ()cYRgL(&P!x=*>hgpadd~uW`^ig`!
zofbY;mz5RiRK6W_&S`eb+K8vUER@YO&jI(gD?gdhW!g*GB%_Pa<?Oz~R(<>HQ~qKa
zdr;x`nTvCuGEr~kwmy-Tw`A`s;iR{2rL!}}8PNm`pc^EJ(fz$F?T<XYt7EocjfT}2
z6LarpAKRwrEwl9&PSze{OJ#lgvn>Y@G)=(qby`Ov7Lo!g%b<ncrxFu0GEllfIR6>J
zH2P!XLE-20CU46z0F}(o?UHUxU_8x0j{9zjoZ5w;&I>x(iMKT0a3EC=*Tv)LJzP=o
zW%;0288iR+yVsea9ynHfA*S)dm?OCTamTtTtD;`abMQAb1-e~4y`_q7hk#1=z3@vG
z)}&<fS$<i{U+Zk1LyV>tDvaP-jmC6^5!M-bGj92X1YTGmbmc@$6!4I>{Nl7R#9~+N
zhUC>g0n+~Xb;~E~nvoO)Zd-@6g`9<M%Vi#TC-K-$9+m2`H|5g}wCf>Y8S8BcDJe?8
zT#Ecqy2f=buaM8xnO#hg1Ee5z863jARQwCjxL!4Xnkl%l7Ci?Y&aRFXm)oJFJR1w8
z8x{A8@k|n9s<^6MGC&Tm^s@7?gG}d=jq`V6jnC8H_a(0s<?m>Gov59IZt#+&b?JOz
z2lL21a32ayHNi4AWtGIc{Ca2E`VdXtff`IMuR8Nt1L^vD4j;-w0|S2Oqn@J#&j>06
zKS7V*-B5*sW|D>O&47)EwY}|O^?ciYHB6m8)vb`!!GRN^4;C)O$F)4UEAj)6JPERs
zQ@rXX=vwM|q&Y41lfo{$hdEwYd+gg-O4~+XoqV5}Rj7x&cUB64Aa*By?fdVnW2yO`
zHz`scSl=S4EUm3eN7RI)G^Z|orX5;tHW!LKN|PoFy)=BU<Eqc@pccLKVhRg6o>8)>
zKU#7{y6cYJlnY1a0powkY3xN$#I7E*!PQb#H6%_yx^kWUfH<>dBIP&fihKY;^~2nE
zLbjvwA*t3?TUC&>51!y>?TBb{{19%?hJHYZY)G<%=Jjc>+a2Q_0!z&|$X5!W%rXb|
za--oNeH&{recG)%*7ep>l|PF0)S$B;@8Bgc1qSfEK#6Y)zBo5H;h7b-9u4?di~jQ$
zGvjBNHUvNB$lt*{;FZ#~S_#wbd&}DUg(uVPhCg3B;Xn?&>|;l*Z}VPbUHp*TnctkO
zR*$+@ChBD<Vrj9ZJ%g-1TkVi{LZy>J+txM@@(J1Wf%}pZq%n`<c(~nUC5(K(1pMyX
z<TE$DW-%VdxD#5PX-&*)p`zbz>h&q+`>?iF6#P&t4JlBr%z?<r3#YE~5=NC|bV(D+
zp7qY`V&arWS{`K*hj>pBV}Bz8Npo$GW8F}|{CjJ+HBLCu%Wt3=%!i#_zv#3ko<-ZP
zK{+NJkdu?0?zb|rvs0Qt5B;`#60>Fv(=3oiC<PkFJ&S%)><G`G7_(n9d4pOFv0k)y
z*6hSdqNmzjzq)+}7DzPlzS2P#=b)3gQpZZ56l64z+EYT@b29Urw6)%`_*jYkVXmj~
zJy8C$p`i2Z6v+1A9H!l%(@4PY7t|*R9_e2y3tq!(WU4?il~1-fW84?8Rk8CnJ@`sD
zmouP;!0jtrB5H7X(c0t}@%+#X!l#myzm+JTyunv(Yvt$d6^JK%p$U(+b*Gnu*=a?p
zzkxfAn)q!Ue{j1I=$jBtKZDCl7U>ioM_LL#UyrW4TR@#9N<Y;RTVWw!WC4+)`|L{m
z!?h|c*x9*4{Wqu}?s#Ncw-Q;{pBJbFV;%=PI}|S~A?;%>jwRe=duC4y&=OtXfy!Z+
z!hrW($YTrNBa5UhZu^xF@58wgHJv4`rb-*p=!Q;x0U!N)8Up=qmMNUfVV8}kETTLD
z0nph7H?NrBdM+CeqA38=VE0?<>){dNx=fiYvWU!@INGhIVwnZ~{Yuy6Dd#1&YlL{H
zW2Y}MRkf;sy8`t&I>-iAAMuSHN3&KDV!&JH?qF@{fBR199;hSaG|rgB=9(<}<y!}C
zLX>~xJ=T?yR-e0{1<q=@61<txXj_`N+dAUby-rofL+6f|Pr6Iqq*h(hqh3d3dsRC1
zrw91>u56_*O2w4!Wmc%>CvP3B-4W8U#Wbyaq7&ql*$_Nk%VWbkt3ug1M-d-(P%Wb0
zb4!`6?Nf}CN(3TAMGG4#*xD-CAT{`RlG2<;)wR>->sfhtG9MJ=jW!&FyKjMritJhD
zh}y7|rc~@NAenJ4A5EwcPs$CUL5HBM*$*bQR+#6IJR;Ef5@MI~>kJJdD!UPurg&kr
zeb!ne+IDu$?<q;N=oNcY?R0P}vgH*KS=xVIJamj1z=^zuC&MPuhK-Jer3z6NjWQ~W
z{XE}DuG~V}!&}sSv%HKg2PuqwC%l|Ccsjjqs)Bb<J9fK%4=MU@Eb)sgor6fajF9a=
z#4g$154UTO4VB&wmdrp3;F`zSPCEwi8I+@eH@`8G?V+>E`pwgz<hFii^`RU{fMU)r
zb;l*vbm@JE*o-7aj<5VWACN!+Mx6TvE>Ku``ffU4Bn^>rP5!Oz>Tz2aRyLo~R#(b3
zg9iZ2pMC<hWjNcsOl*dP3Svm{b!qz2{X`k>OxrVmsGvEN9;&e++}y;`1l6PQ%`pW{
zJ#)ITg77W2<b1qfjE5O%oxnUm0M`J{91a$1JVwlm_quL-G*B(Mm~r`fjbUVb2kx3~
zR-dGda>RfK3}EZA*dg|q+d#sf+p1L!)8rfdA;#q=BO6n<zAop6UOp#EB|Ws5P8O)D
z3u3==(bY0|W4-MxEhHXVIWar$Rc^IC5IGjbXW5J$SDw!ht8&F&OSkp=c1f8ao^Ntc
z)7mnLGK})N8;#O(ECL-oPJF{aAlX~1(pY6CgE@GyF?T^4qHi_pWv>|?@D`lF^gMb{
z79EbY(T_JC8T_Q!4<%6i>h#6)0}3%nb#3}e;BfzI@|md7Vu2?MA==~~Y(xAD^_Fe5
zduizq&v@+isr9jHqH)5;sDfwZB31S=C+yw_D?wP>pOa0rxV_KL+fGiTCl5Yy3*3}@
z?Uduwjcx>U=WY%~bUxKc=L?r#^yXL+*nN-)QdPYlzDMQLQ0~677`|?va9xWu6`Xy(
z`(W?K9ZH)oSx|*onl&lt>Gl)jrF3yaQ)lZP%-^Y!3*|}>^^*elbqcCo@qrU+KQM#|
z!-DBuQ>*tDixsvaYzxF8m}{c3u?`#q&=^Bq4?h_z4k@v~x%KwhY_w1|qforSEZ<S;
zqxYl+V~j8RBp)TyQFh!QVFLB9)H0`|P*#)SlrOVY=`{^7t<Y4bo1mmLaEyeBi{jJp
zG8@6<;<|6wzL~@#ybMI9ItBryVrQq^WbUA`3Q{LY95{cYmQVU<)Pc3L?}PG}AC%Rm
zA!|AFB{5oJyc*9k!{Wg2R0i@$6%rLtyuE|Qe#qkjXfnJuWq-`<7%TnF7v*1)*{@&d
zO)tj}dR3I&O!x0LR|u@A8FoNSZ+eF2j)e0w3Y@3&iYbu|jveW+men9jhfLr8Z_a+u
z(zOmas@%iQa&%;9Xd{JqI+w8`@Dl0!1mga}twHl{&{W1$lR@76lFNFON^rnN-}5KE
zhaLvZZx{FaqJLfAm6ElZ9*DXvIyF7)9+k9A*&w*>c4(q_`Ge;Z!3&#;XT%OG5fzVn
zf^50kPwP*u79%unKT@k0x3OiqV_O~EUo4&pnHo@Gf>fdI*0H#dI?lJsH7^&x`IWP_
z-sYDGL)UX8lP_uKE0ZCucTC-ZkT@bdCj0WTDaXEMBF}}<p!PjJ^TDDVi@L?@v}vy5
z8VU@jay~DYD)5$2Ash0d`J2<`O2<&DkV|TXdiwmE4?So)Ep>kWUvl7v1#RmZHNgS7
zH!F@XHqZ(S$LpN~=tb#|Axf&4ssX+nl~jGCe0<TI8Sn7=+tsQUsw+Kr`FC}qw7@Yt
zisx$#ZwO;*QRzH-g<3<(WW_$wtBCKpy9b+;DilI>F_rY(sU%mRCwT8anP6pSurlH$
zkYVs&_${}Oy!T-kg(EyQ6)SHX-1M`#&fTK75rai~gnhTi&Xs7&muDMl<$a{V63f<4
zq+!lOvjKc=oTVgG-G=Q%_;<s1m}(-yg+va;o-{Y17cv&|ZN3XImRb%Am$ow&NV3r&
z)~EMd%>^=qo^wVHBQCS7svmxY-Lwl#;q^2GdP&An9zez=>Sw(}H8S!mNT&@1_4S}0
z%t<imjNj0RwJUTtX)E(hXCS%GPK!Fe?$rwsy!Cci256=~%e2{*OvF55G^YD9Uz@y`
zT-)dke$(Sl#-sOrADD;iWozRYTUh5B#YoQ1p`mkdH_OtLDA$`5*aa({Hib}`ET6V%
zo|>8`x-<8>n8mHwa|6$yk0;dpdXZhrOwK3OBlO_vO@2xJakGHs`IhR|ctT|zqHpk8
zorX*Ds7~yOuwC9k(yZMLLd9r>^s8QJ64$;l-}>SV3@rshofY)q4ZKV8T!dH?6<$>u
zw1r$S!SkJrAurUd;j2mI3)}Dwjyb`&Jt)5)BKiEKZDS4vSh1cxF@d0}pZxrCEsi_$
zw<F94ztX5B5H$9S$nh#n5z5+rHpD$A_JYc^do^t&ttM5gl5EbpU5MaT*Q9{8bWoNl
zD@K&V0SNJn&5Z}_1f$hRX_;=#6+%31=F7Ew1s(5dJ%9Kd<G+em4XmG(dtL(L&o#Ll
z@rQ{%eEts;eISN%{EW~qPD|_+POU}@Y{0qsUOHcFu1!|zf1l4BF5Q=75(?pvMwy|M
zAKnDSB~q<^HZvQ?Co{D3Ge}CkJh|#i5@@OiPB&M~M^#o*O~H~2zW<mdSML15iFiCk
zK4=5;=$-{46f}$LBva?rzF@v}E_+>!Ud3(0`HS17%xsrrVb?@)jl&0vu%8$E5Z&nt
z5szI>gP19{WGVl2snfy)=QYi0%-F6s=5!s0{St?Z&ey(F;^;kE!O~m)!^v}3He53Q
zz5Zunq=Fd=e(Ga>`+BCbm;#EFEzrgo?u<H<tKE2D#=VZ3Me~H7Y-47Uc8p!Evxjqc
zCVzSxmwnZeY4K}L<Hb@Iu4^~qa^uDzhgjn&#K(&C9XNDT+xrEN;F)HZ+SnN#_!*@@
zQOe44SYc@Fk*%w3<4puwGtHP<Nz;)(yb)A*7G(e6djY^0kM_<qu4I)8CX^;udP`02
zGqe;XzH?DnQ_~+R(gT5dE{at&5S1F`_VR9JKz!g;cGJ=Z(hp$A2s>j4T3>z1{c*0F
zwJo(RgOv_1f%(50?EhNWI3o0q1}(|XzR|^RAJW!GK6~x+lK{@IVdIlL9d%In+TyP7
z`d8mG4l37{R!B(85$A~HYzO0OSTt<MMpD*l5{Nn529HL}7KR2z=f|&6BBT8ZEt=`<
zzt3(~v6j)qm%p8Ik*M)Uf{zzo*}nL;L7O}O5%AiK=M4x$;g7hkblb)X<*nD(JxO)Q
zYHp^CyG-)q#KQZ$L5qlmTDXBe92l9cy7$1!=jZH}tza(tCh9!`|E|MgvZW^%X@vE%
zC`TB!MAON{5EH+@#mlWc`_y=U{4f3%Y|~FKM2!lSv6uX%)sUYE_4o|62qG5R)j7FZ
zp^Q{&Pe$n<+_@U!DlanTN;!wm*9X<|U9HU{6C8Hu>q}>1PqtL8D+%_t?kAPMRU(+3
zZFkpKYLt+Fly4ciS?B%32x&fiJDaG%F%Mp$M{R-Q+tI7q9-JddS<o0JPR_w{eaOWj
zUBJ1{?%c`s01{D~+%>sn;n^<ApA=mrJL5#*S~WW%Z(;eCcKITG9JgVk{V+XSZKtGR
zYFqqbZuujsO+PJ$Qdi#8vM$EMMY{3O#M)WDt>2E<$2r`u)v49s)itj22Y^90sp>qD
znbz;NKa#q{Fexaa2<M-Oz`Z-jEx$9he=d6^mBoqwtr~WE0n$wpYXdcCp))AYk)q%c
z#b06r(sLT)M<b1+;?$4yEUxKLll1i*p5B?w3>J$X{0(%)OPvm*w*hbLTRww3RgSbB
zf0_`aml`WP8Vv{oFy0x3dQwfMWTa$w$h=n*y*0k+Qip1t!Bsf;aepg%7p|h6{LX<O
z|M$BjzbHEbt2%hg`~BTUCv=md?_o{IWZ3PnoW=Y3u$ZK9f4{{`QKJ6~&9H14+<tKi
z`PH;#PKF*2{K2M!4=yP;W?&_T^2O)dbfW1BGoWgnd>~&7<+0b_|0thuyUqhi%(N+Q
ze}CuEt>0><q6Itd*w?sQT%L;3kB9CSSkU_q!^{Q-8QQV<-3W?RC`OFSzcrg!aV!*x
z0+ZXFpphA%-<G908(Ul23&TscJtMVc!w`{gLTH8EIj=i9N<|84hF*Cr837qJwY58E
zJDg8Yv+?c#J+Qqovz*lp=F#pAu6&|SaAUB}ZfJzy4a%O0%6bVYkIs__$Q#&iI}o+p
z{C-e&kV^fAJv|y~mD~(*+b;Rd*vef%d7j$ej^lh|CF-ciM(aQ0-(v^ovL;KOyFQ6Y
z#&9H_tJutsIBeBO-`I<_2ee9O+08FSulZlCk{7jnaiyAqr)X7{+joyX&aj*bUdl4*
z;6{r~S2qRQ8;JLpD;aucUR7GARiiDo$VN$Py*TjRFUH|NKn-i$aTs>&OUyz)1irs~
zfjBRfAJy`VEgL#Z_N2dQ=pJ?U<Sl{nsEG$xb0vfX>mTx#aD5wV-;1a&g|(1O>IM*~
z9NDXN5ymPW^jEaK#<qjO``Ya-M=8uN+7Oab+P5C!0~*0wdyWDqZR(#+AKLZZzFE^*
zvC4D-#DLbsZd~5oUb4)~Mh%rZs!^}*WmS<xL#Z|&%8)^m@AW%2V=U>SsHf$nW2zF3
z?g)I6OnW-+`lggGaYky(Z<{i8Pw9VFn9rRZ4H{p+!Yj-^D6vmgYeu`Pi7bKO=#U}J
z2kTi)?Yg7dv68O_D0nnu=YY}z=%QQCy}ODyp@}ivRSZvA4E!V0eeb}lv;&d}5=%LJ
z4B}|J{e3&8kQq>cdF%vh3JEdx`0P8(CA=)<YV<@;q<zyY=^8y}6?&iY=gq8Z;)s0h
zBrLt%ZGSMm6?<ycWUyh8MAJ^+^%<z;<ROsV7Qkm{w0c3QsiLeEm;I_D{W$2-ikkAT
zy}eeG_FFC0yMM^VB(JZRbODEr7COQ%G%TyCm5d@&n*0o=rd1kkrzUGDUnu8%CgIm@
z>{z5X)_EpsZ^a+sW2?$^M4>Yp;rtUF=T-Fb7k<DCSpqV}M?&5nFxHmK$$KhgQC%^9
zM5O%G+`LVxM$78l&t=NUI=cRDAu<F#@6J5&yI!(+)0k?<EHNLuXJ!6R{54=lXiGHW
z;?PwdT06NuxN+^*v)@3)i1M3B<w;=xaQv46P7fvH!XF&3s9V`>tSDQJo7&|WRMXAS
zWg_3zE1f7(X?tteS9@W(I=>Akv8C*lu4f5C8qUUn8efd>U(HLrxG=MOc?_DuXxK<0
zwfZtsX;XHeD0j|=hk=ntGRiJDNw2S(RSgpvK7bL>Xc*3^N9=>D{2DM}Rqb5A;N!DQ
zBD4v+Jio|6;ndovv+$Hx##mnUj4!3Q1HC72ZdQX2WIX3cNzG2Pav%c%EWFV^OqQ=#
z7klq>L_v*`AM390^($U4YM`$P9xxe-z2A1yBImr0O#Ls3*RoM5r?MTdajOy=4XWCv
z_FQCHsq{3gelO>HoH(<ci0B=QN_zq^m7*(jpYq~f4J={Vop&y+bp4`}(usb6Naq<@
zC{$+u(UtigVQ!VAM?sUnW%@*cn3&jvVj{{nQ3!o)N$Bk{N87UANmvp?GPAMqgD{bR
z_T_KY%2km=vx_##9rz=pw_J$<hOpBIbCS-8ly1eg)aDFUD!c^2Kl}I3t~7+cK0nvN
zY4ko~utZOs%H&14>_R`rvNicLOEgq}VZ<E0JMmxojYF0~uuS!qa<&SOVqSyz9z+x<
z773$4BRx8A{ENMOK5T`aN`DRg@Z9L?(B6u-=Kpy-<som&ACI>mXr@lT#q6K}cs34N
z?Se|lpQN>16@lGqM_u-*#SXEjc?WluDD7T19k{K(cnBsCt%WB&-_|RB{dqaSnL+$$
zpHTpDvB{L1XUVk+znOt7#A7!khK)3|2V~NWYTm6$jPo@FSVu-7cYzvlM~`b3$ZFi&
z)v9uiv=Mf=A->wG5BBzgq2X!07%hfKX->ZV%wSXEMyR<JU6wEml)1VZp-1Vht(&sr
zkF0CH4))WgGhOBg*xGR0)?}Y@o}SNDIB#yEI6s`q-Ol*<5p%wcCFQ;O5O0<4o2ol=
zzeM6U8nc={t`CdCb9f2@Ud1QT6jZm_r+oZc+J3rJn)Y?jpqhmbi3S+PNr1FsD`X2o
z$qnhHx1CwNK+Nstx{Kt902gJqg^=||`O++B;}AMfYWR6?d{T_J)6xwd96n>;vDDmb
zH}ROfe=8(tu84y@$k{$xnz9#J6C|Zi`rI74HH1<fal0@+I1epYLdZd_<=b}2=sB{q
zd&s~<N%r2EJNStQ?UV}|Fe{zne@LnT%La<c8kfevi?x=aZ~x%;YJjaf!eH00;&OCm
z!PAKJ7<|w443S?f_q^e5v<xd0HDm*uI%5De9Qc`5Gy|Lz2P$I&R$FsV!GWBFNQabi
z5>g;@&fv^F?}HZl5<mc{`J+V@Ifdvaa#S<r`I~DjS|?!cybl*<aA#tewQ0G8YVv7V
z$>XoG7WkjVQwI47Eq#n_<@|Rdp*;oad=E{>8*cI}^DM|9tG`6yk+T=GN}dwUOwuF@
z0ga(JgpeR7DB80#gP4+1iR9g*RYOdfO6u2Vfxyt^u|L!%?gQF9w)kZpkysi8zG7Np
z!yZquW;W(fhG?waREgYi+{T+gZm_uF?e{5^M#&5M8LtLTy=40sq<RYI_U5^@IGq+Y
z;&#cWC*EYen=jb9O>rd!`$9>78kCuEuobPa6A$LM84Af~s#T35somAl4ObfV^aHr0
zjZo)g-^4N}BIY&&9co<qJD9rIgY$5ur)@r&`Z_1q(yZrxR@0^8OMNc^b}{bw)`!0_
zll;NzZZ^8fi3RwLFj~BQt7W>)G2)#gLBn~%nx-#81vF#_ORRZGVh#1!$T)H6tVbAn
zFa6N2lVl8CJDI&rFB2azp)<3es1lz(JfE4SK~)$Bw?F#u6PJ|DR)Hb<dqYM$&D?Ay
z|2fqO=T*!8ywydLIlJOlt5GlPLpIM?#(U?DOoi5^e4blU>LfPZ7qgigaq7i$_N6;J
z3UoQb&r}7@#@=U%`JO%g6?I%y^|h4#KIk!DF2Du-4s6uW@|DK{o&CmjA4$)Z?v}x<
zDOo@IEl^?%X$bZ5O>VUsN=-;gl0V;@%70}dBDDMQeQt}_ONf`3-UW@<<)l6m<R`L2
zh>SVR@N&j2dAR8VOGyK(pM!|W`isf5E$0wQOFcZZ8rn(uqrk9lP>$R7)JcrnWQ@ND
z(C>7v(;UZxhmD1C=J#U`F8zlQ0iVCLk^C_S^bf8~+S?-U)_q=aBnYyAaj7j#9TLmj
z+#WjUmAuVq5F%VKKcKvsZLx=O^@i}c#5?G%11CHrm7%J)sn3YYoqfIdJvoRmSsIUD
z$K!%uDh7wf<`KOJE8)pAUTD}?x47qVF#d~T4%Z<=&A=lI_Yj&!3%-0>Dq_R!)qZ)#
z_{8HhT6g9<O{=gipZL`vfqVBx@OH2l?6t)C`HsSp3sL&|9O@Qz?B^j;VuuMrj)gKw
zyFTs47IIvEH3M@MWoBk3%RF0KK3sy>9tElt^YnW8H8?v?Km?8%pUBW9c5=@yo49@>
zKjg{E=g`$k7xB111QX)lo=OUHjzO2}!6q@AKT3Ug2=rdii6bGfV3T0&%9~;`;`vyx
zVy@{Cku_Md8|KTUh6b)sd199VTZ1>;K~-#tut$oDRl`Nja~Ux3mqBnW-{W#6_XQ2y
z?P}PAud>D=-+}`<F??K|?k(||ONxu$h-Xrs<~QgO(kS!a>QR?XQDc8BeFRfQNmHJ}
zs3l_pu6%s)%>zD6O*htG{UOtkyQLu1LzZS?$N7XCV`!Y~9wIB_wlT4NrEjm$jo|xY
zkV?kGzV}HaL+ywJNB0(WFRQj+TRr}4bZ5KaNn#p?y1wVFc}-Z@uR;N$7Pi|`hWv;4
z+FdXp+lsl+9{fq+w@u+MUB7?I$<dZnuz5+vLnw%Vvz7i?9|~AsLO!s*^z6#YE@0@h
z?MYtv*{_6Ok4!28MW&*&4YamcQ4X{^W~QwS4!0~Ra2`5G(}mqi(S^n~tFG#GGY+Kd
zoJqX-!k#?s-f2x^WqN$MiF!3LN6!1Z$^+aARIh7T5k`-f+-|YXI%Ld?0l-1#g}{MS
zF%lS-CjBcsja5Ws{^%M@P)5Ug8|JyyHd&^xjU1`fNGv)z7lz4$xlCS>va~wZ%d(2F
zM}!czz0U8wk0zN_{@yn{T+q|%x(bKh79?>QCN&+xN$qDU@*82^(VIhQ8Z=a0+NHZ>
z+aHj&F$+Jc*@kc97WVk-1El)*hT3aU2i2pJ0~akfz^J`B{<9sU1N)hC!(WD%Of*9U
z^T*W*2^ktx7x@z%1307nJgu_E5YKw90@Ijmy+{-6h5{Z16fgo=cO0nyyynKk3k(DS
z4dfqd|6^Adw-M2jkr$-jfj)3TbwKF2QigcLfV`cmO*Y`KQ0XOT;3z<uhP|_?D2rj+
zV>6h?LRO{t+az0H)Zyppa>GJ@xx#{i*G5K0aTu>F1{CDM792rr1Ox<A2T0ICi_s{2
zpjslGc!rtCuo@Inx+kBcb?bm`zqpUqg=PZuH9kK6?nJ%{g&nLNMC`S2f?B=27|SiT
ziuU$JCno#HfuuHf`m=5~tm!5Io|d!3a0o}O7<JcpXC{R|V%S!cFtsr|Awg!|Y^=d*
z2wNn<!G#7ZIDnWuRFo&iFZv=sZ(JDWzxP~C->9#@KWAegd7{B_VLKCiqZLd@;;$%+
zR4nHiz1gfcZ?1sNR1EldU8FDZ%s9y1axT~-x8@2aOx>(a!O(`vr9zqkUD->#MqpCy
z>r&nFBE3qJM^t-aqtT9YAdmZP28}XmUOx>Nh-BN4C2jPk5AksRKRMs*y8ZvmF2R49
z-TG@m8>nGDl<)vi>bbxFN5!TCGSDeFl^N`OHmgzxZ()@2!}jRYuR)(GC2v3B7|jcx
z-I*<}W(;mm&)c2kfez)%sU(e~;H^8X;S;CNw?(^U2J~)Z7@er)_~Nrn#68~;VPHJG
z^RG`3J;TamXu!iMQXSsFK<w||1?|ZRTU;Ll#f;@Z>R`sm4UU`a%J%e4T}u;x%hstV
zP@~hT?5g}`ZgHkI2}FWquErxxWO!geyZ18Irn5&);2N2>vy}-B;iV(Q#Fo1AZx1~y
zen%Evg59}sepl=baQ(v0>P@iCPR&NUpJw2?zZQakRa`vy1~K65l;}Ma!91Do#?m;x
zOS-YXJ`R-73i`qn?O+Xr2aENFG>F5wV7nTs0?p*lV6K7+apY*7?L-kU-7`A|>m2uA
z*NpKH?=|L!X41TfbC3F)>r;)E;~uJ={H@Gv6ufz2CRXN#e&tKY<Oq~W|LVDU&;8wn
z#>!6jraPEGNbs*ukop}QwA9Qy$|OlPw(!IDJ#G;fhTsgJ6W$ijS<Uf}{z+!lN81nS
z=_^94Zbd?uQ;#s@6naM|H6-%^_X(96(vtg|dNiNi2@YVjhh??gUMx7D-T1+)P0{s0
zGXsy^RL;wMnwPJlsRZ*}HwLuLq%o8PtZy{LkB!W-MJgNo)fwznL?BxurpZ*1<J%)I
zXJ<+ZmU3zK8Vx?#qF>H=xtkb^nyN4O87&ub6dtt2WNdA>YGiDbnRr(03|Q4W=PH2o
z>)lh;l^sU-zn&^L8QVC@#KgxVUcZWdLap#lnCr@q(s;C<BN}vR_Ge?@+BufQ3%v91
z5B5up$7jJtKW@sFkJox_=V~nL+_uz#DZ`|K_R~@QwJr^Tgvc7QwX!h2@eJYR{;OBK
zNA^8NY2}{AhZBGP`}yQEhI#QM#=Tn_<B6@}C6}{ptyD>{L^Sow@!#y&eh6Dc@@8u8
ziBiV!3F)%r^$`}~<L7~P(`ChE(vCDsTl@MK6+5S6aC_VO^7b+Vk%QC&U9j<C{NA|G
z&!0ahJrRdXlkFhCsS4loT{*$orvKqGnhbw*VoT-0B^V?DLXJ8%F3Ec>Q$9Y%Y`;O;
zBe!#7Z8rK7^DFqBu>WFzfG<9$2s&-&8q{G;0P+IacbkJLe>!;@VF8eoh>beV#${iQ
zjm^yxscLo7Rfd0;=^t53bv*32j}Fu#yJtX-J)9w~;y7QIts!ys35kH4lFPB4^Cxo~
zE9c5b=g^3HYRbyW7N7oz{#PUZ&xf1zft1M2Q7mP%*O>7v8U=uT;(q5f{Sc9s1w?QY
zUE!o^$`sw0e_z+pr#n{Ley<2@Wg?_}eMRbR#y%$Tn5uIbHr(%P`LmI7w`?=aw}V7K
z_Tl8=XxaS5@G%~;{;_;sg8iUCvLFRIP{PUS=}sq+WSt0|4Q{OX#kZ5fTktIc9Hi@>
zK=E%N^B?4qQn-TyqPf5SF;Jc!@J&Fu`Bn`8Y#WQxd|8N!x?lhE-SKKSJFS$IF6L*P
zzb}jaa#Vrw4u&XphX09@FLI}lZ8ZAob$wIQ3*Vcot(|a|j4EyI@Bc$Cc#kljZ{tQ3
zXsSbjh7lWL$K}?Po*2Lzt&vD+^PZ^ebWyMIk9P_D(X8u4{|+;MG_=A9yat8GSwpCb
z`T6r_CW$Rr0Ll?)>?0afo@Mg@T55LEy_U5+;zeHlJNEt~7R=Z?_mM$`{0Xz6*zL$&
zj)?H^Y+^?F%Of_+`12~W&LV?)yR)mU%w#o;qc?xm&>f6QJbKm&y~_`7&xv2SZQlRm
z%2p#8YD+C11O+8zQ<lIF?%ia1tO(KS>T0>IYOKF63^*0?E!#POCi_zW13kAbV7_#9
zJ^nvtfw<s_nD_IxtIf24oQU4P;S2EjW4Sy0XVDDd{+Ov9*)pL;Ak;-&0NU4XG)r1)
zrd#K?OmRiU+ZB<&`Vuw?08v;<hODI{cu}<Csk_tVDr>#5h&lejgyR7oXxZYSKVr$t
zik-Y4IK}dpt$=>z6Chgkm6g{vz%f1ON%lh?dI19-JN&<V<DoM4ii~7JU3AMl^U<t%
z>=;L7>#OWXOhO_e2IcbJXJZNnd_R=_hYC)K1Epd9BGTd4itjG==iK*ad0S40MbVdU
z+{AVp7Mya+$~XaY%^cLCPx#w@4A?{m8%EQjO9U=OaEg?P32nXYL`Z>Z%By5!PyOe5
zQ)Wx#Tn2eK{3N8LY+_=IW+F!a9!>n?KdgR$CQUCyr11O0I(q5As$2$jc>3n;@8C0)
zbRPS&aT3=@oqXe%v;=>V$saT4#uyy+N`5qcmn;bIQoFetW-gP~w?pGZCH+2cQ}Xhd
zcBe`y9AWM0C3=;Izn2sr{Ev6>{BfB`fedqYOp^OT!oqsZ3%oi%`iq_uNnSrEd-P~@
z{a5@z5w<tyUri%0kOO!JgyGYkw8FcY3S;*%If`P+9YMQEycT3)!K<B}oqqJ_$;QA`
z|L5`bAiw|Ndj2YRB+fts6wPUY9Y6+Cu6(>6uQ9HzAujiPlfpbaHnwmp!!IVodo_GD
zzSVkQc$eY9|9IQAG5~MT)t=64wUwjaoo|3f1K;y0JvJ7W&?Er&_hu?tTzIjnpB=3P
z6P4%?fN!n}T@}uddzBoXY;0`P_%7Igmz(A@>=oBS)vC4Z8;0kO^8zh<6_6sMX4gkF
zG-@s`E@wSd&ZG4+CVQ0(*pqGnR~r<**T=m(zz*TQIY=qub@)Ak|2dE~1(x96`@5!p
zC<0C!tW{d1O&sL0H(i`^_5Qnr=l)|F5%(|?-yLRyW>;o_p`pkpI8SeF6+;4%<u(9J
zg_DzSx|P2Ht)B7EKm89(NDv2ndcXviHHZj5EY83x#r|tzqMTu)Q+@y%Id(o{TABXf
z>Y%)FSs6f!d|;FM@_c-tM5pvDboOttBVhTge?T~y*jsE`61%|}=7f+2l8QngDVT2~
zk6`gT+s<AjnJe{s|30T};!k7NU$Mgak7ZK6pxu(hOi4*`KUx1pLKGDhl?NEc1gVCY
z)~Z2K#uxe@bl_`Vd@94ZcRyDv@csw(Xf|NK8|5*)@opi)cbKnJy8ClV@)hP95Ei5z
zHmyBC+wSuREjU4<15gKzB!f9QIf`%p*DleY1VEYP0G$TnY0yxba6Gmj89o6)E+C$9
zz|?J5I)Wtp?cP~0LBU`$YGDZp6W;jwzWX3E{J&-!fanbWh>i;^wiJrZ^v&MWQ|NPO
zs5Vdlv03Yh2BwIz&`I#VsBwa|x4qY?G!gIS78Xu^_*nRbd!)hN)Qm0k2e%EutZk&2
z_S0qAVPSZ73k~J2loMIfSS-aZN*Yp8SF2WqGNBJ5egm*^o?QJuHdsv@Z1#j7{KtL3
zEJwgp29mgP4;C6ZX4nA~%R1TMS`*nCPS17S=#O*QFfubU8>@b#4|xWp*Io$b^jy?N
zD$w9RZ!#gdAe4UNz0lyuB!B%>CS(+F3mp&>j{?FTVK-B6J{N(P+|?z!JC=)E@3fTq
zC*}lp$Xx^d+bxc^2TueP?~<!4E9WGe_$ujWujNkZuQByD%+v|Y>t^bT7CN3I<_jk1
z_9kA=&lJvF)Yr{dJID=|=|vMZ{MFXzvHPoe_oy|oC4=hiXU2h}O!B9jLxt*uYsHrJ
z_uOBg1q4fJ<;*s?LWW-PuGhwCH@x3VNXn)mc#3ZKGzJc_!S@52tvHa@bHL(8fW^(X
zM^W2p3cey_+8)_MoFnW3863QAVEAd-8|P4B$dyDAyEPF4+7URa!8jWP##aNl=@_|_
zY!2Shc`FnZ`Nm`X(uQpE#jolEL9XbAslusYTSIhP!_L{v;@k|Z^1vQp+gn4%n`<KY
z>`$NczpFnm1F%uwC(=#e)h_%eqMH&(^JQw^B5O?JnQD=4`A>Jv)X=8+z?o8{?HVug
z!-iJNS<R2L-wK(UMn>OF?_stfAGqo7PI=Y88vdK;en<h?-Mi78V3EB1e67mJ7A>gn
z#`au$JzEXKt7k8yHl}vIx4!cA`(<WO#gn<oWh4tY3xZuzPgth10*}&d3Fa$UFxhwa
zFF`8(H~fyDRC)7i>gokR*rjzV@>$@prser2N^^uy^JFzp#NBj@)o!7HFu$Mxxo<mK
zIx?EnH=F6pb^PJ`-?w2=?1ztL9wbOfGV?c~&RV}LQ%xs0ZGV5bJ%#`61h$xu^Epve
z!DDw4h?E{^9n9Aga*3YVv8p%npVpf!apRHvP2!)#fItDi=KmD-o$+wJUDpy4k`ReR
zw1|=@A$p5W^d5u}HI31G9X%qVM~@!8n=tAyNEo7yJ~~l{QG!8~;W={O&+mPo`}uoc
zALhgHVXkwoefG8X+Iy{G{hXPph2U9N46@G1_U74%<})*nTYu{OF$Mi88{@cgdMK$v
zS6G}}$sD<yhwSefu_*$?C?8)CLWGENK*@ieK}{)1L;#<Qf85p&_s_qCKLI$ez%Q7v
z-)18Y-}8t{rhZhPRyWdt84uzfm38+#l~uLimgMsmIakur{uxkA`Gz5&yqDqPiOo)@
zah>#-EZP;fv0MaAVtAy44cl?0Di!q|mTSXWq}{aVR2S{2%;($xyn+B3<@3UdQO5;9
zLqf%%)TX+8?&=cyh1wKzprJn63NCbeCWw@OnwhnwVhE!A*~^prI@2KwcDRCBy|lOL
zsV{~2<INxW4hVvqR>t%Dy%KZ0>o6`sz##!pFpJ^O=^LOk?ngcH5h7f!!;BDNt0^!a
zD3gAWyGR(q&dmHHwUp$1k&qE~M}Z3$N%41)$5Zt#MNU2+v?30H8Eo9z3Bq+dch41{
zWVCwT@kn&2iK8x$&gjb#q1r4UWOKFx<}^(@VS6iV`9q(0YIHhkxdeWZ?+D%YHFOx@
z(dx_92&u%BWocNx>h^?JsJ#!P`HhD^jQ6eiD!*@5s+w1axi0wHKmeZmQjgYO;cZxS
zxw6?=sy=gBEt?@aBI}+I^D5fRe@~e6{W)=MDnk11q9gE@GjL>)Gbfz!@3O(Oi#%Z8
z>3@0JYx4W<qh7CdfBa2b@eB96)TGK1eXZ+hoYqTQ3A6fxN9jGCY{`MmQcHr<+xVN-
z$+r}Oj5>z+g&WVn_LW_RnWHl>pjuUmlB!vai(Yx|X&}0$BT|FW8M2Ns{`hRd-;W>*
zb9rlBDp3xY>noK;5;Dfb|2#eOWs*y&l=UCgG6j<%D6Zs~BJ8Pu{v(40Tx47%c|lT7
zOY9VkK5Ymq$LiWn2^SYDX=UcEVx#*XCJpCw9F2m;g<rQGdWVyX`LyMv?T4yyS#>*c
z@SQfqMkg_31Ku%Uk(aPF?W)BrO?~wKH`!+E7gspUOfz1^DB6CdQjv}Wi)J00og8?S
z+H8a=uQ8Q-_%NBQAqPP>K;G#>IMurmkUjWoM#2jUguu@afg<`=k%Z6&$w={V&u`k!
z9{2q^IVzDN!+4L(e7$s9oVOr%#x=8M>klS`-r5!`9!IsG$ImG3v`Q9KV_1B5SH6+s
zeNqYPUav$R&#QLmJCCFN`~uxtj8z<X$~F*rqUCyz({3_|8m1VPZ3dc#zSoB8tF~Xd
zAYchd>SO`LD#Cb0HFhDnfW0z%+caKtd6|=_;x^I?6z(+Rn(em44w49*dEB)ul&vDQ
zsT%nuOyc=xE%4ZqmTMhyj)yyd<6DfW#mr<_E|v5dBnDSz4L?)IBg*%4est?XmaRaY
z#-@NzLcRNSVFdhz1jRq5nDdJPWS3U-U#7M_6%+|n9nfC)pZUW-TJmPQyBRY3IQA8k
zneHaUp|FOlh^WkW7e!c8FjDd7d8W4mafwYCif+F4ft1`d?F$|B>9|L1ihgu%*lT1E
z`GiKS-$K?<<HTHA$~M;|(4-u|Y*tr2YPWw$H4rn?AsF!YA+-w&FJ8R>8s578_od)!
z5Wh1s#8Er<a`(-nvv)|Q;y!MtAL9n6^IP7Mz3t{lH`YMe9mk`vRBmqgJ6-xQ{;1+_
zoEd<WA`g&QJUbqAv1<RA<BC?O)_c$pZ9Y`8doW-kX|21(s%I%*Kr86sErw`%#V|cl
zcF(u^h!;={X*Vw;%Rj73O@+rPV%k1Qtm|EZR4XA7E=_}>-dai=7lqk#mgj@2w8f9!
zf1PRilFnnm@gX56y4*%Kx_3gMl;+hv0s=-Q6?xeg2kKhcW1o28*%}hSxR>?Pd`EK3
z&?q(vQMG|>i}jgGV0VY@j`WscMjh^0wZv<z)m&N+IqURGeASpNdRWEyIH=s#NW+9~
zm?PuCrfw1|KK&U3R5|bG^(8An*gRZRevERhwKWv`s}nTz!Wm)ov|3@lW~*&59Y&!G
zVp68?HoHd@%P<Fh>0Tt1Lan=_DvLwAJ!(jqaU9saO8;Z_^AYgVwQf$35c<S}bHlT|
zf&X_Bi|*sunyM$oCZcIUYabh)bg2mAw&ur=?Z$jR^SrrfdF7|Y>?!^1q;dgj(Xv3b
zI`}*n=2?^0UAqVg<X)q;IV7`+H#a+cwtsCae1Imo6W{&=-!;+)sHWq3%Ke?}`KzJy
zy36f8S=>w5x$ixvT``n}YLek7Uo&eBS0=G-{jZ<{lB+4E(<PKTffpRF0Wy4nu`zPt
z(APXIB=K+68p>?wa*BVieQuX9aEvNHv!;tYu@(>A#&xF}(j@pkUw#O`mF~&eUs6jH
ziI`<a&>C_4zhakHPWn@oZNSDD6WO+a^lp@Ud^$*mvO5U`o2O^xMh3$UgR3MrNYC*m
z-+GR;K^sXrbxg>^jQ)~`{C%in9k#yAh?0xYSeE*E!Ki%P<lLLBbQiF4Kq4atoOuP3
z&apwNjuu_IThXpjy-1u`6XK)u&ULkfka3eu3@7Dv$h3<@s{OHj!=Tl1?dj61u1PDA
z4xQgGA_hwePzsO>V*Aox7+_7<<l0Li|HW9}e6K04YPsd{(VM;3Vu=X_#$@%4otl?(
zjA1fR_c@3s2lVUbkA5}5Z+^65vcj=Cc|Unm>J}L#C0wZR?pof_&H}2?2a<O~0cF@f
z*YyGCy%G7|xvBcWE6DWBFPgJZ$!BMkw;O<|7|H48OM7<udNJefPs8&8*_63Af03cO
z0=)3GImBItFs#%dqsb^vG2A`<Yepo;P_4~S;Tr=lcr%#3i}OitbW*`!(UcGEOx8zL
zp5)GTOjGZ?B_&-CXj8H%V{dRy(N)&>_>y(=yP%^3Y}9ixx{(IwlrvIWWw#Q=LT6D<
znCcwT|7_AHVzeNQwS6_==SKD&hbxfkDg}q9_G0l0G~k_8K#uc*7u>0)Zz>_emRyr-
zp)G#%_j);X9>^mz#Py<GW&C<OpqoBY{*%1`Y}cwsOMKkGERgJQ?{OH}UQPkaU}u_7
z7td2A4=t^1H(N||G})e<VMiSGc=Ou@g!1)<&O=}jLBllhS(0#cjsW-u=6kaEs+VxA
z>2M*1Tj3qGxHHYHW-nfucOy-@xF1vRGPX5wb4WM#@S*5M#s($LbUpOmnQ!CC=dJ?H
zB}aIJl3i?IFsJmidO@9WUsowgZW46WiDv!V?!X|X7HT&B!LPZNaC#C}bLuLyA^!Gm
z7^JL6V5ehyPP^noA+scF!Ac?Ck)iH02ZY4^pLhA3a+5h{JQ;f``-%n2U*6j#^-eDG
zky{~J>w|5DR6{i#xTjEx`x0T%qDo{a)Y#TfSto^e?g;Abt;WK!F-d%l@NlD|;H+8b
zxmr%q1FC>B8U3uea<BhA{TavG6Qo9Q;V}ydwgWB1_82>-zdihR9+J{@N@6lllLF^H
zQwf*~RU;pyw41kDBjXx5yu`JDnOC8+<dzn%eC8EmUVh(dI@kQw+=%o(<x<&f3_+vX
zgSA{~b4-(V`OB|pi5-3Dpwv=MXKf>V5$jK`yVaghr0{IpYXc4gqyujIq{f67(&^qr
zrDd7>I&;P?x7Z_X5(917p)ZS=O;7?$JXVImN89ShO58<C*Ajl{A=siVHKpyJs?L79
z^x}P3*As82hMdqIpHb4hk!;T5A$*FFac-O9ha#G@H)H5r;)a#mgw@Kxys>o)w`*1M
zpkWH0roQpnSD8a-5)D4(n-T+M8%~=$6UJ||YChSlpjUj`=5f+#<9axQn_1ZToq^Ph
zhOxy}km(NCPUfK<IoEW$3fjF?b0I`<6pS5h0`3j_TKQt$Obj46u&+MM+utR=hM!H}
zu%tciPZ~d(=Ux%Y7-%{g;J&Nhbbs&H%WY=g6g_Rfy&K4sqC0#Old85K7Az<k?0mVR
zr;iqwjlhui(`%N@_CA6%0~J@R%V{+Fy6&U=PWRr33mExxeK|<yyRL_@w=g>t@kMA{
zX~*y6>$}pBTgYA*aEzZx42(d8MT_huOhk|B*1zAW$B`1l6Ht9QV9PnB)cDL3OW}R+
zw`jcAfgAhvM@Q(sJEoEz2Nmnt$L2M-tT=E{ULUmH_^yaF^EBNHee+mfr&fX2+p@(S
z0!Qfyhgnj4kGx`hDmW5mPIhGmD*Otps+Ci^n|6Y<)A+6{ARhdCio1G&XY7lE10x6Z
z<fXNX9^@_KzZl3ju^kQZa`)fuI`Li4k%$PI(O;0M1+P9qrBF4`^Lv}hlWtXyUqKyU
z2Qjnwf~RA45!b&M(Yi`Ih$#||d-xFo4s%Vfq_>#^C<O9s+CJ*??Avjd`@K_a@?quH
z$TF+BpYv;d;Y{8=u$6x@arOOC0%p5uRMC%1Bdyo^(nh1UQI1cU9F)a(%DWg%=`FB&
zx500%x8nZ2b>&)}Z5E9Z=SC820=@n<+nc~SqEYko;+9_$a6=xU?2Z)i?dYSt^{pnJ
zh9R$aqxO5D#RjHV+7YD+_bJoF3r?11#UHB3TDE@N3ppOnBL`9!=yXgiVQe~zl;r6h
zzA<c^@`D2>KYq*S>5+vCP|KdH$JDA8XD7@{e=>gpZgGRtLmvHKGTn<8Fu!+jT}zQg
zl(Vyqit(-dC2jDt33MpSVZ35hrKOS9E^ilF3}-kf6jcB*@%UY(Gdt;@U|hzBjj9=*
zCJX{XeXmNAN%8{R1W4E4zRjv8bm6^ba9NKN*&jwcsWZZk>acpU>EK-CGOc>IQ_&ak
zv--MiGPHZsUcSv&MLyX1{)<uAN)^Pn`g)zW(LqWT*5GESli5kVc<eLxvE~)eo92~T
z_<Oc(wZ0iJe(|oM%oXIBe~5=je{z~J^uY}TBpCHOHFB`z*A=;s7ab?ES}g+Mw*3nv
z)7OO~q8}ZO-I?4Sc{}Z)^@gXpFzmI0)ZvB+x-S9z)?Yw)QJey7?(V8pnC{UI6x`LL
z0Y?{cA4Fo8utj5mrgZVM!$kArV8q6{zb+MZuuC`rZtWiHql?>r&}*+gHla!6*n$B2
zhv({^)FVZ2oDd6Ey&#D}<dD&NzI5+*4K%$(Ub)IkYo``dyc29K3cNSj+aDAc!tXL}
z@)<eXOoUA!XbR-+=hWt16Y8I2>AHx<d>#vsVl2sX5rmh?9Rnup$vB)CCs}Xm{8_=-
z3l{=I)b7P7DZB$FO@?9&<bPs03#qaMU-hj&;<W=#ajrYrs~H*9q9&bl#eJ%2s6g*y
zt#Re(gAygVNj;DX02Wm4B;C)QUZ3<I6zauHL<yyt-@b*utOA7z)90O)i>Zr8rDgrW
zgt>ZQJt4*Zz7!&w<?=s*m1S1yCkG!t6wic@M(0rx3uhZ9v_7WZTMHvDzlaDTdT>qz
zNu-NDajJJ-`swuPWk~In_25hg2emhiXnFCOv=jvo{Ac-*w}91VYpS+xMujxQkhx5`
zZ#Y(LIzR1ao|VlICefj`Wc*M4VX(p~GN2idv0o>5(J<8cq}fP`8OdHjHfZ(-NpEFa
z-kgv)XEiF+E^C-1s{)M1^1XTT#N2*j32K4ZEjobdm`KV{Yn7UPCh8Sz{i=Ny;#SiC
z|2+zbj>>yvq*ih%-e8)X*DMUb#B0Z|0k#D2BUo=&02dqBUP2W};gMpEcpemFR#CcC
zTqj<?`5|o90sOGubo0cI*ta%+td7I#s&8W`T*Kg8z0J%Qe=OHzkf7?c4mFEtN)jw7
zj;Q5Noiu6P${!xWk)ann^)o<7N|S%x<TJ->bU3<pg?`+C-CxDgFSRvolw!gvoe0^H
zZel@yp?$R_Fh+cCp@tAFWTKxFE;{peR009wlXNf>*@Y<p1E5uC!R^K4GKy?V2S^nE
zgpD5FW#yWo%bZ{}^iIi1e{_0vI5S6l0hm}Yw^*AJvubuQCe+68P7agU`qEI?x*?hD
z*W>lKri%+pA(%ta&KDum$y3ug!gdQ)AU}Ar=F(ybw&v_{PBp3W*x^{dn_J_t!A+B~
zZ<TtL?b`rtV{bSW(`^LGncdm|ejR3TB7M<3gY^Xturb&&yFXm`$wT&+=tOFa9OaY&
zSB(4GlXOudWvbN)g$3&v4j!jj!V8-ffijE)3`c=qS%SI@_(~;AIt5a}c2bJq@ksTj
zbek)-lA2uB^4iA}e7NKN1;IK!315iZjU|rdFkzr@N?fQ40OPt#Ej$!<GU{-Qop@ab
z%OZQae*SYgVc@d87RYaF&ntBM48T{bT9g}V0h27`ZX9CnU^am06dR4KP@VV?S?j;f
zZkqhcjjx3qu<{M6v$JLnjLAH56Cs94syUdB8;EZ>ruOiW%NHZ>dU#RvKV@xfM04`p
zm%2-JAq;4Fbs0Iz6(peOe#$l%1>Lg|djZ;fmk5*gO5T35srbMg&BDwtfUr*>fvi9Y
zt(SN6?{<^=Xd&9yQ9twIM6b9;n2a{^iqw-3MYDY-s;W*n<to;R918>J5vM$5^mk1=
zhn`+(C5nE1p4Us=MU2HqE=@BB1>|_Zs30GD7`lqP3_bnaQsZDU^1;6TYrcEuFa7YK
z&nVc@o=)JTf>2cTTqsa3Xk~;}U6kJ$8PCB%w)<Qc8vdf%C`i?|ioC!plqSAaV0B3#
zBJzaUrG~!1JD*M4C8L6;TxAoh7m|Uui-s@v`M-HZ>9Tq)-H3jB0NP1%lYJ6yQRQw>
zWoP;@CG+KukM?D959@D%`aU^WmIZaG>dIYPU)r35;9yLfQJ`t~HNVRFTL^xChT|5f
zD%5!93vDGQi-)|xME5IQz(!WQ&0LF>T%ub2cj*QVQ1Wy|3SM{*U)czPnlcng*1Mma
z{-U?;$L%S_OCLY5txI}&$b9aJp=is9cTrc0rP-MGI9WB@kc&y~osToVbIr4aUTsb%
zW|NF#NPv*G@V`+QXvA3~UE(a0sN$ThI?`1*))?38Gu{Vph5r2dy<M(Ag|tS8e7t<d
zzw2SGY=#7i)JjekDsA_)IO6c!46Yqlj=k;fRDZXVJHLjWdq3$6)fC4Ot^a@)Ckf%D
zxF|;l1Ox)RA#%&&C3q>SoBr{_@BI<0o!Q5NsrUu0I=Gc~g^k{IaOPrtJVfX6r+fWR
zGp=zti;0-gAj2tkqW(-ve;3%9J#~?8SPLx>0Y;+XPntTth04{XD=+CsV0V)11cUM0
z47w<8RZqW_e*v_6wbC`{`D{st>~qaj(-EMVinia-4{tK`{W1+(1ZXThy?qLzy~r!a
ztK2t7CmU2KCzWX3ZR2h30SP}=+C8^Krgu;iY$s-zTL1t8!7r0v1rUlMAp^IfR)6=V
zw5qO<dG@PV9Ep@YP%(6z$|f~g2SSo6`n&msGy%J8i;Im_=o7~uPl%?RdYR2B(E@MY
zCCOw_)al20Lq_wx!D-omtFn?rMSA&GnhZWgPY<?&oNc1!7m8u5O<!y0ki(yGiOVNd
zRv25z+U)f8=er)@=mnPx($`)H5;IUzJ>H+1se;~7+u>7v<hj=ByEgdfLDe5XGnA4n
z3227&OugEj_BVI}{g5r#sCZa$y|{<LAoWrTP@?}f2JjuNJj8ODo0CQP<O(0AVGxT2
z3FaeKMFr1I6`S3c9FJbVpsNmR<&dKy2J$}gts~6^90dW9>a76TW}?^@N$rfX-;kUJ
zSjv-kwPOFr87|RXv3@2r@*7d^bY@J>)J{)x^PdcWtFr8q+;|)e#wmfGXgLjx1}Y$H
z4<+0aU?Ie0{Nn{pJs4=)!5_7ZwOJV?!5C83b8<K$OPM=|Rgw<Gbd4EnZQ;%({3kyG
zn>D?>t^aSn>&ADXXb-or!35JhJRaiVJ<*|vUvGTY5TvHb;CPx16l{Mj;x2j-A(u~K
z-_|n2d^2zHGM*Om%hZ<8ztwBH>Vy>7D&zhgxa>c>-*`Wq2~I<|<I5rqoYs<FOEN^q
zic|wp4KBt6w`p8I%U(Bn>GAt+C!<1#PU&^!#qUvi>)w*l_??D_e~DCis_)q#rK#zW
z5`+vKa4S>VYw@};k{so~dQMd4+<nTxOEl%yOTkO&{JfvqRMhcxt7;XXiJWj{54Vk`
zpVD_VBt1N9fYAqMGLW#t=h`&^d-;wBPunl~bH>y`oLCj>EdousAnxsk<yTTnO-6{(
z7qU(*i87+X_uwKUWBZ!-Q=?>j5)2Hb^2zi<IVOXtO1!RmYx*$-ha|9VBepn2yHS5a
zGFd24-2U^T0p)dLw_~U;QKZ#AVM!&t;o;-(FkLEcL67fT#8Co@ynJz&?@?f7^~uLa
z&a-+GPvitBs&J%qFgh3m-fRq9&w6-erSL3166e_r5v($C+o<Z^0J?g(rL|UTVZZJV
zma%h)bGGN-u1$NxFV}kgwxPC0k*wUJ$_?WVXMiqzpqk#Mu4fW`Euug4wimUwmJy=M
zU@qjor{`vkYuC@dg7Uq$tH{7e<u~MY5C1Rps!hnB<nn5=;0RrCf8T-OI>asH_Ta^B
zgbV`XO^g-MA3xEp1t_iJT;Pmbc*NmjpwBh|jLj@fZdhOS<wliyfP@Py227+MY43dr
z!+rm4F<b??8~JY^AI6!yx8g$pO~WA$IOaNee({o_&bsq($1$n4#1(?@`}QbUYVK0}
zf7>pN{mXWV|7SxNc)R=#ZVuJ(LT2e&Ysx@mTE}og5ny3q#gqdAML5!5_zv5i7bRRV
z1X9A?>h?w-0P=K5#AZ+#7~kXontbMj4p%60pg&o9iwD!}Iud@>vZ1;}8vZWo_mzUj
z-MWB^)KC!a&6VJ59%aXS(rje)TIZu7*RS<twXv_Wyxi~eqwS62ZO1dzk3Nze9qgCt
zy7rX<GQ`lAURlN$y^ZZaVwt6m%Ky1EoNUWKWNS~>wNzFT<8w1pLd}G&bVReEdfN9%
z2;rRJ@h>WnROos>%dJIq&2ltSFvoOgtal(87VH|MDH2(tPz(>dLGR?hsQ#$ty(#oc
z@QS5_;!Wg-5YdKSLBQ&##EK*6{Y9HfM0#F$Q0vumTLLB<Q_ddqVr;v)%9^NrYrfG1
zO^q=}9|=yBM6m1gSlj(8JeTYclmcr)$qRQsVcZ_detP?(yCE-CcAeFV_qc~jG~Gys
zymv|^gda7SPGx9uTyldIj{@8osvzw5E`Zhnq1OzIR5gsB-|ZC&oDjN$;YFos7A~4T
z*FlRW`i4!NxUd-SeMsObcwd<4yPB3H<EqX1{kWwOuxGeO%#;Zp#pYec=F@hFJA#(k
z@=FNnC|(?tm8A4@Q4PBujPk32<4wCuFMIM{@bxk+=bhyMJ(JG&9vkjVWzwGT#b5#{
z-Jr-;;-`h>ru==2x+N(CN4oh_W*zXluSOowZVN>OU{8prLBem>iYAL*8Rq{kIqd`B
zrN2#DUCxx#`$K$;5r^68Ik+{qR++08#Q+m9v<tM}5BqHt`_rb8S4UqJ8rld?g=@L<
z-3$NezIUoS+@+D@^pgKegLp_et<(=JGoXc>-)9Gi%L(_or-|RCk2Q<o@SN04;rn@@
z<*l3*9vd0JN;ZYo+4zIZ?Hn8}`<{a6Mai&74ro#CVt!y_Non{$#16SwsaYqYy$%I}
zD%Pd(cqH+Llrzvji~rFyat=K9^H=(9{MU0L)+N@^jm6VA3(<N>uJG8XO4+mGl1A=W
zz<l9sW7^KrNbi2fwwbl{f_g|ioYfA*?lqwEyMo_9{2eAfiwBp$<HPr1wMJx!>OFf{
zjthtSYhK5vw?C3F6DBZOy3IS5=U@%@14ivk=u7U>@xzMCOzcv<w3h95`*UA=!Bv{6
zCcGbPB39B+(ojwNzL&4<OpgtWBBeGNhJNjWa?Idgt{usgDeW)Ut<E299DduZ>*dO<
zmKdf2%kBD4H-CWbcdZaQ)1ZC!<G>FC2riQ_{?Cv20s<RMlD07?C+FT#pUFghff{k9
zz5Du4V`t>v#?St`b~=lNuJwp<iM?QAx~hW?ZG`{H0mkT;30<n2gWP^<?}Xjr!U<rC
zZ#q4YS6p0N%CI$D*R<#9f>TQbONmm^=lCy`9l7eJ_B<e7R;B{MY<ugf?N9enJc`ij
zv+32cwAm+UZkIvPBqz7QI?Rp#wL-=K87mUrEHhfCuDYaD5$O^_hIRDJ+qV$Ebe@Id
z>rPpyl;89lhN;8D4O+m+;wmcFxnV9ZB(<sxW=aKBFDlz~_u@#$x5knmKT9br9sy)p
z8%*zf-QO6ONZR`zPMN@M@y>6oS!1i9vhce7=CO=wl`MwO(RKGAq*WpR`-w3=pSY}!
zNpe4lg>1PuxM51{4}GLYJdG75Evp<Bc}7Ydu|>!FEPH$vbLm_DLm6y@p-890gwe%q
zeB_9exvqeVm?uZB*;!q_-e~Id%KJ0Shw7-~o-^!5hl%mm9xxb~WY{S_(qc&Pf=Piw
z;ucb@@yF`(@GmImFzlO!0@6!Fw`2(Z<0J3|u%?38$G>*qaIq^vvSs$R&)W=VEQScx
zw2D`p>oy`yBk_1>Q_ZLx>fN(uZ|t=5EJZV^7ArnE15qv~ZK9npN$D$oqT|VtVm9=G
z3+Aep6)i#@Z8=+^O-4r6xpe^iB+R^FhEwk$aWFCS|BRu(Ar)U1fQ7roZ(tG+TRw(_
zR6DiG83Ikk%9*V{dDf~z#^}-|3->eL{4K?0SKOa#BWU!B8+)4&v~ic5$JKa+jy3Qf
z9(91s^-EwL#f<EsyI|H^Bt)(|!!#OD(~_tIztO;^udMki{0+$^iFdzLJ&~?dM{m##
z%j0Rgk8Eeyb~Oq<VPms80dXr*s6^96;2+|urSjGs8Q%Q1h1IK9G7Ymcd7GOKu3Ea0
zhBni*H?-KFW3_{^?am_0wY04DW8=KzZX$7-9??Z=5^p9|K}nt_wx<?D`@R+6A>svR
zCP5beND$Qzn3%6<sS?BZ*Ceyato{9Y+Z|P7hc;weM--^I%7npN7%j$Eo^N*lKi|W@
z&+yT8-mFY^o#n&MPL(|z&eGBGIRb&Owz7&DcO1DNqbyHJNjU)PPT3PI;T!kTWk{{T
zYIw1(x*nv$+aA`GwZX^sDWKtH-H+(Ko6~EyV7tq<Hy7SciHV5~PfQf|^=Wz`X4*C>
zS+tB7q(j`-)p^%lAGcStGqBPs{AWM`@#R=baOWtBjEr9HAO88T2z(_QAcUiOqRjLU
w0r9UyGiPv#6-Q4;EB5i9w)gL~miI0GJ-utxi`AV%1i+t)f~I`=vsdr`3;Gfh`v3p{

literal 0
HcmV?d00001

diff --git a/doc/src/week46/DataFiles/ensembleoverview.png b/doc/src/week46/DataFiles/ensembleoverview.png
new file mode 100644
index 0000000000000000000000000000000000000000..dce581ee65b3cb552589a0ad6361ae775a6183ed
GIT binary patch
literal 81932
zcmcG0WmH?;vp2Pe3Q|J~1xj1ANGa|ul;Tj_NpVRbcz_U~XmJX~t$1;F5-3HAI{|__
zL4t$?3%o#|yY9Mo{n!2WemY^!IeYf(y=Ue(atQjUBu#vm>Mj8R0kN#idsPC0o2LW>
zHzaOfzo0+`q*w_Eya{CAzf*Ue+?)n$>ua=h?B3vRCXhtKnEc+lm$bgDj@MpD{HcPd
zuQmX0ZEWZwoSmIVo!NotvNm!PMFn)s+lLZ&{AM5tV(1&ciMU@kM-n77<Mw;#xNDLj
z+X;~Qvz4jmqZ<d83<<pF|9oS-{6TQv{o3W_8Rm-qcJ0>XU)}@vE-wV>6jv7s&dbYx
zBog=;UOjRfc6Ir8wEsx_k1YNp@qhCDUo8Gk=6@0YqZI#<_#dVCkHr6z#s8G||B~YW
zTGIc;;{S(e|Ha~eY~_Cu|DzQDk@z2__>V+_|4WMhF9+lQ5{-z5Me*`v5>;oQ8bdRn
z6K2xZiLBqPZxQ==I@;NNt3T|z@yMrxF;MvqE+k#{vyC9kQ2Ev68@^eM@WVjKG_x4<
zfyb0@GbN%plS`lfc@>)TcegDR+~;RFU0~T``IdKy1ek*M(dJX%A9BU~*L}ua&vLv5
zy}(TSvw!Lf%x9Ihw`dSoYIdXKV(1V95K6X@Qvdj3j3;YPy4?C;bKAWajGLSFz;;z~
z1)tF?%Wc5>0M;Uh>0hIp_OKthm+DCQ$$zv?Cwsl}M!*M6D|e%VpxaH_i9S~vOXQDf
zfhCWJrhiwdjT=BdQUd*R@y0*eE>B3ATOzwy7khkY2Zt|po2WFu{f)ajS7nmEVA$V%
z4EU;=`QyRs@9Li&L>DpU7Y`8#G}xQJ<6K+6Gz$L+eOZ%z_E$A2;bn1*1e-T=`EOi4
z7QpU)X+^xRwcm>M#{ZpW^XvKN&ihXGnKYlmQ!j1Ar{=+7;h!xf`q=)?y=pc!Crld!
z>DC*f6rueXGA>UWv@Z5+KWVkVlIcosy&pVs!!x#X61g}($`Af6DE=4-N#<bNY_T^K
zlla*!$N#jjsJj23SYhXz2?S}M)PTZI2=6j!S$MksJ4L}Vg!oFe#&8rdda-RcE@Bn)
z+UT&&u_O%GbX%1=L%eq0@#HfgY&Wbra7;7G?cJ5-`=GE=^<xyPDe13!I}SVbQ!S0A
zQw3Bs+S&Mb(|u!GFfIU57`q!jwD7C%7FSJ;94ZL*_vQ<}SkftViyT{v+Ni%2KRHNo
zq$j169IWb{pV41?^v118tys_LP>|)?Abi_%yH7QJl!u>U<?+^pVXTp}HATM;x8#d#
z2@T!7;co%t^zWSx&4!<=o*HCJSfZ>Wa+f_VC~Z^Z9Z+i{0%^}Eg6>bYw?v61Rq*rP
zhApWhM1&4AWL0F(wrOqN$slz#oe$rhfsT4bHJjU@dp)%Y0e3Cc!<>n4N$(-(tmf=x
zt~Bz~dkZ=Igq|c+Ia>B%Y;U8pruuhzgkV%h1pHIo-cll9dtz+Bd8&t{hW*~RJM`b*
z<r9MQw(EC}E%_^m9!7~USh?w7Px?1?kVVh8@0p*O=zNYp`f<3EcaO2;)Ok25r@>~z
zQxWubJdD2|{;XM6n_R%Y&u@b@1b%{EA*~;sK5sdS;3Od}e=6Yop=&QT|A6!C)SZF)
zSf}N$oBBsc4YjqEjZSofc*yoHf>!OC$S6c%&2kI84newbOtz^kVn|)rtmAN?JFWiI
zJfs!)IN5F*|H#5Qfy?>KtNx{-2^Oo@xby3d!>gFeuN$%1Z7;>JAFUCs$b`V(aiR``
z(`T_2Y=#))npWhmUa!8J{0%3;6I<>Q%9qaX^*{mw{wjNL46Fa!I!wxie|MK}*}xg9
zG(YoV*^QlQb*FS<*myC_q=n-tKLfrQzU^=AA1P4Zuuha>8+;l~cvqvibMO%~l!fQz
zIx2{;n(lS>Ky4r=2@aJEV>$oyi)&ttk%4uxt&5asOsm?l7HtDoP~wB;vrm-K10C%{
zzJ|lsPO~0MetB4;Q=>3_Ffy|hU}Z}0RsULMjLWy!V*`$UlOLQS8kHNN6S;}D=$L(T
zWyhQUgu={ugA*_j=ki0^3GFJkQ?$H8MklN{bRV}Axgmb>i}g+dsbD8dH9!Yj&y}fr
zMBM)GdvGG9HAuG>iRXTrOfSPF^qN+~=+505q4PkO+$r~o#UR@NSszBJBRT;q0kd&@
z<5Ez{lY>@B1p2Ue@poSyKYJbLgrv135KN*k_F~kD*ly)a8Xz8P>zbxj0bbngrh{Qv
z=_H4gw8<;LdWI(uQc}R;)83}ShONM3Yz=SVu(IdE0;KQs*{VUWso|AJdAn96&%xn&
zjRSA+Em#?0c$7CP%Q`czUp1$Hj`dVY2n-zH*G|f3-yQqxei|6p(@yIN?#EiFT7a`=
z<{sNSNFs%2U#W?NMl+Mb>lcxIf;^C!V{EF1d|KK9LlrOPpDB^v8WUr)qqq_};m=$e
z`xJCp{q?UH>P7cE`Xl!Gxo?7oLD%cMrrbmJVtA=N^Nj9^C)p`posOMv2oivA&GXUT
zu0DnFVTr0E(CLmUY*|}saOJs8hbtHi`2E%keOv7%rw~?A3eiU9QH#|X*3Sqe;~~c2
zUdQB&H=ZnR$B5f6QK*r`37<F-#Zf9Zx`e`QcY9N@$0?~RIAMu_zn1&MW)+}HDhaDu
zC7`o$uDUVBn3MI7%(3<Zinu?o`<X%fq!#q*tXy0p1!!jkkEY%e4b9vK^RXfO2X7Ac
z`Etp!bY69L>5D>d1sQ#7r!Nm$1c9wl55J5^h)q-=Lt~cL*)tY<fHgnvrD^jD@BgW<
zk{0JwEw_1e1r*0?_Ix4`DQoLrgs-Ist)wHOOs5!zOH!Ub*jj(N1wI8=9^k&bPV?>~
zojg0u|K$8*!r?{+IcI6vBr#nR*mYrRaUm=X_OQb+MgW^Xxwj?0?dc4b<kov|<-{rL
zHctg%szDKkfWhA~TKflU`VU|`ckA{`LdY8qV0-Q`-w7tdZWUPn1Y97d&*@!JNWuGA
zM1`x+s|jZvlIef2QkgcR)J?FSN=T~Bad)qp8nlPjY3UKqNxoUdDvjIG#ElK4m~d&?
zl{DTD65RVIsn(0U9GaGF_gf2!``8pXf3g&7<&EfaS)mSs-c$%;#=3`*wF0ob%(}XF
z6KQfTI<fHXx4r2!{hzHL_lXwOSF=^rH+#HQ*}fOC0*|D`1v+;0h2Y+ZQfd1|;1{<J
zbV^pp?O52a1blXoI?Bsq_(a&fEW-Yy<UED4?eW0#un&f@l*?w4o$h+%hz!ezEs?Jc
zM5d>tkR2SlHq2uJA1$bck^zk;ei^guF$pK`=Gz;oaCyzXI*y5g7gJ11^<vL(WpdS+
zifiD%4qxo!B(4a0z^7zIMTs)$YLrT|IsQ1f0rj<cIy$^?msL?mm5<T=GfqceLi6RH
z!tYkjItgt#eR?H#b~Bahu5AFv@aP6%@EEYCf(Hy@yaFw)h9^Dm@kt6r8epQ2{9{fj
zX2+G!IFUX@iXjEe!ehp#Porv|S9i0)3(l%7%j9QS8cbxQOupo)2?-qqXj3Q2qc|$w
zG>ViIce2Us4m|&h5*uXabNSPE>t%Aev=kF9s89=#3H+;bPDAry9}FJ}@a(9T<v?x$
z1aMHU-BnwN4@}c@@V98v%v?Cw;-yksK6mO|T%2l9pcdfR9&3n$9?GD?+SzCh0_C;N
zHxfs|p-nxTABI}maEbewjlyDUBO<W1&%bytgOz4)CIwiER!JOH;w*6&&qR9ii>8ox
zTUR<REPS<vQ&eANWh*XzZeteaovLM6Qe$Pm3?!wA7XU_aS?7q^^@b)MGbT*UDjHWQ
zK+)aD)~o#$aco>Sb%Q#xk~sB{L)z;4S_Lt+?rw*K?ylhPtaLQbF#aiAvZ(^w$I|Tv
zIywnjV^gfS(6HSDz8J&!eJD*zfyJG_K<dF;;|)K}hBj1);bBqvKJ7;i!(FvUEAbCY
z)~77!>u0DZL@0!ijkUp3(qiH&W_)i~HbxU%t9S{qv4{gI1bdvu{v@|vbpp7j-p!Bc
zeb&o>_P~pm?@Q=R59}6ampO)+Ub|+Zn!%L7WN6dN$(c=Kn)Li?mtVvff<Pq!?e0Zo
zE_ZE!oNBBJHNs>0Jpufl2aC=C3k$GtZVij@{sAG}71mDar4gH*SJkc-(h!C!7YOU}
zus>!;7DLcGU&w86Z7@d!&P&+1(W~OaplNqG>s}gswIGfpHqNRDY9J>yTDA!?JX~2n
z4qxw{J>Ny^-QACw&K(=05#Bq#G24_?+EfxhUCcz{u1k2aUz8ug9;L0c1q{SQ-|xg^
zWGD@vI5~h4n(nhZ2j49)d2v1WeXZMV`Vn@ibuvO?eA64NAGr!VJQ_}3FC!Jm*I3^B
zW=ISo7xGbsB8h)$1xUryHS#p1))@#D{aw2=8u~{O<&*a?@+8>XcGk)bz3egByo9gZ
zKi_xz0dV-5l-tlMb=%C%M$;AwpSE=_(QgPhi1>@S71v;qADSZrXXq?%7Sw(tyMznP
zn5l)^rPTUY7Mp;w`d<kBR223vIxYSd*K=>2_%A>wAkxG94GhAI4Vw-B+WG%}YiPw}
zcK^@wX3u{eKzw<93va`02F>44>-ChSQ(_IjSmqG3<=tjaJ9qBO8N7y1;|z_QmAox1
zrR3?0u8seDw657TE)A3fA_RowY-M0+V+5Oo7n8bBhjnXBK6t~CAU#H&(U9g*yo@g<
zS~^lID_5PFY?Q3mESZMnN@qU-#3r$-296r~1LOA?3$u8SUtqzjr|i6Mrr2VcNziVm
zF_|ZpRDZ#4Ln;2SpK!mrGkN-g+w5RE{`g(0wrTVnkkiww8Wbn<Jmg-u-W;?$+y22r
z`tM;@pc3fLS95QQ#!pykUtO`k80NLg#K*m(Vub&Ft={_K{uh5<N3LQfTX^@#FKo6t
z&mE?^koPhS*A|V9^<Qq;w?UjL719QIO68Skm-hbR`hZ(0N+;NMi|?qhJo|e&ZE2Fe
znC|c6kB3h(h3P!F9MkU8-bvqI>NVTvY}6Kb(r!6I&h6Hda@ZV7piIvrm{8(j5D6@r
z6U((C(;bvN`it&g>q6#!pLQ+6_kLW2cUD|J;WO}C5!i3wMvYAXw8!oN$i!t=h~B4z
zg-62tE30C!^*)v;S<K~w1jI1x8{6S=*609om8Z8#G?w&gX}w(p&7O#7L$}xFTqRNX
z;=XOu8Bd33aI4awJ~XSxSbYKlNI5SOnNXWlRRL#=?ycVRFN@a9(>NIDgDMTn|2wby
z85*OfeK}t12le+D4aw+4sv+AMLB~TGgH*HD8q<1()hw{FN&O5&-RIPt+<#Z-St}C~
zP;A8-K|%=~Y%M4XrQPh=__WJ1j?U02Q)_8FHGN3$_hC>y#-JRcHFGRfVbzG|NPO^g
z;ibBB)k#gPsWzu3pR<uuR8nhQ_U=G{56s{Lqo1K5JEP#o?Fn-F$|8msLfleux2^RW
zeC!QcDsP@kNkT&(bMWQ(A9_Hg@oC*R2d#7)X6DHwM-?~Yw;!Y5yqp{t=5UTkc~CDS
zV<z!Dk+sH@hwqQbN*{@`Q#2*8fZ|O=-lOY2>YA(8dh2_V3=scDwAJI|>apm~hFB16
zMdIaqLw;d*3GV{++ln^5gEw^2o&<mkOO`3V-FI8A954mcC8-x^{zQeahHJ2Dm@jUS
zl6}_Lf&7Vd9N4gOzR-#vbS}%ET|vW5`pNoqa7+wA|G@U6sZ(rAdR`EChsH~%mQCHf
z>Xp1bW}tv`6JT*@cNUcvM+nw$t6!-I@hW#|TIh~F^LY2^ga=jG1huPK|Gqf)Q~cci
zZ0If1s|xkU9tR}=KwiaD*R!LI*ubh~1+!mzUj0<~wTwFRgW;?gt9HU}cm3@8wDa1P
z33Zb@5)cd;jo;1ENf7<$ftz@EkPOe&L7J*asvS)OJy$iC0h~vYg%CtFyWrVaCQ=iO
zuX}d5(v;fpW+V?+Qc!W-cfOaJzZ9gqEc56>K7#53Tefqjj#m<2$+$rkiLQa(a6Fmd
z%7hntU14A=lDsvFaSDIBkVO0_4)L8Tu-hU$wSb>|=qQRK;i)B?+-Ux55@i|B`A(^^
zM<w^4zuqg)jT0v~ov0}qVkH7ypbK5bnd*-PL%3sgecXM{iidu>CLR{u6;8Z-)u~l`
z>_FP;y|x1MTB@u6{q#(byT{~n*aSs7Y5jQ}1C((FD=fADc8>eYaMERxQg0no59Sjs
z_*kV7oFEJwse`rVzgs<DZR$&$I#Mxk#<cdCj8>)2WVUzo-R8l9-9Hq2Se~YPWo?N@
z!eN3|+4GgN1AhBZENII?3TW5U@^<mWXA>mUemo1tVeE6Ptb=RLJe(wrLcWH;Fd5%>
zIRf6#i{g~IDMTwBVQ@Wy7@^K=-LOa*{{bGJG1;lpX@<R(1sFr${M{m+Kb$QFLn=o;
zs1NhZe(!bTa2g5sQz&HRF#%XMwl3yxw{+49*hjyoeY|b-o5!Gu%H+rZ$nmWK=x_Pz
zz5R0#=u4N4{-!Cwq>xy#&a*xBpg~BC;oNG_jU$JV6?u$LcQ#9Go_JJ}LbC#5t@!aw
zkv`v%SAR?KDr>PhPD-e?8cVn9f&=ib7SH5RW8`m(YV|fXJ2B7dS7Xc2fY<)x{Vrek
zDLg&h6r>k|X}oMk(nXo5#YIPI9jc{Ip5d12W{%NpYUtC}!S)(wWW*+yUxkJvc|xp%
z6VCTsY1YIwO`~qPBeTiNrL79SMR!9QAZFcBqN5od@SgSY1gPlfckx)YH6i5Y82?Ho
ztu^F4S7SC57?IPP4G`b$M$J%jzrd%5V&E&~W>nPh!s^gsGS#4gk*#BN4XGJr>QaOA
z<h0fqciy7){G3(UuLcLCa6g_?=FM5_KqPJjPc1|qcd&@Fs+egKWy!!tFC2bp^3ZG}
zaWenw@IFO12KUgBKo{obT%jnaEyo_W4%s=ig*)}|Pvi4p^B}}rE-{$KlE>3dPWK_j
z^Y_0&9Tj1|OWf^j)je+AOw^D--q8j=O4`Lor*R3MH1aC3P;0qH8md{y#wWuLh-^zn
zwUo97_E%g$D>C+OH<e>(#O=GA$?q48_Y;wC5P315Cukq0=bdf+;tUI_=}pxFf98eW
zCfQmuCnX8x@qDC64c5E4c>@A3Sa*FDa#z8K^e^JV%BR~xA&cBXDK7S1dxP|g{h!S~
z_%kK7Oa)bvb`>h99*@pv=8(>B-Ptm2=j6@Nr*yV0of!Fb=blFxy5bGOi3k(Q!y#2>
zgruI1`xb-sU7zW7u<$C-5Zyt}Rx=ld3s&{n;RzInDQsrEn$93`)J`l)#hkCjh&l~y
z^#NgA<OTP_!F94RAM&0I$jv|YYWZ>JLfO2<p!JI}`Rzk8=Z9u$#ne$Fqe<hD1dGTV
zOrujl$Sd|K1e;J_yK~%0jbf%wMbkoiTL7&AuB@W5jmjiHDNUeD05G3pQWtQ&<HP$W
zgk3ob@Yy^S?U{U8c99%+EPM=AS)#O*>ZvqJ>TtbUb3cIf)o*#!$a7U=!MN6OR;9Ae
znT6Um$Bc}|0K0;}ph-aLMEUbX3~WY+wSoTNT3)7l^7=@@7$xL>RZLr$s6v(stN&W<
zO^r-`bt|=%Z*5<qg&jWcNdc=pXH&;wmHATI_3Le#^7hOv{QqfBZW_l{@##U6ifzX>
z8mi;j8JyC(u~zP^vlzo6t==$(IL#GSDOE^Z?dH@iQqVaH-5!ps)gw3fyvd6AKC%?;
ze&-T4d5h-T6rWlmz5MPsCE3fDg8gi`i8bIVh{w3f!m(tbq8%G_?1;5S0Ba1*dA8|{
z>2Jr_>zNe)fDH5)->Vz1*7ZJFpH0(U=A9%Tw1A7Wkq@UC3U4SoUq_l3p5pr2v-7sV
zx1*0!_D6^p!_Dc6A9%{zFDil|<Y9VnHsB|%zh}rrx&+Zwu?{H8&#_nGUS%}21Rq5E
zheZXyDNy!`VK{oC+!`YwOdWVup1Bqg7g#gvFV8153h>ZnSYe>-w;3}p64x>29on8f
zT{`|}e#&Htz4K?+lRV^~N*VqN<%j24t&VI=%zAiybkU~#RmQNH1n+UUzd7_fwHS`*
z!^?|j>^(zt$r=lV;z%OzK~qKtUd1K6$&)k7A8N<=#s&8CDsZ{=)D_gYEAgZv^}Y5G
ziz^J4is;0>yVXy%l2N3DxuU&KYI6`Py1mH|Kw)cP+e>ose@4FBx}E1T4fAfw{$+di
zSlz5-`aJzUNmH6vg#HT4CgTD5*vMT;P*J2Jr6{`FK(AhzThOd}ouq#`8=|0L&3-B-
zFA7#aNvMH82(e%kd;+S@5{$=YyY+J4RSwn$NmizQkKg<K3!|oUvaLA($>zt5@pTw`
zSYB0n3MBplA06P2A0nA<^PE+Gv_Jep2;K@JiKkc~n`3*g<~SdeFIr{olF1z*O@8U3
zG?{sWYxu-6N3F;i0a1^|CW=w9sSp_}4>}=iPJ+^OjvWlu^?sN;sLRDXAq~=K;pnN+
zp@(DP@8=V{*x5CNT*w<a>bnZ^K`C|{YhUtV@RE;tH7qI(f1t`1iNggAvL3a{9gmW3
zNKd;O$#}Re$EFAZ`Bn=j0CEde6n}#VO%M3~LGW?L>)Cku+?m$|^Om19ip54so$Axk
zIuGslW6mAVwr6^sX=>#5?!eDd?-g@r!;T*T`3u-l72MF(syp%qY$1hz@Y$8}Rq|=G
zs?cA?zdQStWO0S3h;w)oxT!7EQj#eGzUQ;+xO_lOdvavve-Kc*J2h|kytHjAM)4W|
z#gB71=Wi;szCU#M{Zms$#&i`vJ_k*OI$4Y0^XL&X9OapqwhP1j?2d!G)ft6qW<H=Q
zQp!|e$p==D(i-4x&Wx}uf5Qz<l9Z`3{Gqba!&Pbuk!LXi2f}+K$Gu7NGqcr3d;a%?
z!&l5QiyIaiVtUP%d#&p%yh{|u1k?G@DFE?p-IOj6H<^4#cMl=&mcmIkq)L3^bg{8p
zXrXMxv5Q8_#@d#dt+;4VMjs}6<Ax|qn01beQHFC*>Rh#?h}~R8qVL;YP8PG|YCg_C
z4$9k*-U%zB3I>S_o+o`k<_flrAYpw?wuk+p8??bO+TqbM*%GWVw<(zF8XCMz4|}A@
zs8f%RJ5^nZ{tLt0b7mi<vxTbMwx6h^mV{<(RYVo}D~naxq#Y4Q=}-B-xbcGRVTlmB
zH=?9w?00;{ylZ-!cfbvCJ?IQ*c|QU)p20&sxBY~R)!IY#@aIRh1u_Zdd=%tou6%#N
zr$x{m1Gz&ZT)+9jH+N@s$#m2z4|HsxtzAf7wAV}YV?G&G(TB_}p%}<q3n4k6zt8!@
z-;mO2=pOzgDGj8_6B-s@|7o)gt3P`B0(X|Tgc4_KbZSiVLfB8%7n;n%;rQuE+7(6R
zlRV?o-Y@MD(xhG!!`#ktj<j`E9_uOhifQQ{PKk;zlWEi)B{d<>%8z{1bxkM{`7y91
z7qvGdE{d73n2#rWi`0#DG6rjAGj;jJKP|U9ctm4nEqWE5$9@tyli+_dfCF6Eq{_Bd
zg=hQFjUJPu&Qm_JHHL}Frk4CAU-JeRraiD_^s7G*8&;RKspdHZh8O74#1iKwT7K%R
ztd*wrYU%jAmu^T-UQ+rb_hvUS;*hmFKZ2&Dx7vu0lP&=Q@qU$8wVohkIC$-mMB{a2
z+E~G7YgW#@NrTDXp~dANLG-A{3`J1&XCJ?==t{|+%UhO&HfZT7qLj1r0JAeW>v^YQ
zaY)0}U|lK5tV=UswsF%#GKXDY=T9btADlMdVkUmh(K_%%AJ@i2aXuiRt@o?JYrN+S
z2tB{@uLQmsiH~NtA<DPyryCoT1$Rk0d10|53<=o`3Ha0^cuo}@Fubbqxv!wqYAALr
zDt5TwEr6j?t3CEEuHu3<&~P=fS;pi?KGMam;s6b+kZi~(eK)WyfT_*7EE8HPhf|Fr
zsnZ56NB5&+bKWcEz8l~VThDtx6S8734fRl1i0*C)L!9PQvs#sLf4M5r{pgK@%&--S
zKH0p;lvXBL5sznJmeS3+;P|KZ58~~*Au4fZ%eiJ+6jheVcxcArbG7H9%Q->AF0j7o
zq$_vWobb)C)?$CK>Y;S*g|^PUbk2GADS8Yn)+ewUyYw|YqVFv81N&1mo0ce(b0|G`
z(myY*5+4?CC4pOndS~o@wWcWai_-<Arnk1_7w&u3-C1@5#@{2tKV)F;?6WeN*<|Z=
zwjuG5^D3VzASpJ!Ct(6KFkWkwVu>51S}SSXTs@1$Cl_oSs{G3Iof4g5aA3;&_^RL)
zDrB#VFXpEe{gqv@Tlp-qXY*m67M!q(Rzo}AC*irRqzDA+>KamQeG06FTxL-|6h}=)
z1Ge1sZrk@Ji$fZY2A?(WNM2QhJ}`%A;<4>s`BnK7nf+<^MS;&kR-z35b`m7(0MO9f
zT1L)+SDV2JH6))q-Bme1@e9D7TnzeRux#N5^&9v%0XdNV&W{jvzEXcjo7Vx9Nl&k2
zS{uo-bOr4@O02~x!5O>6jhef`_72(KxW11sx&K<fED=-Mi%lI%K;;!B?7G|*iz!3*
zZiq@MYc3D;QNH<-n$C0mY9PxB-u=MJXkmktWkl}G>fCE)CVxh01yP+CMKzN@S??7x
zAGo*t9q(P&MhJZTAmIK2O3(`vLr5~Dn$ynyBaiEQPHXniZPE0r((?;{&0Mz&sLC2<
z{Gla-rEH%N^v?dBiR$y5T#C({Uap&}&vQF&g{9N{TQ&O-48p+xA7i%A56T_JXnbLX
z1uIkQY|iPuT;s{v99q3~(Joi1luudRtfHsDV=8IF4r=l#PU=&=>41~=sq(dXtOa&&
zfmfT8zLh<MmIN1OM2Nn1Rd)V#2vp&NRjlj43~Nk>c+0*>Iz3;*9Nt<uWjd=lV~P3j
zOUwJ(Dq9TRXz;ws#{4fIzmp57hDt0j_Ko65c3u9+0d_wXG=g)beTE$E$QGm9-cnKU
z;`0T)vLX=VB<`?O&5PA;eqH@yBaXECRBvSoyJJ>z47r*1O=u~WZaTVxE7K07WAIIN
zx+~7-Oq!oCAsQH|rR~r<Vzqm3!6(Y{A@x<g-Fv>UJRMV!QgVl0vk$(gz%YY75oG$N
z>doy#t#bh!K8+8pSE4_=f5@{<c^D`3!jqo(`rk76{Njl(-KvJ^ao3HK#9nTLe{+7J
z7wqxt1Gyw5;g@^6!j}8)-#ZG-BkY$W<swqV&w~Bt9#h@?o38rZ!Gt{5o7Sc!du_d*
z1+*modsd0)AP%0$=>4P$b(cjS!lyJ0Jh%UC2na&u1dj?EK9DGkchmb}LHS>EGLpze
zz#7&kG&S~H*Cr~Y)1-a>HV}O}Jlq|~y+~rsY@ZdZoF$AAT#eMc^G)kp-S%9Egutoi
z8El(_mgn~0*#_Lnu0Kdo>_<6Gwk^go1t|S^M?fiKb)5j4mr-}{$?70zGN3C~)Bi&O
zV)$igjIVE`EW<V1;xe(+N0O)4rwy#44Ihv4*<73+U41oG#8k%6+Qh^FTD0$1uMQPz
zEA{b;u!XuzZN#a<=+baXi6-B7zmqltk!73*CxOtp!fq3>H!M#p-?NKFL|s==?bBO@
z+q=TzVlCZm2{~&SIkIK$U~F}vDs1i(!3`%<mZvMBck3M7eM<ECT0_JlYYxJQl?7!D
z&d^bDzOWZZ9GBfx<9Usf?GYpv`~|!`xBiK2h;|yS^`oTY*x7I@%)-u!(^*?;3_hl`
zOD*mLr|}yp{OpfpFRQLYb(Yv(9B4zqlfEMpj)hm3EBC1>2v_8rpiAQw_j>1LI0?U^
z!b@D@Lv%C{XH_Ybc-{9stNkf*h($6Qa^Q{+5f^i9d}*V<6pHMJhPD%44v>qA$&Yg}
z8|Vo6wYdEf2K?}5RZxqcW&7O2BGX6AkcfjcVZTDF2Sa_jpX{Y{a=HOSyVf6{*&G3B
z-*7hAME<X)8NJ)bbkk4uvpC>;&3<!bo%3ofcw@_l5}e)@J-JWn2)k>;DSUy(rL$d2
zwGU=)bCxL2MrpLPd(*!LsjtydWD}vD-_3VOJ=*TqO~FR2iVO021VgJO-JrXqfnwdP
zy`Q3+38=`TJm&q#-WeM4NTpqx;S+&pmbMJL#<!oc6&T)sM8)HSUe#AeRx%X$(KT}M
zRGY@=3kjO<Y_Ap$G0)B%Di<kLeA;AiNpc;&hx07kV6TdGeD%7yTyA)uHT>MRU6JL1
znpN10EG;DT6&*pc8}e?~9xBn2km#CL-pd&Tu(kMXvNBp)9ymetng=6MD<~Xx4@aG~
zCqBK48eG7qiw)MWdB`j<m-V-vbYiuf-HWGw{TNuUTK7NkX>3_TZ_!-j(>EoQY+i&B
z@Kw`Cs_6InD^e~fDt-0)#VokB)K`=sF2kV2&*3^CYaD?<#UUuX$~F{f?*zQdIX6Zm
zUD)rX%ZfKdYhtSF-L-(vCe)n;0Ci*!pSMwI=V;A{N9}<WeXH59Ezy2hb-y#<s#(aC
zy7`luRm=uSF`WLu@Uk7Kxwy)@lIQyPc);5Vx!<G;7gMlp?jE@SdU~lK)e$CtWy9B4
z;}iYe7SvSXkP{yPQCQ&+rUmurfoB=dt#k@+{DOx((R4dsxmH@oClkt3w{!Nzq)e%k
z<u@g>{Y{8>0^_T}xfI*Wt>#48p)D}K%Sk)jCcyG09|QyU;4>%T_ZKGJz>xamaU0Q!
zZr8UD0@O8ij&B*?+PXZN{e%logKtQdm)Psx>NB@4LKTIbk6eJ2cD5%Jg43x5x})0d
zO?*UKj9N-Bu1sOaoIpd-1YXl>tMe23Yj^If@v4S~^<xhJr5r1t;H6W4hFK_#RjCCJ
z529a&U#+kXO2U^Q)@$bIhVt62<}U&SbA_Yhy127guY2?ysSXByfGM3FuQ2&LSLR6I
z5#d7RwLUqz&qrVr;MUr}osbT8aXl;_ocGdJKBXeq1g-^brf;nD?m9h|SfZPPoK&N}
zO_3*s7b|j3Ev_us<4;dEu}`E3t_l~DrV854+Tft_?vF*)`DF|>)pZh3mqP#Cip+V%
z(xwfWGd}r^yJMSf^0&*Z`wV8bJuT#`>Y5*3+G}rahi78f{leYP%Et}c5gIWi)5*KC
zK`;wI*hJEUq%<C(UL^H--Wan*m=H*{=c$X)WmDhNuQw>yPjo|@Am~S~CxiwSB9`sH
z@Lp4zZq*(-FR&3y{E%6ee6a^}PwI_ViZ*S2>mJ5*p9$*B#3}}d!w(nI4JvP6O<i=r
z39i9YGm__IgMO(B0w*X839=e3EOor%6G_Zw<e4ZkX(L4BdTp&S*V+HN!J$<Ae&!1q
z`D^jX(Yg(i9~z@ok7Ib3_$&kKe=RA|{Ce5U6i-KV<qUG&^#!koGxTb40TF#C;ZC%p
z9ez@xE}ik8OfV^8-0wV3Ef*(B<zDPmBYV37aD3uSexB=D8iUs|`}f!f&kkoDa0~vh
zP|(lb@k(4(uZD)C=8NGH4c%YLVi)U1xlNcyFpl_V^R4tkAw^0$?)`2jP<+I5@5I#W
zw?%*Ry{T}|u@(x8hzkIv3YB;Eu-iLY)GDta_E<R}K|!x8r!ASS!ryGK6>!o`DQtYX
zEb#@@b*iN2?ZWA4Icz$Jbrp@=mUDW_@x=9NoUb`}cv}#&zdL0MfGVu}-PZDCE7^H9
z%VWfvgr6ndlpM=<%K_5hjX;U;=tUokK=)%vt-Fx$-J;_3GWr)K3;WSyua@UGAhMN#
z(E}^1ai^M@K)|n^w%u#Lio+$=@(X~~bSykXHGQLtiFb@Hy1I%V(pPppV(Jzb*Vke6
zQv(z{q<(C-{nMKngs6VvUtCpoV5!np(gEx`E>2(EEK^x+bBELm+|4d8t=-$WvCPzy
zLrQYD)>z?VR!vAH?5ru9%vL<C<mB<^?$h0mXD(V6xD7Ql*v?Ksn7ahJ%2!bl7sEbE
zjKX?q2K5Y&cDHS^daq?6ONu%V=S)YfRz3!Lw(D<0i7z~W(N`~9Z|F9#%-%kNQ|Y!#
zHD&zcM-DTg_GWG!XuTw@vHt6#Ezi#@cvz;&o1d6uvgZ!@HP)PM1zgkV*j~L=AfM>e
zGCg>VQL+WIE9pE&zt|Ac8VI$&!(TN9uQJD3@EUX&h=7K8zu9!C2fv;UIGEHbd7AbN
z%?5b?Pj8BV`RKRb2oUpnYH^7msJ$vvls$?6G^36k0QIVsV)rXJ;N%csTUo#NsKtP$
z;Vkhr-*MtTParO5x}u`IgUXg1T}(-_ZC%#7zryT!=5pf5G%8VxcW$e6^f5I+r9th^
zxAX$n;DF_Wwlwr{+Wm$62`@}EVL6hd`m{3e9GtQLZs6LrSGKyfBF&W04rT3Sk3Fg7
z$g}179|SjgE}GaraY`>Vx)M(S-=nSFk%l%~JZ3|!=j`i=uELI5J!{|++2Acp^`yHa
zfnmp{ZirT3m{N8%EZ+cNpaw$3Ne4HrA4(&#_)40F0VS#6Z!xXFC6wcEe!{(M11+8G
zty(G-@`S^K-bd^YOH>ZW?a_S)yh&w6k-*r#jVM=PnJJD2*LI@*ZGb*f-=rh#-0o;?
zI|v~oWQ$tjjQxP>g;VWQ72{!23%pS&;9U&amun?6LjBKosTdTJPlsOe!vtc`*?^><
z7Xg69RRG|A+LC9VcWOeN6?atg{M%uCM1J8ZX1%HasF=BIl3PoJpEWyE*+ttty9B_^
zw@<!wwy*nq5Esnd&u_~2@c|9VJ5Wgt{fvLb=$!RDPkCuE&#Zt?6IuNUXTcH5+m@HV
z;M0#KBhNQmRibuO%1W}TZ*1Jt9CTL~{g}&-s>d7@i+FkXtZ(*CMme*un53L49;mT|
zuf%M=Dg=$lEoA;)iNJD)@9OgYhTz90x8;W_0TOJ8yMvl=M&aBgzac9JiLXsuOM%J?
zy5(jYLMFbgVJUVt3KoqcJR{pZc@I<1)QWO!Ve60}FAm-;$J+d5JD1&bx$KA%+=$=K
zCzHg3a9MX7>n$4yRXyUP`OkDsdx46=nO>lIXvD7FN6+D|&K#FltjwL(-$UhzcKP|V
zhAIa;Pe@IP_Q;Q?&q$Ru&5qn|0t8P;un2A}INeSDU3ZKa%}Mr%2J3rL1K8j%3dLZ#
zJuuF)37cM3h}!vbI=hzyfqk4|5Oj93PF9YTN4kt_YnNL%VsrjUkJC;X8+IJa8j#8R
z$g7T?P}3QK`V7p4L^Uo;Ait~aK;bcR#}Q$=JbXN3$WaYTy`;o$CGv}&(=nu`2P^RA
zh5sa4vKR19pG0f}9@_}X2{hWA0I=ZJT-b6II#<fW1=<1o;nUHsIH3(g#6I+GX{Uan
zW6(EGwh@e+<=e#98{>R*yl7U<{+K;!AZ^qZ`?RsPwEbXneGUE6Z$yxO9P}$I{-yG;
z^|~iA*~E>csy<)mg&(ShCNNEoJESV1tv{!uq(S6Yq}1<4N#B~WSY6?tLze#P&v~k(
zyZ|ZiQM>cP117^-d#r$)bbd1b<k87uJ{A|HF)UO89Nqo>%JU76Ot{^io6BfrXKMJE
zrkumMn|%(MC2Jr2p!=zb;@!4(%tmZa6CeCV-b)4Am_`dg{V0_t+0@DIgWt^kwOj!G
z$2M;7$1&@<!ZU23rZq?C{6>wU{Ocu`Rc|wuC``2PA#>6NMgov{&Ht^?zhSDW@Bj>_
z^;Q)XS`OakA9gWX#?D_e;-3*Y@geAz?_d1Q)R$mqU^WOopFaIkb?@;MMbBwR`S|N@
zog9w%#yx~0h;>Aw)~TpJ>$11jYIeW!QHU6{wifv1L0pz0#Vl=H%!5ACv+8ya{k`#X
zJzk@XkYysZ4}emA1GvqJIN~t{Nh{U-X7gnj|0AjN@TlF<o2*S*F{nmTa^4XVmhf=_
z+-Yf{7aV$zuT61OS&u#3009y<^56*O;TVS!AYK@umLiJNa0~fvoW;rZ($*4?1>UJh
z*&46m>fHJj{!ul$)B2sNUh*Ks4}sOdH;a!nOgTJ7*^1efHCN8hZkR0CfC{Dg$egNy
zel@#H-tl=k6>)m<bjO{A??$R{O$}ziqAF_XTWvi!d!Alru*JgG&RKm*gTic#V{?}+
zKvA0bqFh$wS=Rx7H5|`0Q!_I)`OOuneh!YvW18g}*W+`gHVDaIpS}nZFZ&tvpHO^v
z!k)Ke)vRhELI~=WB9rHUa*c`PiikVL59uJxRY0*#g^2lX+PXE{bK850Dd%Y%K?sb6
z>23=l)1O|?MXtV#Au0%nmnyr8Oofu)uxKWu*P`EJx!pDhacH7FkUY}FiGAwLTi!r%
z35A;5yIU3-MZ@MlRM;+dZa><bNCH@}4Xdq~4riNFnFOni(pi^2Ef}*t26U>ya(Y!N
zW5uv5{kUviT*3HIPM=UUw(_IR)~vXorm<i?eT_grCH*oj4TW_3QIf9q5{=Ne;i~2_
zY4P!t+^tDQwM?P@roNf96qD@Kd@zaAj4i8BOLr{Q*>|hENA<Jtsa@HSwsFknP|om)
z%V$hybsRA1mC5xEwglJFKNE<)sqD0s+0cyc@rerC5WZ}`W7x^#*ou=GmyRXV;hri6
zVfYjVBQGw3-x_opK5Dz_CQh*ti}RZLO%VujwhdA|Itf3kMQ2s1MJa->m5$pVWJDAd
zICAv=4CIX1Y7~;xNvpFzk88>S^_775NF0^#Dhl-SE(Fh1+`dT5l*akqAi$bfSFVq%
z3qjhMR|7aA`Umgy-$$Qquc$7*tE}i(2b6W3(*A+d&iL-*-e;<^R7%KXIapNC9UDN+
zZwi0}^H8v-q*^)|C6LIePO0135<NlR5AaO#*{s!X)KxMca>e4p@9y#*{K|J())oa0
zb?M6*<=EPd!Z*zeH2!qH0?83af=xdsu=9>2c}CpN&uF)`;4JVr-{G%Inb^RkjI%j9
zG+bcT@bHIc6~ty+S;f1l+iys%^h)pH(q23F_#F1JOIHAAPwsX*>s)j%V~EYRy)XwR
zG^@l*qnp#Nr+d<zu&!LZ$=jClco=dolz#uR|2c2(s{gs~a9SB=)d}sSRi)Nd1|fBC
z=?m34>SRxVL@G{l!%3}LxOeiA<odJ=FHuxL#apRER97Ig5B1xwV&K73x*Dfv!M!X6
zYG1vvl2+?&s}V=*(I$d4CZ@`4vae~4H}2*?z7bCr{(0hf^+i4wIe&;h$dKi}P9!Z>
zzR#K9u#-*Gq%AJ6*9b9Mn9`;LS#)@olxLyGUWbMq@C%3NnQW9bssa^qCs$4bXjfE$
zWD-uv%=ue<igdhABS_8jFB+rT*okm0|NF1Zr?=4$M24`{+#Ev7eN3wLM8lND*ZRlz
zz0!0sn0pWG_IVZ)`n&cx0pc#>#-IOfn9j!2n}+Z1XP)jWRj0YnLCk^!?WvZ3X@Na}
z)jvf+P)L7%cE9p&6jmIAYVg7kHQD>@y4%_G9$UDaTZ&0ZwsrJ-RPzeHWNxTjs2sa}
z5!6st2mV7KtcQ!t9-LJFtkais?6;3>h|In=2%32kAFHFnpx?oCGhrLVhPr3?it5kF
zOF`KbKR7#vZ$#_-V6m4J(C{g}gucstx{-M<4$wf`81O#AeKAVlXq_h$UShOGr=D}j
zn&psJ<mjx*qJ&WojLp$4dOhW!lR1i|d34<HjR?b-HM}u1aTqEw{YKDbrpHfe(;BNH
zn)i{`W2=GbO+$*u)42ymjUP43+;AP!-GcT@aRJ};h_nj3Q=pg(8_s(>e7y|8;ZU_f
z=NunWgc!BQvgCY}*z*#0r1@9CYDG!}&C%2(+pYENTe)=~uHru<$7jXJDo-53Iiur*
zB9APoo2;5_`3W$Uk->Jc<Aa{k0A?l8<lqMEqAaYT>=`Y167ceO{b7Wn_9-x1%A&+#
z&8z28Nj}5YdnqL|wJ_C!k*{_89ZB0_)7`&ndPTEyY)v&3f#;4t3p!kxz7<vT<t1r(
za|z0%Ft}4&k24sM+AE{$Z_uUv6j)iLwL3@_55vH=hz9eMVua|Q8cZekWPFkTTwda3
zZiVoUqa?TTc$Fk@=8~JkVK2eH6fN3N<y#V$msTTK^sHBxy$i+jTz?3%F)=bsjiUIZ
zt7)qtl~&+oeT6#+1iGEE)%oFRKaryMBYx`FyDl&X=^T;A*`V@+?-`!0jGG@kTMZA0
z-st+kPRjT6+|@vA0w0G@(Q5BYyT1kU0K|bYlOSbXeIT{CyT_^bbVF^r6oX&7uO4cH
z{sH@>A^gT8iy#)(ZV3_p(qnH^t=ztDE)8Abe*3y*X)lq=$zQ*Ak4H|Xlb1K$EZ$>h
z#G70)*r)1i`(f5+G><%p^&$~d{t>k2Yke3FwWO2-y8BhGUu;5d@84N-9^O{&+XDLZ
zoh;<ti{M9otW(pyOjpc>(n((vD1zF64e`A+*Od9V+wdvVy}3<m9~Z-{yGT9*%Ce30
zBf*ssJyKr-+G8=n0S^y86`96qc!IP#_vD3yTs|m<5>KQV8b2l50M*;5;|pd}3(?AD
z)%@i~n`KevTw{^DS&u$I*19+96;__mL~l+y7f<DDKA3nAlN)w02FlwoJada~QAn@J
zDs9hsnvMqJ3H>Cd_xf@yoZYA07PUg}%06*9nbg@&Q)Jy8+;}w!|CYm!F6?e}m;iDM
zbCeJR#Gv!@1{=aTj+Ow<K+o@mgVj`%@3Oh4iF?~LGSC|nLhfoCI&8%eM?x#8WT^tp
z1F>$a!UvaLb8a)%0P*AgyMRp|>%I8(2HaM)v*yHTh4I5_n$fPDG+gS`^rWE3q-)c{
z99Izk675-e?Y7+!OPmXeYHxp7_l2;MdTgK}g$rhVAv`>{5BWY%J=<V)UyMng&s*?y
zLutM2%_pj$)<c!yn$0rH`5!+?IfFLfjlCVO9!EEt(l|1(BkL9$IK;_o17_ZgjQ}Pw
zNR>8%m!t25TL3V84641H`CO%&5x`A^dyd!9Z3rHvhNn7Bpo!w@j$i60mII;nM<OYX
z)+Af5=4zR1yMoh`<vrGm*%j`?c{IJma@5vYHE*;*i~2c6O0;kZb<D)w-0CaMB@h-J
zs@i?q$h;cYRcm%_D;F<Fc)EDJ+g~hu(wVU_yned$vztlUb!>I^R58`6r)^cK$?S7S
znA$4gdf!pwfWU`uhIxrBAn|yogNAMJz2XgluOn7|N*7H9>WN*6i3V0ohTZF%8ryZ1
zo=<zhddn-k9jFfFHzLuU_I;R-cq+BBoIuuFQ=<Lo`d%6fQpl~qGmE1`N{f&;2ck8l
zGqEK<;4fh;Ryp)(a;K@a?mbXC8gBjdcp--6^g4q_E*<)2Ot=xI9_|r?1?PU^VSV`c
zxtp`JCF#8gSM0)4A9rPiqqx?Dx+K(#?;)PRtPaM+hwhu0V~d$Cox2vaSSs>-G_9!0
z=FHA<bjVRS^$rc(wPy`ul8|EJoHD#jffh4DlHu8d85cz_Qt{JM{uz15mqpOZ)G_O5
z-GuH=r!Bu$@h*e3iB&h&+{()(ZQc>&`MR?g>}bE~K9jA4DAbJOtNy;xSMH)j6YV4$
z-=Kb5O1@XP!65=Qjv2^Y9a~L>m~}+^Wf`WEPTxPyJwEoK^4EQSaNt0?8|C~|()CCx
zL}VChB{C&M%AUx^$&9z0P9H($grW9sj#bw#GRNGW-(-txXnP;P$==-nbT)SZWT%8H
zLhfxax;;rHVv7icx8&{YV^dCRLWA{;TZo5S6%w)XP6w$&=I3wEvj}S5S$;au6P7pU
zAmR0hy!F26G;SxKDyFAfaRw=AJ=%X3%fB%cAMRp*(6B%02(q?jwHb1EtvhK*q(N8#
z&+*Wp=`-G9tNcUL%g!RLo<qq<Ey{C5pQ~BJ2o@oE(r>_n+z}I6cj!0q;>VZ9=;%Mo
zAX?3UD1N1b(J$7r-(&t+`1nphph{=^`v`n}&PLsf@<_5f<%W9womG&UBDS6Kaa%_X
zzNG8b9-n&-)7}iZmmu?s9`td9^IRjq{CRbnz8>_7F;sVSnU`gkbXbj6j|X2T7({U_
zlqS0If|_DseqoK--pMHrA|<?gE8@Mo9>5){gQXECkH@+Ffi*@M`?O0ROI?0+kWp1t
zj{iEb<G^?CvuaSL)E6ZsbHvZP{&JavD4yx4LJ@(LqmY(^Ei<gXE~tb<e7%Ok?8usB
zT-E%zQ`)mnJHe=ZA^F8@jFk|W642yk@<K64%VZ3Qvaaqj2=6m!xJWiJbvqeZMiz!j
zf=m-<pDUoy$OKp37zVvyvgy_QesSz-7z5QljUGqD9c;NwWB%mo>}1_wEk+DzDK@4v
zTXd!h!>IGoZ)};=zTUs6Mpfr0GQICjjCQ>SK|XOm(A1M!=TFEf>bDtk`?8&*LKb!_
zXE-`D8uYcw#-7ZqLjmb7vk+YDcvO6h%xL2n+b}^@hXE)lMxBfIIh;255cARNlYLZ%
zqp*PaK`#%}BkDLRE!4=^8rzZXMSFU4`a3|$+d3e6)CIh+D`d7Hxl6l37=b+*+UQYQ
z7%FI+z(=I``n;P#PMVsTvw1>h`xZWgRrx$FVz+vdfP=-UeAe8APw5Wdtg4c$xSo=x
zE6A0xxT8^FwAJ2La#}toF3x_~PwEIboX<G7apNo+s(g(8Ik)@og6|1s=lvuPJ2;zH
z#*!x}n&dZO0R{YxaVpbWm=B|amA-L}_N84rvIRC(pJlV#Yk|(8s4gdvd6=rjObgvv
z?RuRkgW}|Nh`ip;V7h4g_a9aT#;f{Pfg}cd?<V1LK-|a}Yj}=Xp@-QjKRfxXE1LK7
zSiF+OppGEDZYCG+*ZLw+WpLUF%hqsmIT0s+6L1Be;N@kpL;9}d4gV5(f@gR_vV+QM
zd@!O0L5ZWy<?^au+O^*{SRsB&0p6JI7V~&{(BAhHdAe^u%*PVc1nm@^Z&wa7q4l)w
zBuf<?S=2U)U573dO84WoYE4D$M|?&5G<LVRh8clB)AMLK9&so__rKJI+?evYF@358
z471a;AJ{mOC1_3nZ}c~8m1@3AzY~#rNHz0{jGc^C5*+B;(9<^DNGNB?zbUdK<%>^y
z#RhZ1w(2<i?oa8Zv+Wr#rb<U5H&zwt%1$PeDLe(Tk6u3V5h3Q+BbPu9Wc^_W{XwmJ
zb{7{vntI93UZhiGZp2}U$thSMq6u(=xENNzs!u%l4eYnkzloxmXCNmvHB{~#MaoKj
znfg-l1^F`5Cp=N)Rr!D9I-~B)&txJ?-fHskg4nEICnMLFm!2-Mc2TQO(=l!?qn^He
z?^ci>xJo47VvyorTF<t)&}b%+sm7Yfzrr7X_YT;3%dNOtBmT+K1FKq9UCVo{)L>1V
zYemTn&{Df9RZrpJ_Bz{$C?ZP>dFGJ54)rt}s{o#746kq3&(WE*Fp?}nxB5sq%U4mi
z+C}ky3d4oXEzodWc41GkT5aKJ2U$UAUis$YDE}4@{^Jq^xi23Z-_aD?p0QA`Thuz*
z{abwCcf%rG7IvIFZAxpru9b;4fw%J%T=Dngsv_jkDLVt*bkn@rV(t@C{e3&Xqw!eQ
zu;`{-HISpTUhKzeb>4dkC3>qSW#!=6BUR0_^}=s7$7tSJM;R+T!PkrbaiN#|M+_S^
z@x=cPX%i%;P1@=7K}=5&?quO@_H>EgitiSrDY{eSWqp@9Sq+LZU1_(*b}PMnQqrti
z5jo`kQ;%-x=Y8cxuD}Jvbk$gaQJ31}%GB|U+(wdIOtw&pRb}Aa0#avc-4fMN+-rHL
zys~{wW^+dD*i2$g>2YM0<htR-#|i!^{-eS+z^EDEcz3?@NRubQc1qC*^UOuHGDWF@
zdAWQ(#kK!>lvq=dGESpdgsHi7@gL=ouo<i1ue>Fq6>S?|+#D5H276zK{Cd%-cV~1(
zxxV-hbZLJMSNxTC+FJ4A^Dj5i;_8*=YtEY^&(O99;|JT}(-A$W)9j(%-r3K^(~!rR
z76C3P1<!HcB5oKq38h!aaoC=br&La!ZZYi(i(iDH^K6&RAe+kqQ+@ruF+y^^fi--Z
z=TNOxjkg@Msb4HQ6JS+$rQTKKJKs!0_S{jjSYx6X&bwcp9a1z^YbhbO*Ck`Zp~(_r
zRW`p?K~th}RckkVa}e-gCqfa_fB8|_aEZcQn}Sg*KGV%NN$tk!QfI?BkAtCKdv6L1
z&j0&^F!%23HJ=-{VZRA=>-&bkryTVe%${$3-360{Wpo<r2r7$|Eyqz+jn<_K7@d}V
zFcL)crFgj-b+pS#)r4@5I8P+diSF*woIJi!`oRS!L3+`5<85&03rn3j_1nd)Y`(DS
zCv#qZ8Z+JC>ftlrA*tA_MqOSc+St!U-)!Mi(@;_n5;*d14HFi;QW#tx#tU^aVEBJb
zePvV}UDGXugpdR$xO;GS3l718ySonV8Z20Fx8UwBgS!TIcMmc+!=1_Vu64hE^JBXE
zbe)o2yXsWGy_W-B-^5C1I=8AtwKiAQh}il**5BHj#^=7yjGA1fg3r%?wYC-x*m7jP
z8(J0Ee`w+Pr0d`U>NeB$3V>w`Y=_Gv4EnRc$w}~~gdLGMW;x5@6~DWPYZ82iVs6%L
zhbFX<Htk`SQ@S@4c4dqyCx<5;lWlI~Y!In_Ah31&(t25WqXuS1#$UPU_>SN}0$9Pv
z2cJ{8cpSzhi?)-8zG!V+cB6Y$<+3w>r&0+I=asvdX)a+`bx|W?H6e-q3q9FpA+)Bk
zw5p`3<y0*)^Ymj6EkK!!|2aslKn=4XnSlKz@Nbg%r>ULVH!o&=UT^F7)FN7T*I8fn
z_OFr@WdEYLp^H@5mRk-@R!m-$>B<{Mo%Yi$RPLbs0wJYRLB>D)#?U2hWQu*UvpHQw
zjke<g{>}W1S)(*}f0UJnUm?K%eV_2AYxnT`sL*6QnFxK+v~QEg2CsN$QFd}dGym_G
z3{G9!?kdf~1VO-!MX$Xnq-8{LEs$-;*u<>p&AWT~h$u#yF^^{pkmWoUxw_2OiDm5F
zue;;>!&6R97BidN#1Eg?kYn6Kvs?F%qZ@$~K^lT*g|RX2D@#6uafR)<M>j>5^yTZC
z8f#VWxUkf0=*z1cGrwGUrdH&7ng=Jo<Cv$x^IEx0Lq4_T{nL32@`6sr@ZYkxFsO(-
zUymIyC@u$7w=XMtu9aWbd3n?q)eVmD3^$yxA-nQKkMKP&9T{@25-1hTH!Kuo&y=|=
z|N7Lz?pd&emm1RcR_w(qAik}EOoV)!Shb93)Q%stY-IA}*?Zf*9prV;-#1i2?8`o<
z;@d>?lIZPAg6r$I1-h^Ut4GwbP#S4jO;5+x&3l8Ot?(0t31#z|St}F{oo?R}Xf6nu
zP8{j9N}`Wc<f3TH+F5~xE#fCQxTwtJW24;3PyIFL@9W0ByiDEXPsd_7IQC8)M@O9}
z>WuQt-#f3L<^67#AWfjIFJcDbO45ZI<B{h7l0WC|MIx^EC{TTeC;jbbXVs;jxP{Hl
z&FRfrk%so-hMB-FGU2G63;oLIxQ|kbWYFrc=<zN+CT(zSe&wE&gKbiHmAU&P=x^E%
zWhlRnPi3YXR|PDzi0(?u&An~qR#6YMUg%vjcJ^OQxJ#o`05|22PPG=Q<zTXcZg+St
zuv7b&ua2W5MD(+W3{$w+xx4t<S|-@%dxeH^67eDJru;VCEu|b!UzfM0i~$jxI>;sC
zBje!q8iBSOfM+_G(!yEf_zSxnX4UBkpOI9lqzstfPe(~I(j^^_VokU5$j?@I|4e0J
z`dQnRZKEGyi?-z1btCz8ay_$)!(smw&d-q=HO~Isn-5$+5sLlalNr&eRV>dm9^d|1
zYixXJSX*B?v$Jh`+OTbE7`$D#b89$jbZu)$p4X;Z&)tS4Be}!06Okerq<H9*l0rab
z{@MkTeP^xL9+zsq@3Ehr%Dgc)#ugOwgvGztk6{RagP{Pz_`^AH*(ODIa`$c+lCCxC
zwCD<_JyiLQ*34)V5ML}1sJvyU<G+p`M@Ln*zN0VvniC3=(ex`W9u}nH>2F{#WZ4&N
zz#$c>ECI%&E~Ad}X$UC0&AGl}9tJ6m=yNL?y6v~_>^wiZ#lLBd_tRHU_F4I(Z5@FK
zBTMFthq+xdXfVqZ*+!$sY3=x;FFNtbKOY;WE4NUW@N|zSYr)^=0_1dnwGULZ^V_$o
zep^*D^R*nMb>TK4A1x4s8m5G4;mOo(;OEj%;`JcPu#ub>J8+94uhMv%HM_K^VToA1
z+I2c?ic;5LCs8nGd;y1J>uvTfbX}x~g;6#8;o1#_4VNA(5Yl;s+p5ntUUfu|HKWwU
zI(M5kCs25&0#7(t0|Q;rfat(eQF#W9FWFruVtGN0U{PkRw`HvcL0Aki+2(SGsBvMR
z(0uB(9&kq+y#;c%C;p_@u=c~$k#x9F`pw9P3Y_`f7#gX+%(^vxrLpf77Oi6ow-PQ^
zYQl%l6Bz>MxOndAoj-;LRZD8MX=^@D+zvlxXGZ{taM-9Ydc`ax;&T)=`94MzY1Zy#
zl{Y`RxG9aBxSCG@;}1HN&*JQ#_MD@NrX;S?3FSd=NFalANK6zQ8p2`>n4-Od`PMhh
z{koK>s~%e$74+4fC6$V(F54!Ts0KoVXE9Z8Xg3|ybq+INm|)`t|DqWxFkEpaCb_2r
z?o8>kh&)f-gBv6*ZsFKW=e2)EHJtH|syc}>2uaNe9KuCj2<!}5ckbU4%<rxqslXHb
z>|H!!{)t4z7(d@)Gmw_>6A>sZiR=7WDGczfxWE-X8sA|6fcB+5fx;`gk*o@=TKXyG
z>^L+61#`;FQ`SO(6Q^Fho}2M=vb>{UXLdm!sDi@et3vAPfc%-E+`pSXs!``)7isgf
zHm*kF?>_D~SD#WZ(UPi^3KyIiOR5EJ`d8F%ugW)bbek&5d8SCUEqwo=DU&?-ak(F;
zH#MS9%#sjmzYgc;`KI@9k6l&1&}=Txp0hq>Y4Gsewav;2?To%3Ks4P}b<-i<f4B~4
zzfd-24fN^Jc{DYtoeZF{X0(;h87GhO44*!#ux#YmIA`487_hT?-TMv??3Tif{EWLk
zj!OUB^%kl&R6ro*9DJUoUG?H*;7Mh2fNq0oX<pPaLfBqA6`kx572Y+^`^t@rGd8_%
z@%kG|m8Ui{_#pT=R)*!}i<;qcrYO{C)F(QHqt#|q>5F6DMl<zAkTl2a)r<N#`;BVj
zS|Q);vcvOxX55A&awTJ|>@Ukrdw&Ye)jFh$rgR%j@N#NW2{_}~!lm0{b5AkMz5)Dd
zml*}looZW8`fH{$?{8pZAVf$0d`{6X)j13iR`MT%((!U|H~rF0eEi401h(Ek`v!TH
zH1%&q<TD3+e|5!hhk?F~&j;2Hv%6`@2H6@)JiT)R0xT<^o^8alat`bPcB|h<_%ul4
zcZdpQ<xP4-Z5en0@F{)gDQ#Ph<1N!6qh5^Obv6UXm(IUs%~_9R&#Uc1zE1zMd#biT
z(+0c|RHj>Gw5zqdDdfU#wob;Ltf$7vZ<$lIWJ*jcsqg3f9FKY5%hB2bUOd*C-g71B
z-1{g>ruZ{m8q^^uTC*3N5mtnB*y3w>P|vIcEYwfxh*lC(d%{xDp>V#gzf%CVodgr|
z=u?G^xtxhl*o_96W5Gg9#(!8Az)n<JOc_KUmCynrm8o(=K9>U_BkG@9^S%2`W5>0u
zR_#Sl066Bu5SokLDgM|vTjr%`q`6odvMu^iwZ$=J9;u<dPP*7B)RMe{uyQN*-rWWy
ztJ&qu^!S?)%hGkfsd4MtMctuW`qPdY(a1&z&q>tA#<Ls0EdmU57v8#+*UFd7g?@Zs
z>FCKR&{{yvls@zXNZcLC!7h+<$Tn)wR?De#&1W4RVJ~>rFwJ=BMs6nNo{KtUCBHg3
z<%y;sJIUHArVB1PViIaM({gyIx>efd{~R`8S4x0zY#x=l`$Yp2hKX*f%hs$Y<W+=S
zmbHs0>XNzw6;>M{qeR##NA0Ew@OZ&--P|}OpCj1GX@vLf+kcA-GG&eRDc}f-!Ot_D
z*w?fpp|X4I+=kDxv3c5c4m?{+EluKpg-u!@jhzwv+>{A8yTSTs9Mmq1KK61Du}`0c
zHB}Uk&6_EW5I?cWteK@__p_6`uHm+ddgLTTuLF>g>2YSCV98za&TR>H&@kQj<v~7y
z6@C6~^UtgrP66N8Akm*Jug?Ws5md7P2?DCSj&};357%{Gj9`H0VLS6H&S^9ZGv^@!
zEKFn#+tt7Ta#ZJgH&6enmw;w`Wclevt`6XZQsr7?1(NC0>ABUO`-g*|x67=0Bj>}C
z1W$Olh)?G+4G96PcqP8-s$QP&1-L}{Gp>N@;cQ(ajs(F!miVmkT+en+;@`O$HjRGI
z%>0-bC|qDQT9pn{<e?c6XEGzwKO+JMkV8&!&o)untndHOgdY{cpqYLHzt;B}{2_47
zeyv?$9caGr{_fQCBx#|L1r_z#3tyQV(x_r=3y|g@HPK9nl+ObR^4jw`dKlaO0L1iW
zzmJk&D5wG`n<USl+MN4w4pf8sjv6g%=tRYV#DzIbv~&+%qz~8A8pNX_o@yBlq#W~F
z$y<j<POo;SEdjBE+q&48zB0{Cd)nGxZjs0tWJ1(BI#$7EJ9?@p-0OvP2UMnc550I+
z)@@ern|RQP2d|KceGj+A<db+SQriX$GGYd+xhOZ%wG$Zt75Ctq&WntvKSAp=F+Ua~
z6cAAkQZ7)9!V7~O<UDBm-u~0##EnI*sX}yn4X6BV2YUiL3-C^*ac`D~`FcU7#|<-}
z>bu=SGUU`1&qitE?jY0o1XiuIc2NhlE|yeW1<Gv&ee2Cdh@S4P^D(d6{sBh!s%l4)
zD10<`=kJ`vgSadCI2T6hCv+JpsdRZjb_ojm`nRX<0E~4TEYe4{Ob6}?js@f)56NHs
zCKxks<vo%5Tw9k7Qevbr;=8r7FUPEf1TIGlvduyY#NSgH6lD<bwl$ueBcsv?Y{N=<
z!Wz6;u@xLo$3+{<ouU5Ya8D9xT48JYGT@4HLkNuj8a6cZ7*#HywQPsC>Au)HJHeSX
z7&D^o;`+WtwCKWK#r5}Pqj>#ceXm_bg{wac7TpI8K~;kDvu@+GCUH|qJMou^=fYB7
zN^xIbH*wFJu$ub(#{()jw1*c#4vKGGWC@MUI^d}#Pw|T@OSI`R8eeqSP*cH>ssygn
z$z)<>ME9@;sa+p+qwhb2vIi+VD;i4*&cuAjC*^ogRJwCXvLtrUOo4u>N0jXXBBCk#
z`oD@N;0nI=i#>c7dCpCq)M<M;#n!G}Uv*!RB)|d~*?Hg3pNOxO<{&IwRf!hNlMDU}
zTE!!rs>1O<Mos7uVsH4kQ*K^gJUxB)P2{gv+;VCrZ<sB2zd5*VOa)ETA=Z;70-SAw
zLKl}MygRtji527Jfckt5A6YeLwVV1z+<h(-q@;gpbpGJ=yE+vOJvkfPXt0=n{8PC@
z7B%}l?&Rf&!LTtf-!itlber%(llbRne73zg{35AEAQm43tNYS(>Q&tHU=!xuXz`N5
zMwY8&zIryi*+i6|lJDLW6C*aN<3b#}%o(T7>rZoaag7Z{;n&57^|zj6Gqp!$Oez`0
zHniaVU7>Snzc*gX6yKZn+6siOcHWKsf)}S@e1V1F3uePJPU4SdLErc_?yrty*8k*E
zbtK7z^w-j)Q<92bCDD!iLWWzLj3B1D{cu??fSRH+#58G*S`wChy|=ee-{9yCIjRvC
z7bi4^NVI!zKCsHu_HvzMN7oQFt>`)GfTqrQHY4SF!Birj2yz^&U#eTz`GRZF5;xRb
zap$%Ader}>EjGGhyj}G8UM8C9gFbz`&Y%5z=}ior%n6O;%Ts+n_Zy9^^WDEU+?K4<
z>%+fQ6Ia5&z{Kgnq;#e6vh%peA@ZI*{v{#6<NjUtTB~wp5Og*zp%+QNxCJL*MaS5w
zTR~}3yjNvw4Li7H-25?)=3?2rofq`bbL<4dBi7Ges$mTSX=y2%JRe-AT?ovdJky9H
z#E+ZdMFt5|_Z3^5m|O`wAR=d*R;sgt0Is+IWAO*KgVKlu3PDCkx$wT}Y@QkF)4BeJ
zrgPilN=h=YdSm3mpN^B{p1cOD%IJ&f-M^aX3M+V+$#d$L{Cr4}BsDRUhjy-(fk`R)
zbw7DKjdOhz;XOjvGivQtO8k6%ch91l{7$>~jrE-vVmHED0?@IhXtXAYO2(aZe)3r>
zYGk|*T$}O*f8Dce`|#s1W=|RxP0(9NFkKok;hU<WipS~KGs{rAAxu-ogpj(Qm4%N_
z8Jv_`=Ib0UA*h1UI0=b!)V}nKIix|O3JUQ9p3LQEHDmHlicvvgp{vPQoV^g^3bL{t
zo}_uKGWDK}tk=`Wklvz0H=SDNCt`7KEVCLtJ#qY=CHB4|m5_*O|5NL&xv07APRnUr
zc2#i#cY^TrCzwCR3epWzV*8_g>pBx}1<;90b{-CPD@adveu%pN>dcrsV|cm>_p_Ks
zu6k~;W$=&qrPEsb6)k<ryk5k-Ej8(KI^1GKqNJ2j=hZX^<hk>)_HH#q8@nR<XoAz&
zOhJy<m9^+K04-3SH+}tn;SwTwt{O7e%!iQW<RppZByaHddC{U)jD6#)B)<%#wHm|L
z>^%!!7Ajjnd;v!LSHKEV+48H;?Pz+EHtepgkUlM@^qcb!hR+6iR{_cg$Qj3_C&0nv
zN!H1nQd*rSTTmaEnQaHar|;XI51yVkzIfelJ*aVMn_uOLUL&$tG5}mZUyQdAf|9(o
z_g}TCIFqLofHx0Et(86`;L3Lj`JthAPwJT|pPL*T3T;j%?}$&`{2qE*6xNBx&Xz|g
zTbFZbm`n>-ZWB^s5`L8uuBe(AE?+0H_Qfc*?A3f+5qd>iNO2N-deY_;b6<X8`zW6g
zU)|dHcP9Cl-I?F-SoR%|J~j1SC8LM@>d4$<tEeV^XZe>SWkKefut8R}a801u&kxdi
zDc*jfdv1W!r>_AONYaq*ebRj0A|XXx1#${Do#W=Wj(T;q%Og<@iTs>Q%PVX#qc^|V
zG!b(bPVi<kO)JhEd!<J8SLT24;as#W*9YTLWAoxDaOh}uJk<7h2Vc4f&Of>HSrfg~
zR@b@>M|TRb?&TC*2kmp&T&~S5nlphekXv8Y{0LvwY&~q7%b7O<aL}DHo4tn9j#rV2
zJU5KWirqo^r{gVx`x#3Yx0a`QN&U-2tTx}X0jHR+Y923bK1?s>D@Au0V6Rw-yoyox
z$H@%=$JX8NO74$i)$FG^j4B&s)B&v5GQP(NH@fE0H=}S#)Daae{*}rzpcfKNzm<)d
zc6hxnT;Ahn+FC-|$)X@$lu|Yegrjf5JaKbO!GFD|aXxa)HSDd7pb8WmJxJg=M0J;K
zJYa6jOqB~OpZkd~G_!aKji0Xo6bF*6bKejy^ncVzc}Q|x8DGH^*fZYV@U@fQ{#t<%
zx_Wzj24(_ywNze&8A2dFO#EcPG+t&mH`^bsp;5T0=2CgxgZ{^fh&(m$awAsE?XMBo
zC@EBFS@RF6#H9OLl31Cv&8G%F4cQr$*~z#D9vDQ&a*T5z8VB#eFs3;l^NeGLddAe{
zj-q-${9i;&ien(ORQ*pa^de))Jj@li+`R{Po#8a6JZGn4MQoRT$T!$;E7GpJ)Cbp}
zt`6H2VQ1Ri$xbaR<Kn$!37eMVkl@9cK!RTRlP~oFTM;%WDCO-}fgYt*O}4L7bm6q}
z`}0Y~NExV_5>_hppxNTnr2LzVQ}P$}7_y+@Z8oIOhRh#w63#w`gl1C*UMx{>F~4RR
zXZrMUu-<3E8+D+4Rn~HCRXT{pi9IzIpu<n7_O-hO$IoA_qgAb#w`+-mw|XKzOY&3N
zxCRY51_%i0;l0N9cycsv)F~{Pwli=@-tH0?ACttQHC0NbOdNr)r2Y6gu7fe2E>8rY
z!C<x?UQ{hO1GSc0gtrB`7uKFaQ0+@_ZC|_9rP2xw@!8Rv#9k0khhE(jUgMWNu$|jy
z^l&$Om*-H2lcOgon&qHqux3T-0-TN{hC_dheLRYZL_4jEt&jI~b|EsE&G@`)(ys|p
z6{4A;crcUA_B57%nEH|Cq)+5^ku!!({inmVLa6h$YO<;7TrlhyH=B|5WtQ(`{AZ?-
zPYJgx%}P6vqoI_ePub=4<|+Le<?b`X!?Dmi1>b6;0?P4EAGLRVj_psrdb|siW>*(n
zYuqwfu%ke@L2m;2RqHC)OK6dslE}j6e5QsBZ42g?33ZY*6KdFRXg&x0=vqPxpq8Ou
zP?4G+J3nHrX#5)(CE4iG$s^hhN9^r5d`yC+_$A$Dyv@tEkU?`q2=uJovXrzS9u?}g
zSLO9m={270ih7x#$})I|x%K)9D$Frt_1O}TN@O(ih~#F+`}Z)+L^{0AhYQrIx$8$%
zhG@a9PoAC?-OEU|!n(YFGF?6MLQp$6)LpI?VB*(&`r=X!(=%-xjE~BwDg2~CbsY!e
zDI7b!nrvyMqn*p~cc$caa3t&^ZGdiNzvbIM4_C(2m8S|(6JoMDVjk6hziEA|-=j`$
zQO16C_RiKtxA&?Al_Hcr>;{}~c<iQk0v+mlgXe%gzYQ;=k~Ev&z>>n<_3GR`G~vs;
z>UWoyTdN~455>+eCSGz9dxM`|&70cH{oJ32o+4)2>ehdoDAfj{l6~}{AAU^7nQL#O
z-r=yWcrw{Be;=PO{4h7)hWsX57IZd|Sns00!m4)y_^2%?kPKcOvv<lO&cvmmq3w?U
zxQb0k3gfS6%ASgxFo!CVBqoi91cMWz*BPeBD}wk@_d@vWi_5GF7D~NXLv26rdsF+r
z<-cQJ5sDV>*e(6$Yb0G0*+Cr+iwbzR_W(5;-Ov7$$*9|BM}7<Kk_e<58TT4Fb`oE8
z3LsO8B0vq?H}Fq<ns9kl*EJoCw2S*=(|B`@+781cx}Wz6yf3Q<IL_QkB}A8=fUN6`
zY2h!4^SXcHH(ui7XRFNJMmzro=b2(2sNB^Hd`C3`c4#?m9#X%^VCybg${aq}2-yJ6
z5zQg!-#N3?(&A|L+da{x5wRFt7opOS!@74!U0D|fMV;(BPsw<GaibGUZL0Khi-16N
zL;jAF!$!3pLgu}>(`2{BXl^f3w};@>v7;L6gCK?LqazMZqxn@o_Omtn9AhJ24N(>@
zD@FOGsR&GJL7}Vp9hui30^=OTw&wW0b&HO*vrfzWXiH|}$ZSTo(Wzy;&Mke0)*h=;
zsb6{)M7zkuswnI7bIYoJgL~WxC=0g+%r!V@(+OltRS%^R@)jbMYC2s?ON)hl=OpCg
zgSGS>mE*FF8{*erxK6?r=(~4Ubz>n%yl!fm^3zaj>A!KhSA|gNtRuoyN>OqnuBh7Z
zhsDHDg-XuTYj=FJ%<Sk|oJV1s!qT(kS!UTSyy*O1{VEb5i~ih($!4y}l%7$;%y`?A
z`i6iL!}0EU@Kq%{Ti=J1&QDTe^`i-p(V=*A+Q`AtZ6B-CXeE|Vmr1jUdWCT{kyv1*
zsH9x8Z00<h5R1`-<vqAcU~>fd{?$8%mAoX@p#eX?ylL=+=0vc(s(tlb?DVwqCuS&<
zF2B5>L-TUz2bfJO?2~lsWxn>7(*zPf$R_C`_YJNwC#Zh0s|N}9XYj2J8m3~_nKtG(
z71j<5GyK};opOp`X0oO7TP#yn)xJKY_PXO*ITs)@Fh$zA?seS9An^A;0CJuwiVQWL
zDx@!;RM)kb6w|G0TaT{xV7*I|3>mo9S;<Il>+NW7x@mD+c(HYGz9D%i%4={k>nBXc
z&BP;|uwH4o!Nay|PTJ^0B;2tyPxrQmOhk~FBD%I*$0sY0<t5U84`XcT=erhBCzh0<
zmXy$Z{5J>i#2>t%eEU83EKO`LAR<$qvE_GzoG3*S<=~ifir-D2qPdwdn;5TXCsBj*
zgWR&LJ`rh>WCeE@Xj{G~nu&(r$8{F8)c?N8#MaaE0iPWY>=wDYzB3nH5tqhiw<E84
zDA{30_9?ewffGOcR(<w1VXw-^aA`WreEm56YLqE~OTEd`g%>WB<061N8IWkKPfi*{
z`Ip?}i%{8j^Ez<5{!&q|>(fz#hg|>SEHuA2Q7N?OoX*{4!L_XIDoY-yiia`oWA9ea
zX}nR<wh8(3oJQ(-8_PbCOx<a7+EK2Eg$}+YWV2Dv`tq^LyM&ELTfBcy0EvYwVqAT=
z1`d2HHBzRZxjsdgxmTWo=CE-3ok{$t6%anA%)jnuGEAR!`yAHZgrcfdWk{9#dx%5X
zeZIW6xUp=)b%sQl;8FHZqFlpgLl2XX>Q6GCRpX7}fOMd(p5^O)7mp{$rnQ*nk6z0|
zFL7RWiL=v*EL<MPNZQc|W?EIYc4dp|7V4~q%StP_n~mWWOs6mT3jh&Aw>}VmvvoD5
zGdOF#retas=%k539OCyjd$+W5zAsCgSj99k^b^PZT>ymN{hghuPWyp0q9?FLHU#7K
zIR?VM@3{=qTD-OvQJhT-or@A=z3$vqd4De`x8?E-<A3pR6_~=&TEDNHW>3%p*Ou82
z`zSM*XQhZ7MWD!J+bFBByBy=n>H|8&hE6I~cCm%SCvGWfGH^?VvkRz3_88{A0spLU
zNsQGo)Uh|^*A|8Ra%xz{jZMm`baz{A@^(m1+Gx;tNhUtKZ!=P|$;~3D@|w=}_M41Y
zne@%VVdeY&`A4R05O105=c5_pv6b9t8wV=_nMHwMUbijnC)W}0_6nvlaNR+?e9^+h
zZeKVbJzFWqEzS2TB~4=00fnDwFOV#!+kB*~lw^&j=?e{~;X#Ye^LyM*d&qo>$Z$_p
zb34&b5%StNRKn~qt0!N*SCPin>FgV>D*CaBSJdT}#(VcZkx}I8;Z{}df$r*Vk?iR*
z2ETr)qFJm4j)A>V^uMHG(WIj7Gu4^4>q<*V8sE{YXixUf(T7nO%zrNN8Ia-lEbtuB
z(-9u+OVg+=s5=YP_$EmO765Wy)@mNra2+GHj)!vMM!6?k4w;%KYPUIG%X*O5dfr)&
zKlnreM~)C1E3=lL1bD23tzw4u81R93PXIS_naPz7EHblIt^)ml>Y)wbZCQmig<evV
zqiI27(X*&>H6baau5xJXCY?AXHZAhz9RX)9EDX8gLTMvmsq1(7g0JQZ*?Qnfwj%sl
z<U(9tro9a92AY$Vb&Q6D3UiuX5<)ip34zYdUW~!AgoM13G<ncNL~_>SeLHBk;(Uog
zlTSvA9sd0|cXlT5c}AwfXMpRo?R|;HR8hOUYDJa~@#C2j(O=NVVhh@34_`W8L%auj
zvOBTYgH_Ii(?pb;PZ)G|YpeZ@mpYvdOjUlAA$&v83KP$ad(~rybBh;sF@9ddxAEw7
z=SOn@3a5(TmbwM{rB;V2r5f|!i=s9RulrkOrRpBO*|ZeH%2|&)nH!yryNc)BFTs}6
zU3=p&o1V5>Yb9y$yiq10@A-Ue`~g<}EPSZM=2{U<6Dum*bAiQt^@+Si2Ug|JoeFdx
z(%JOM5J9!ik7mWi0KVlz?t&M`m8@(59(0jit25O~a<!rCdl7jAHXbRQ^IkUn2%GPX
zZdjw3$xmDgYtL7HC_vd1<-j1tH;w+oIPJK)9xm(s_*YkLSd0c3Mi$m;XV#}FHt96U
z{SXW_19*9E(>>_SB1)sZjw0~$O4B~GiW0)IP?Nuyqe(dK3kLe&c@&ZWY!7utPh$G8
z1wU%KbX2ad_*q3(>Ay0e&vb3DP*Hb;${K6rxo||hdL575L?)cmOT}FI-4RWfrq2eQ
z><l+y?mKu2wpG=-GBn##4yyT{kv7zp7B&PM{S|vtqd{^kU(j9%pCZKr20{x8(x&>-
z9H#8!AFr^zuBc88xnUwHGjD6{QrL(pEVz$#H2YO&*7S<&8ph47Ga0;F8Xp|bzA$(j
zg~w7bhIfL#uDhk9055wOE?98sPv4V&_x@dN_2{fQiC{XB<vUqw%kX;A5dy%9Xp1-g
z^~LPhmqPQrS&TxE=q92L#$edIy+o%(XnI8ZW-0@$etE=M7~J4Y)xv2=ZMjmLZwmf~
z_n!aX0`X5aru^?}aRSOO9697{7yDN686m^Gv1h)AkSDqX#q^_f|Ac(Z1963c=HyM#
z^z!@_@uY8vMdbJM;58&yz%`TaN%6v$J@t)^ht40lz6&B{;aBT^&)0X_t!$z`*>41z
z_HWaj8(8yHjf+4;u-_I=-q)3vd+G5x{}!IPfPV+f{c=%<EvW0)?xwTbk5Bw~JYILj
z8|ubx!6EnXZbk3O)0nQ2`A&juD^3K)zC^cQbptzL@F5~AC4Tc+$IZ_TH7AheI(wxu
zyZ%J*iD|<S&YvaeC3Gc30H7(_jlYu3igq;?E*A^yuwzlN@JinMf=3BCz>pm#GuNo3
zPJ@fZ&6U8!yrHP)(Q%$YiLuKPD8`>_QFWHheBCTeCh|FLS=iHns6KMk`m?7~e-L}U
z6+w0|=r#=lAGL={Op-=>`P(P$Djg0_LogeKNX>)Uhl+2i`YX=g<XaZrw|j_Ub=VX1
zK&|In!jyA$c@k<Zg%W2z?&UQ;49pES<1bxaW6*??zU}bx&9*IZO=`yFv||&0)$IKA
z(XLCgt?88jlzcw4=DYc2et1|8JgzthOmgsuQrBF<+AG-^Q5Fr?(-c+HgmWEn_>jy6
z=TDbeTxmUXZ1QHZ#Fq>Ljs#HpSqyOW1xD7noouTW-1BO0SCn7-kP%@ZZ1-jw(YN%W
zj%hrRq1rFFzj|3@SIO=3+i{p9kB|rR*%p|*lZ4F)JD#0{&pe)+-a(78_T#&=9CIbS
zFZ|rn=t2bH#$V?@!Bmog!ZcF7QAi6XQ928z+(<}Zm?@ee!+Uhh^raj2?DTk}I3p7v
zRrEKmoH*HbQaUa<+wPG4Sr`=3O1`||UYz@*N$<bXm8u70AJt<fUXq(RX;%0RF2sZ6
z7c89o%`e6I2-x2HmAymu$JR)54cUodx+g}pte#F?7$~n`Z0qI#DW531Gp7p=25`fh
z0AV&AVMUH?>wM)99=3=0tSZo9VPUAZuFPQU1JYbU1p{iFfgQ{$A|dHlZ@M(PCg69L
zdMvpz>4elMI`8+U?ud16-n@AOm61W@DPtvB5IsVod#z@#_M34zVoL;!KYsu`*%oEd
z+DL{rTs-RMj{kDG89(s(Rz_{_5@kspd6m-!$Dl9#KSn(=N&YN#nF>Ct@e9BE;-^ak
zmN{O5A0Ldn0{hd($|eu0mpOQC;F*)sGpw->PcZEzk3*P{J6}MTC`D3*p9H*DIS$0<
zKW#Gp3$n9WaaMuD?0kzG$%DAJA{5>zV(0?{6dQ{oK0tdJu8<l6$Pb+k+@gN^u;lUc
zKaH;kLwt5^RZ|E^8%_Rl$Pm^`_N|qGOMB_Be_TP_o-D>witJms5Eh)mFdwRG!F2Tf
zXeP!KnDkYFBL#Q+cE_^XXslDmNz8+4cSy-NodB94BOoWuSs)Gc!4Q{pIZ=$J64SE0
z3Wmj#;YhgsPUUJGZl7e7K`1a(?`2kl8*d+u8c3qQ`Ky_J`-OBDqJcJQ9L-xDR9J7f
zBqb1yd49eV=4KR!Y-ME;7Y<6|=m%06vI26Q1eHx5A$C;K;3-ZacKaZqKap-|;2)+R
zNQVZPBgd=hNhx~tnZMyXeG`*Va&MK`xO~{+TKfyX?7l>A1A{Z*C>1k4;O7BR+f7f&
z(&J^9NBday`e^%I8SQw~vwpL!4XltU{Px9ijt^6dPQ*Y##ocaU0)zT6e(q?!Z0=Ik
zayhTDhyz^>uPuXhl=H^Ny44nQ%aor=WY}Z~5Z&nNkEEIb|7Sj_GFXQ@-9EIo@PYE8
zt|5l(@bH<pD$9t(D2%RpHU4*xgx4B~1aMtnj93n<Hdqwsan{liA9T9cuKi>X9=67k
z9vktR!oQky&AweC@P07DNhcNNq-Hgs?eKuWZvTkPN?V)bKX2Wy3g6$iTBRz#0AHJ#
z`y+aLI++kF3ER2Lg&1X1k1wJG7RYF3Ujy6N#P~TyVDp6gjQ#B0B$N&{ZqhgfoX^`+
z@NW`aW#*~!h7Z*oH27W>IWg2#yNZq3c&rbE!nqNDTtd<U@{cggh}w*w8!bU)+dZN|
z-q?)u8}Z3lWQtZGRogT70n$i<D2Zv&z^n`|0kPd+i^ce#{K-^7^~DX8NVa`JnMgVf
zgHuVAK^?w>?Zq%tDXMtn(qoKZ@ycQY)|9D$59b(2u#*rl8mIgG$(0$WKl4{iD8QYE
z(Va?!MI%A>Ug$lHs9Q-5=%VN<h+w`f!i?=G&v9I;LCGaw$G4=TrBQb+r|7TlJJD~>
z0vcl^Op}(d9Y==fCH~PVh`IHOEhk0~saGLa`>u%+(Qt4$61k%X5P}w{)YCVL-YZdc
zP%5lb$#Zw7%fx&*hHdA;wMe~MvF%iPzbSP4DlTbRUE)7lle&tO>X(AyA|L{V`Z^cm
z7}%fESJdQ?It}|?I3b^)id=;_F#d$s_*t&)JFd>r-)w88wsQ=o`h-3gM*TG2vYR4d
zIZpEvezM1AomWsu@F6!>4wOuO&@4N!!D@aDF}k-fFqJ$E1rL530@KCix}3(Kahj9W
z>R?LH!%~0eu4-nbvWhh*^M1(SumnG9bo(7b;ErjXnWQc@E^25D^qx~!mrJ8OSB0jE
zwc5tLs7Np{_Sxl*Gr!C&g81ELRt&+)sHia2UoXaCnAxKU)2?0QWO>h8t4YKlZ?Gz9
z@^uv{tSpjhXI~U>NJIEPpT~XsSsho7LKTl>d)ts2a;{2GjcSej61f6U1_qCHLWmxy
z+ZCO!oXak@`;Au1O@=cC8CmAmYf?RL&M#BMhP?lxK`|-*6$S?4Z4k%7wRpfOZsl$4
zp}Vi_ki*SL!S~_}*$;Fgve|^|`NDpR|C+2aiV6@lF&#yQkhsSo15oU2mTuZ0b<6aX
zlUKJZVh9d_ye^Ensp=5l&Fp7=!=NS@=tJ)Wx(e9lA)#v6fRJuAQ{4p*78bfXdl&fl
zqpzAsq$ZXLh|oWBQKJQ~#?@w_UiHVi@z&EqWDi+fmug0NC1CFGM?3%aGZ6A-F!p{C
z+kmK#k$@A6b<bxXh@7B*Gms5BrK=zfqK}B1LPdrQ!$YxPU-tQwWyaQ0B!j${KnQtR
zh@rwD9*XEX*?NVdM8~7=Mq5B$)_sd9W|iemSY!x7JX>F=K%yhWcfb&=XlvP?*K3aa
z-V%e<AtI?ExyyibylPtGE5Qq2E-W-~bOmf<nL0#F$M?&|zazned|=^xH6Bb6eddv~
z-4~YY1Ca^D75jI%Caa){RHC44Gvh+xS|s0kmap5mR}kve_;9rLoE!qezyA?As!H1V
zd`cJA$H5AME{&Um%W5j}k4Gji=pb}Mg#F(e+$<{V*pgAmVIB6HkwTy8`(g#&@vi*q
zC?~H5MizrJP%<QY3&V^WERn3c^0g@Te;_f^T91EfxqeWP=kIHysUIG-+%2}$Kr`&z
zxT(jSdF8CGl_G})2+y00*@PW0bog>BA`h;pUq(5<V+$dyE|x7-_U9WI?i)GZBk*+E
zQaAW+GEy9(ECz2Dy$_?5``?qn6$;cDVc3=vNq^r^tGj;oi`ez}Db)Nuk^(${>-G-8
z!EE~&hpF?KH5Ufv#pSTXFPkYpvJk{ph|ppKH8d)Jnz$qC2$UgAOad$>RuVfztFa1R
zAmz!KL;*qf`vogCjtmCAV`>DkbuU#Yd?Dad(JRYAoOuR9Tp-k5sNhkMp*=+CBEGy&
zsXB^0@VFQ1gMzd8^+74j2KEoXxhy()Zw>9{EKirA*kR~KVyF8`OJr#J3k@V&guxhQ
zxBTXL>NxUy!}|~^Z;IIybFOLb`C?Z{t}kd8vDX5*@HO+<+(#S9)fQTrW*;m&PSo-#
zjuGlt|F>ap>#QbTI`je(<<woXHh4W7p1tkLkPqBhBEi4eEc^tUFrPRe8iDgBJ`AF*
zoZ0({|90@wUZWk-^eKm6d+<;JxmF@CymWt>%!84Y715aO^RQD{noNx~In=x|6S+8x
zIlx4#sR12OU6Pezi3=JF(vFho-7l9*5WcwDJ5V!CXR;aaH;fG7UtNlT=xIjNf_Ivm
zbbf-M+N2}I)Ns7GipoEI9^S3M6v}!T+xyh16#D*icM5Iks5mM#oGE`y`@Hut5j#DR
z776MR4Bi;CES|7t<;)yDId#&NOuC)*K!`Z|bF^L!WVtiYVhM!0^jH3;=a42fO+;@Y
zi2A<8-WM`Gikzd0cYs7Rh60DJxN(`@@j{$|YCgJTyDsP!^pvWVUgKlxP?3|X{DA7b
zNwxRBRtdl)uhw;?L;(#6r&Ct9go%4V3<0LgOx*#K)I?b7_xn%!5F2NvfrJMT?j#r#
zI*6<f_Apy)FX@)99R4ah^c80H-NenCuTA^N<C-4if%4j)i6Wu5-$gscXk>jLg19zf
z-!|nU*C{23ZibK=;(wmdO2HCZ6Mh>3^dR`BP>J~H>cJ63jivanUfbmV`5M`;w3%Rc
zRe`7&e6sw1d(diZgy*9>z(!AXu$%aTO)DSi9dIa%)4PbW7z+_hmjMKe9V0@pST+e0
z%D?;Oz!?7h5B^S0xvAn&zInrJ_`3=ZhJ-ip^1C<#>28Qr#{N4P|6%K``t%8P0;Y1{
z-?^GQBAuJiScb82g*NGh)f+9=PCD%33BQBPYe;F8I@&_k9@?$6D^;LVkeBQIc1lp>
z-@_R2a@N{5Eg4tHi%z9nhndJ_JwjRUCw{g%S@yIRDBB?#Mam^hQgwi&k-+N#@f~o*
zAJCC-5FqlCb(M&>^s|IGX&U+KH!xlK2cy^<EN+mg5F6mf`%soRvDnZHglRqVhhQP^
zyHoeklP^m5dWM%c<us-~h9qxsF>;~7nX2|QZD3#CmLxkFbu$984KVrr^gQ~m^i8x+
z5#$&Ajk}Y1+1Iy5%sLSnpETDSqC*Bx`y`c~+YxfUMQ~nJRt_9&iZr+58q`9avfwyf
z9w`>&CT4C$LR;OJttu{$7TuLw`*OXf&I{ted*m~t#jy3F8g0qmWtfZIMa|f_MJefL
zP|futI|`5sVn_a4qqY_$YNg^jx2FayP}4Uc{lDnfu8P?0jClJc8-rid@!Dc>`8S7a
z^c0>;vnDnq*ffu2ZCi;9Z6nac|H6RIdlyh5V~r!nvbZ7CM(*S(cd`&Y3>aen;h{Tw
zFhy}OGD4U;W!aM593dQ9S+4Wx4kA}Q)9~BeBu=9ESR(YFP=VakaGp)kUA5C)MZCZE
z)%fgpqM^FN<+6paUjQG>944>V_7n%H3ZYX0Z5!#kqah6tO7&QJUmk{8*IDGBduTq9
z1L3$4MG((a8<sTHk!ixh92#7M-c?F{jJm$QIsS%&nCT*i?KFatR7R*`Vw+Q1du$Nf
z;-CGBRFX48ktHY#KUG@hB;z+>SMF-5drpjGu03=a8GU^{V+m(Uv}TM$WXJ}0fJW7J
zf;5+)pIp?x{u8pXA?9uni`<346~_30f3LAMjJBEyUO~C*xam!0fKYyrJnuWJ`V2zt
zy54nRA-{KgkPN~U{9%?*$<tZXNiYBDuPw?;8i%F@-G@`a(#lg+w@Zm1kcMp0S0%xD
z-etJ#j_|QaLIGjWAqw2x31JFTg(7aU9QL1a?>N+6Z+zzuM%L$UtPGQ^ddE=)uUcGZ
zjElrg*Owiifx3XX8|&?#$BwTS+*3WXJe_q37t(<~tmm2uhCW4054~$ovx|h1Zy|}f
ztP<#W;uC0`{$HyO)M~nN`eJyljgk&&&LG$eM{Xd;q*C;7toTBp09bU4xOVWsEnEh!
zJr2;rO;9zs@tGXYSTG+5#=W>5zHDXj#%A*Lh3_m9d1~n~*7r3u+#sl;hD%tlcj#C4
zg==8mY*e4;U*g+opme;>YH7KMBrhF($=lNTBu5oE+DHinNVm5MDo~JzMY359#XbH3
ztEsKsp-hYb`c<c`PSGKcP*?Io%NP%aR&?^#B1yOP3Yscz-`$ZGHJ?eHxMuzF`t(tz
zt7tBn;KmYo)rZt)U31!zhF|ftCy|0EulkPdwj$PNmjD#t<6UMII3?RUJ(e`%ZnK<y
zkiq2PV=J+T%&_%j=9rKkbgp%%(8Yo@4P)7VtF+!r-euU^lj9(4Bv!o{p9Ag9$>|2R
z9~lK>x!wHw4B@C!A`>8d<S~SgTty{?;shA@Zc~rdw#^bZSFo?-MDhnxr+h&)`68gG
zbZiwTQy^+!Ec0RCpQM-J%7u^PC3}s^A5hkAzuDXYf5S@W`0`*z?Js1>Y@6SjSQlvP
z@qm@@(lP4J^}#mn*fWMvO_RD7ZF?n~r!vI)6AkT*);upVTcyc%x+k{2C&WI=_eXiA
zzqE(g>$6vC4<=iZY`8+lQm?1d4*kz%nLLC2CXOVM=k~x;&7Pk=F?HTHjp|0Vzy4y{
z9#7VAn+0q*A5x!sG6t~$VdAGCd<<|Z;qxK&?JhKU!vsbkrPVOo6?9A43f{n*=4Z&8
zgf1MN`zyDx7AJ{T+4wcY)XKFs^E{<geEn)5a_64S@j-9CX4#3D=(Wb&QsuTtNYJM+
z%JF;2>r0(a>x_`?OB1nv$6^`o@vBdkw%s^sF0053K~weFqO}m1Mtz2Dr$pKB?C570
zhf6?b=;%N7Uh1jr9%K|Z&bu>%0uwU45&LEcd#gzhHq#DSqkN0)EugPwGcBnW(p%lk
zzjx;3aP1$DxR-tU?4^BCl6G8fbs~_Hl_j{QejRQyua_zX+9Sq$R8xQGzg;Lp^3gsy
zFin?<cR3yfN6`c87v?Q7vd_+{gVUP;$NForM%pN9vSD8qRb5*A*fVT$*gCyQ_6R>f
zs+Y`1CXj@HMzlQpiw-0ag{*};aD2CX)O|ndOyL6fxATa!WNTZ4dV1Xu-$1VvxYV7`
z`%XMRQt$c^!S;1~4x5&VTPrEB#VAm;9MHm6B-En>lE%CV%BII@&s6cs%o|u}Y(O>I
z2um2I{NKa#mSTI!pguk~vJlP`(i2_qFJEE)jJ93EMNzrzw3R-tnXcCk62)>V_xw>`
zAmnWBv}*u(&EkSDVzX}BdkSUjm~dfGnLnVF{y{1J`MpFOU2vXy@{^=-mp7Kjd2u&{
zGK?P`KK2ZjpvH6`c<>^JDuQS&>%%(9MRiEy@rC^BlU3LxYL@1#`FzJa3n10)j)9);
zWYi!5QLl-kU8NqTE|Te?b89dahNuZ2jnyAZ$vVcnXo=2Nf?pSWGt8V6hXGJ+`xzDk
zESTG$X=jD{B!i*Q3FrOC;K`2L9e`iU%}!+XECe$LzRVDg2px8>r^D-)4tMj#pIvD-
zD4mF977FeZUk|>z7H1;5Gp>AB|D$u{{OQ9~BkhNpU7)RP6!7A!i!IdI&%j*BiY|%$
z!$6zfx=xv;Gs%O}AFf29h8rThkMFTF)&MC{fi>r~@d2+XquWs;X-0U$w==50Mf1pl
z0KTJW?(T7wO+L=N8s%p^HDbDYPNgbrHzD{p$l8a*aTa_tGPvxzrV@G(DfS*=k&P4D
zLIXPblD3jKf+jjm!HfkIN_G9EPxn<X_LGe$1^ZFIvPjXgt;)N}1-cENj9vet#mBE8
zZ&vOt1Z4-4x6q}nxK;kZTRHl+6<JKvYbnIVDt-9-tC9x$k+dhDuJV6Rx;?qY{V(Pl
z(pcHiS=t`drCyr^{P!@F-$#)U09lQ5Y8>3=R935-eT%HtRpNnv%;=`EA!BqAe#jOv
zxyEG1xPOyf2Ap=i1ZX^0k%!cM?AaR?lHI$>qId?mz&2*oK7W;NiGZ47Opf=w0CyG{
ze?0;;w;qHRm`{3tEFqrWLu5&g{^er}k0~MPu8`!qqV<YoGyBZXZ@hR#>&h0R*EovN
z^pJj&v5>_Qzt#F@`GDR6E^qJh{$+qGJO~m2V5G%_RSqz$k9V5d;|`aKg1A|(a*FI8
zjdprA?BT7*L|bnsn9dPTeDg6R-B+AO?Jle3Pm6igHd|X)h42bVy_cz0e2NgW>oEoQ
zPij(gYT|H3P66a0UXbG#!D+58dZKjD<P%yOgp9ejQ+%#Cp!<cruxBd(?tWPi*kGz-
zLE?fXYRA`$x2&cg;RL+gUh}XWP`a3B34Zc(n_Nxe#7w?8=(cYjJ2&-qGg4T-+DrQU
zYb)`l%XWotxy<@=1w6dM;U4uZy36Ev_O^noDjcU$LKb+ST{E0hU}i7^z@>M2eN@pn
z-<k_Chmd#%il78+_EIC$B>S)=`d6|B-MuB%V6#EDGS{-2-Y5k5dl=b|Ug2eOLHVI?
zua@ZlS;;OW^@s4#-1O_IaiZp<zalL(BfDodiIDkFUEQ(Da!hu(3gleLn+1Ea2qyo;
zIkoC-88jwEgI({7eNYC8^rkH|MQpw&^d;?}iPVpez-le#G*x1#&+_dQ$iH3<(l^kZ
zKq1A3h{zB)<Hgm5<=vX9#m(XpwEnKz;A*8XijT{#W#-Nd1(ut%zIp*=`}|)K9*!5i
zMkeOhOO8>qX3J~<Y#IkMIJ{4?bNJ=hOL#ZDknv*WFNn&#nJ@Rfk{2tEEJeM4isnk`
z-o1j)eql3ddeIlM(>}G}jp+M4wsmHgRk3kvnN}1YT_qB)gSX~5pEwsS<X=orUI{C0
zkNMP}_AdOubLOh1K3am2(EabVBei@w*g$^GJynkHs8QUxX?zR!B4#S$3`5oWsz=N7
zU5$@R<M||28Ol<@EO@%z5nq)}QiT)>*4|UD481;iKm1AR`5e(AR%H48J2I7HixDj8
zH@RMz$?>g-X5S!o8Up^vBI{Dw46^GhO887Y^-OKAiAGd3E}PAkF}BBn_nr^wkZN8j
zAKo6r0UiMd(@rbXMMBsZ?sES^Be-K^RqjfKIF&116ar%_%u7{D%PKnSpk7_)kI2UT
zR}@%M-o)BxN9~qt{HgbM;K&y=X{Iid*_@&*p_qysM0{)t>33NWG6lxpnr3#q%4X$H
ziSk<0=Fx0Q|9t-4z%WKY<Vi;@C9ulIrvrASkK?Kv$cOee^iM2i*V{W8kBSi1z6fCK
z<r@e^10P#-aqJOoBY5-Zr{Mb;pTBoy#*czfGs=&<0N(CB6vLZ%23{IyGC&Bd{C@-%
z4y99WWj0z;;gpMa5-90Iw9LDI)sKA*RoymuVyf-I`)ynRM6!8@>es{vPQmXQ4r;I5
zQY2HTW(-WLGqwlpD})dZk)Uu?1Cn>?L-Q`p&(OpRCX$h<^6&XW=ZpJmcE>1qf!*>d
zzjYwFe^X1phI?XHZ)gJQyPNwoJ(P_@sk-91$h%OWq#BMbThIP@n3i+teHVluh_~uT
z$#!4+%o;2%mA<V)iLUi;g=Wy>Es&IxCMv`La#3V(+rEI#m(K56*$T#Txw{H^yp2q+
zID$U*+7Q~sz-R<uQ(>X}fxBF#<R7nygHL1w1|BJpA;Z|cgY;=p6$(HiRw`PzEU)hh
zWbrtB4%53gi+hsa8p%Jy*u9bLg~9l4nttr~ze5aA<_yNzW-j9i?Jcy3#T@R6t&S$b
zDndy_uQLv@A?A5*dZHgNn}YUV5t*mhT`U}r(^8z5bN?DZ?R_(iIrUEji+>xWK;<81
z_Fr4~w|+84e5eFq{H2khYf5mf0E@XmlJ<TyCFoT!@I9uU4yj@}mc)~ETbLw>Xv|@1
zGpoNVXEsIOK^76;9dPwuGpu^gKUNfy5j-4eUxeM9c+NRE?r<^9M|X1kY~Fdq#txA>
zi^Co7KUEhT5(-4Yj=A~uv$Za_;jCRr_4gD9Zqz-)hmZh?BBy2>Ay}b#ZSLN9?BnGE
zYZ_6tmJC!!QHp<CtfPGHBf~U0KSwzlxRWFYWIT)4CO1VCsXxO8H@rN^B)s%>c>%yA
z<3$Q&y-|pf?eaI9T~4rZdN?XYYI>BfdL^xq-@6cAVez8UkmX!L3tKBKJN<=Jsag^T
z9hQ`Xk=5rR9h3FZB-7gd8V~{K%w)K&I-3+w+ASWbc^kQNhZ@SOiz)Xs7xJr4vg-{6
z-r`oN<&n_~h6<-3z6$TZp#U<c%r`oja_Hkk@dmxoktrGJ)KMZJ@{Vq@-ycdgWBy;K
zp~Nt%HblNSle}R8s7ZSq1s4Jglw{qvysv%N7oRgo`{;NrVD)vA9d^dy#@|VK8dgU(
z<sa++go{MaQ&d+1RY_f?#~uKeIFJ}Jmr0nS-D^<ZK=@|+sHkjetd_mr#97GKmh2fy
z!oO8CMVoJa2Ke5|j4xFrGjLrt>X$sE#&q?;9za*?FJIAKzVCF($h<saC3xTnC+V13
zH~gGWCNdN!MS}zj`AE>$3BC?(?f-k~#L~P<n&|nwKB+jJX^w{vo8%}MR%N(ptJDk!
zP9m@JTL%Nx>69T8X;TMe5_50*Xci&81V2PaW2EH$3mZ!vwFx87e<RK(!|T2dpfE!&
ze<wMcx%W*~a2v*0v}8oxEy?SuCxq)MD~?vCXdDqw9{Q;p*;=vM1tC9arTcI(&3dkA
zNyN;v(&gF_<M&ny`UR?mf}?0+q4WwLfBxNix?GYq`mItLPWH6Nof2&SaP-L-OYG^b
zBW}<%jrPh2Q_|^>W&B;$87-xpZp7Zz%uId_5n^U8=rnej--t}7N|WswPDM*PrHCW`
zGBO1}K{`dJhop3%*W>J4rTRw8V}I!XKR1C+`hU(%ya@4CesRiS&k%R!nA<q0r}-ds
zU3tV6^w*>BcX_$s<<1@<UfPWHTPe>qW$P85x3B~5@%Vbv+RLTw)44Y$EC-`+&?KPt
zm~KEl?*6hcJe;976L-*8p;J-Cm(XxI8jdbW)1oUhLnV@O>BitrCcyvy*!t$UI^)0Z
z##**k%eGxxwtcc~d)cj8*0O8aShj5}+pgz2?fKpJ{owqyz0TFSzWVt6d4E$U>ms{y
z>f?%83_;$lFTb+9r(+h;t<}*dcGwN97Q(0Z19{3Vy_b&(4R*wn`#P$3tP!@IT3B-w
ztz1ide*~D)IY(7&+-kj~<jy*`$#wD69$h`#&zD>3?4m>&mY`O-P2-}(G}GTb2M(O9
z%-wv?a;`!j@tmk+E)`S(F>mD#IT(#<P~a+Y)fNX<AKMI%A^Cn!#FI7X@X;L#)a64l
zue-Ba3D8A`kNwUAT(G2ryxmO7Up9JszTS#ejz(ppyA7feg3lj6wA1QuTzpR7!{4mg
z2rO^#9F#OFXPy`A8`#RY9=vaK4P`PqbBA|Mr7%@18TqxB{wu$v{G;c7S-#+|OWAQz
z_U7udz0qp;tT2w<VsTv?&e%gviKQt{zu80hY>Yv1Awda}28ex<3qPXl%`9PiqP3lX
za{hk3>@3hJSB?N93=165B{C#j2>{Ms9<ndiQXTr(8Gu0%=^{RN2o;@MuJ6^%{7fd4
z-WjhBIykW8iY-z>RZ1NyrIGv0PEZ*>RbXOHO5H?4!9(^CZ27_Uk9R%A7-vf6@mQhd
z&gbGeU#yFKL(<TvLvy=IgViyHMv$t!ZCO0kG-l$PRz{h@lgInKfo&&#GXFn{KRs1M
z`l`L`9WUIk3pqk+o&PWJyp<YD9{JPR)DVXB;mIUKC?aXtD6_2pA0X#dONIHxD;>6#
z%)b8hpaqQ=d_!!1Vgu-=rb+n$+*dO@br&KKM_5*_R17%~n;VfWu1ukO{XXlfJ!Ufw
zc3$(7p;g1}B7`=K-Bi8`w=h#O56<78+q+@g!8@jqJ)ic5PJh(t7MHNYM6K<6n9>Cq
z%kh6Nzsfb6MKG^ZBV{()`RUPd0jf~#;Fz$&HWLe?p&sxy)#oJ;JiT4!BTkOhs(F0i
zJ6hf!60x~?lmsordhg%W*;Ub$hhZa}3~1lnYu>1*-;k$QQf}Y25(eEN{}-U%SmpEE
zmM+U9(oBngv)gHcCV*heAKR6lW*##&u=Pg;TeQSf!OSQUwFjrHWak@$W}{Na42#D@
zofu~FHg&HWl72RC_f4Z9h5Z%YvER4C6?Y}1D){(0`9TJYLcM`K9V<_y&a6kbh5g4Z
z7bKASH@~1i_7-ydLzKI{BFa~+j{XBEzj&L?L_{J)0{>3wAACI{Q+B3?s%gSywgf4h
zb7>NWKJtjJI5RxYAf&Ag664?aa=hNu&fa0P+9>9*=ii?~(dk(z-cWxv=libPG82yU
z+<s0b<VW3c7vqO@?ppZuTQCd93=aFf*vnfX45x*AtL>^M99rU?wJ><X-DDPDxge&N
zvZZW@9(gh$+`=EAbVYEn1wwLRQ?zhy^<G|S3ZJc^poZHjM-p}YfItT+Yc@trKDtxw
z?|GORvN6ZxA%$AQ-t*;S7@uWReQvXM=&py@&ZeoIUYTj7>;aUO__=+F)5JqOJ<nM(
z38qKVJ<Muc{3RY=3=3dJfJ$Z1z*gE%-J%R~#Xb>mXhdo`Q=DI-GQUqnSBKEx>UnZu
zIIN!L^m6#yhQPdCGIN%?G~@9?;~a%4yYxiJHsq#<WMz?%ukvXwue0>1lY+wq=7=mY
zRh6#E9Mruofx(5BT3o9uYW{u}p4R${hkGwjyMZ*81Bz3X0QP5s9P3&M(3}09#HNO0
z=7IKo<4h}{`=zFX=UR47w;$nkFAH1guBS3V2(`QJai`Hat9ARm3EL%@TK(UPKXZ@3
zlfk_)+_lMq>jV+8amJ)t0MKV04Nf`D)x7B-D4zWL^=BktR`mLx9-#_9Umrf!&{2c-
zP2G#F2k6>lS2ULMruSeCW||>4J|6SuWSx$VqCa!+ZWnG%mWLr-Y51YX+)hoFCah9Z
zei81M`<ga<f8j-<3lJ=^mYep$)#rgFRRo$Z-lxebnY6O5%OjZ$Io}mfNTd(8#Dh+v
zqo+}8u^Q9#pcHz`!YD^cQ1=YRwbz8HHu+&>J=FLbRU+t8*=iiL;#GXybj%i#b&Sek
zH4@mKWxUl%!{Nf6){vioX+iW4`I~*71n(H3Lk)n5e9m?>G!<GJCs+Mah|gYg8B5Fu
zxW$=Th7$I7pX932lDhvE6f)vj9~2b59Ny(j{$1MYc?z8Av0lB~o)tlx;fjFkq%g(K
zY5Y!#)N8+`q@*em5~GKbicn&Dszgep?yRvKp5uJ(siNCX|MR2>P(tz2DSOP=PKsbV
zcWklex_F^S^nasP{pduDe>(<GuPixl_)h(Lb6lY3dWI7omr$*$VB0e$p>>4mDT?xU
zK-^Ns<4f{JhsGZv@DjfHl>mURlCqY=hLg!&%W@D))0odDv@h2&M=_epOOwT`qTA|Z
zw#MaZm5!TPxq72tD}v_MPY(0aqa_#lXNEN`sD`dU59jwPdiC(eA&*DN8N4nw4@?y1
z?~anYuH0>MkJNXMk84utI7DH}-6rP8HO|=(*c4heI!~Aqh+PL1yqhl`X5+aBTJ|Rb
ztB%>H-6Pi1SJodx-`l_)QIwJTYv6t9(pwXu0*;e2PbALdyZ()rbWSE7qTGjDt^bME
zoS!U5j$-;+Trzu>PSFSY%sI<QOMDg(#jl)JUVL+%fxUUaHb%Y875a(bu|5a#IoUx*
zK%h`#dYvRZ^z?{>MnmmZAk5T+1hUWy%C<J;ptwX6#tPJ5(QW+ypgu$Yp+0SE;X1X6
z!12yIpz?Ep4dQ^k@A!wafS~^SVbtQ&t)8ZZ3_=2q3}2+@Cjm9(!c8lG%Wn0-m&Wq-
zl<tjVg+?5?Aso5zQp{X3N##{Fs6CTeqzN8K73xToFn5|%vz4cXJp}=34jq^y`QhRN
z7M0DEava3gYLoAlIr(4FiFRiR3i1kvlXBr)m)Eg7lL&pz3}nR62#9$EH~Ks1z}L!+
z5m6na&%)!HjLCYE?V<^}X0)8{-ETU1GTbb+nva>g3sbe-Kaaeip_)F>@+<wkdEU%u
zTiwu!Ux@OayOj`1ni(tq_EG3L;0EvIx@LqR0VgLwlhaaye)pc*^hd*ggE$A|E~}IC
z9(^s#JzFZK8*<_FBi699k|fzTkLQw_ub|$BSNyyk_LH<co_ao8Gv)8oF3Z^J+{D)a
z25Dvf#K%~sG`*r_ulJ6B&VQHOPy%GCdw>?y?XrZAL;x*Sn%yOd`HJ4$6@OHW+c}U{
z8#qn;M)*apnk!vsgO8|Bq0iV3YJ7%8`%8h=js*M?e<quXWN@RB0$0~nWm(z_Y>2Fd
zmwLH!Zc(`<CrN(aST=LGBc5gon}P1Ezywg$?s1-YaMbGmbH1Qh=<*a!5sU`N96$(e
z`_UH&CI<yX-Ece6V*J`i->eBwaXyU~aVFo134!x|z$Zovhz~MxsLQ9Z!b9{jbm(MW
zV?u~_vE4ps9h@<&@V1|`$UxWBIx_(N0Q?EWHOZ596QI0$ykVl0CN3i6;x6<61+L3L
z+yr^qLo7i8<<fet?EdPuYM_0GSw0YKqJZ0{_C6%aj!dz#I-0z@Yu@`;v<H6g$J+V!
zx3w%(8yI)-qE|Q_a~BW@03RTwA!t`vQ6pZ6ZaZ)9`wpCp@v(hpECZlFHkzQ@v4C_J
z6hYED5M95<4&m9GqBvpzt&q?Fj19|H=(X=QP*<$Le=Vm&gcC9I3#`ECH}?|&khtjU
zMmo-6KWVa`nfkEX*xF7hC)>kNaW+Y?766P0ViPK;dvccx5fXDy3jP})DpmtryUXO(
z_W51+Oc<dnGZEVwn-YGF7ap1@KobEfQuq;o+tBeqg7KwH4n_LUpSjc6y6&Y**4q9+
zc0kYVx0<b$>{uJyn~C@_k{#tjyV~9sc>jJ9ioxU}RAN;&VyEiLU7{$0G=>xg(Qa-o
z1Q<w9g3EkKXdV?16DGZ%w|M_X?1a9aTz6wR1JU+rIt&Y$7PYx7WeJH5o@(34UdSzs
zA8UINYj?>6Ay7;xVpf*CrV(ZO$^pB9Kt5Cxux&a^NocOiqdTqqI66Z)g0zAF%0Y|C
zB66<KPJm;S<6Dw2cS6rGt|8871-q<rF<WxXThDjgw(j7J^eqY&Jqv*Kt2;;Kf1{62
zU?6SKuGj^PC~-^YX>LpFW09H&2m769wrydg1zh6@w&}`ZVJ>+`wZ$~0pJ{WCf&f+}
z<zN3FByx<0LEuH6TDXg9y&QJM%@rcoQ3OE6H5Yh7IKY4uoYI1*4zwt(%!NgTrR<gZ
zdF+gcW6Q?(%z93eRyxtn;dxlPRWNDKJA$AAb1g0A*?~9oxcShe-DL`gIjy+)XNXVS
zvjYj5xTSORo{k(_N>mhu@L)gjCOIrCwD1h)?SKiVtg--<nr{=>{Pr|($jwIQdf)dx
zx9yU{Wxb*8HL_zfc+NCudPm%{iIm1O+souH?U1(La?}XV9}M||uKkTWp`RGICnr=L
zm5X{4cW6-Z9BCNMrW=-%0c)XVfimFyYm-$w)6kbMY;G`l=W|-j{K0aQ<E|AoN3xL9
z>w?AHw;iCT<_<MjQif{h0y@^(<EoReu>O9;FnMlvEoS?a6p&LvK{4O&$l<1O?YZSI
zEoIH6j@rewJ@fSpvCVfc87&Wy<_C=rkY3dv)bZcwi5^;>cDSFQ^(IAD1u=OIn`ID7
zE$~XkGFv8-N~M?dBuPyi&Z$Y`SogYbe8RL~6@4ovDOG*t&OrPg>c>}|o}H&gZ~VO#
z`_a1<_wJ|tRCjB}2)&Bri<EG~HzCH0L+1P(Sk%IZ)23DPH9hgjRm)YGoN3P%zc}M%
zD5j)H!~944LlHN=0y~TO<a!c=soRktvNBNIan*t-in3|6!g(Thc}BR0OT2?dt&Dl6
z^EOHds=}Pw*qp{x%=rTA%jd8JG$gN<%=JnBEjrBvn6RJRL}<ZHtsFe^9xc<q%YUCg
zSvdz8*`0{Rl|W}Xx-tpUb8w!Ipbw8G{LNUrFNljVAYJaybf;zK;k@rj4t-I=JoqFQ
z-@}L2kK|BXRckeAc=heX(y3b3v<i6>54A$=%Sm2Pn6esN5Uo<H<vBr4$L?y)><@Ob
z9SmM`YzyvBEW3G^^`F_Rt*q&G?<X`I@1$#J=(&2!7&#hWj}3AN{d%F}i%K3Pw^C!<
z*Z5+K_>@u_D45#_eqAQtF5g7zy-TeAu-L97s82B~a_Kf)QvTJXrp(W9u^(^(>s;Bh
zxR`Jw0`@nLoe4LdWd;mgX@2}k!r=E_WnzGVe;V}PEm@&NB(8TumiVcC9hjS-E`@g4
z_a5{q3niT)&I=>{6mfO2_wh7Hz9ck;G|IT9sod2_a#T8FMPduXfGpbH(0b90c3NU7
zxMgrT1rl8b86cSK%B_B8OBhD_r7cQJ{@LtGSPs$L&4E1<P`W-9#A2x-)F0MDe-Xa#
zOc`LkWAQd$5MHO}M5e5&y;GIybd6mml7VUKUQQ5)^I{)KJL6Ph2}KAK&uCM8()9bY
z*lslQ4(kh6xtMcF^L}nvE}ha8T76i4<xWnh@r^Q-x_fDAa$cZ-nPr?vAceG~lW{Yv
zAFax~&C1_sl)zwX-J40PD(MuEg0rOgS^Xk``!~BeMLs<dEMPOlq{EMfT%_*xbJ0e%
zsaJER97FP@L1QI6`mkeb+t*erTOr7l0lu5y8?PaBPRAw}RYI{(k?)YarD19u2aO68
zv*g+`#|B%TN!M+FC)p(%5(Zy#t&5}ilj2hEB~3((b5*-gOdYOD7i0EX-%sIFc`DAY
z2$|E4?X?yf&_Kf&u`Y0eP*JMJp8P}<hA(oZJ{);}o;Ke9Ts|t@)WAarhpdTP!fA9^
zs!Y>M8{f9V>#y45=_0<=QO1{5P$d3R!1`I8PiaJ2!n=eBB`-G!5^^y7<B4nQ*3#Al
z#v(|3N)Yu)TCbqsi70)S(&z~*3=cHGk?4Xv6C*o&rfd4ltX)w`BsdM3++Cg<)a_s}
zvcF^g_LBH*97SV^*r4IWPEV-qEuhXIAl|=sa|=R?!0O!TI%||p{l#8kVpR?*V5#GF
z$t)ofFjtV(vLAXPOpKl8>9(IaXvzzOzjy;&)-Y1>=iBG*Z@71<nlEWwn@2hit&&R9
zXKW-Cz8Ev+r5ae)2b7oARN?9U8AXp1*+*E`DVs`^H<r1(7DhjB!e<>0`TN7Z(2&5f
zl(RM*NXdLcfc|p$<;0#l=^H~?XQyoO!nRtlbW7*74LTV)hip@?>#!n!U)HS&dPJb#
z>&8KdonY#nj33P>{ZZDW^=?|^iNU!S;GBhZFe;$>=6EF`zY()}yqArV(f_kPhSFZ@
zh>gdXp4DN+Kj|CHS%q?h6xwmKn%Z;FNB0C7sYV6m@|rX9s`|sLFV23HN~{gm%R|sc
zBF@@HbrT=)Wb(~vfvc7*-{h05bwVUl^CvN+i)GV>MD+k&7{CXS9CKwUs2#zYmDw|Y
zjAguV$m*nS*ZY_UZ$_<oO3<>x-Ku2<i@GOFIc9}dJmHTfk6}k@UgObu;EB}rG0a81
zyfZcy`ZL7z!qHn*DqTs#Yh38^8av~Vt}(r<m>#YNs>wr|ZBly_0){Uz98aSTkjSGW
zct|E^9(j|y^R6R5>EE>!kCC%7G{vL*>@j4tIeZ-!{vK`|bs!OT+ESY#Id8swe7S^T
zEMxcac&+c)DTu$*nqe=h-+a~l+ykmqB<UN$px(zHj|v|fwD08dhJRQSs2zzJ(UV0m
zC%~G0zQvJ|kO)Z2sdEge(kG5$j(=H<yB6QbG1t)CV$B{MMLbUNlU2$iZ|wfX0=lIp
zg$Lu!Abfn<Pq%q5D!)n{`-gR0+aejav7E8`Wgg48R{AYBI?XR}^i#Y@MXCO6IrWNV
z2r~Uqv;Q#pJecEOukbZUbM;34xph@#SrV^M1A{`r4_$h|gN@=G$@Fj4{9d`xZ%wt9
zJCo9J%pYmQ{wqJ*DQY-zj@WEo9abNouQJxFYA3QkH!j=W=bun9n6C`XHg_W}ea3CY
z_aC-x8-?*^?^jgMi+0o;Qzmm4GTeLWMaUn}iMdmy%J&|2xL=TOTHqITUoo$kLZLfw
zyhmfp*i0I}I<(0R_Rk8?G**aHYTbfbf@c%6X<YhYSFi4>ByK_fu9p-%bBLIg=#2CG
z3~RVwj68hP=FbWxcL-@MjM8@b!OLf-TDntd)5#drn&|YqSNRdw1W?hFFEVM$<%Wz6
z9A9`^Nb~tvY~cyheG5cxQKRY_!^3KmE=2~5S0*4G=#Utu=aDk+_VZfMLLCcj*+0JS
zHc&)qFi2(`dt6-Pu8r3;Br&F@c!@N_i(F(-Puz82L?$#DJrpFMaJ+ibB;oq4ewTtc
zq5r`_es&Xm)c42xn^pMy{)bTVm6=@}MQ*jBYhyH&?E8kdN^eqlBDzuKK2^-4(&W8O
z>8vBaGgYc&)?Th|%Q7;|@)zp$3wxdMN0`~m>!hj=m%b~kq}nR`?!tTN8Q6!GfZ~t}
zq*Y6bRi$Qh-l65&D8pYBU`U)=nYT{L@~>v&Zz6FkNg?6B!M6-3Qrm7JZBepbJNs#;
z;n(RGP#5$;f+lYg`>2fch&F8;)bsflE~yZ?C>k`Z^5XX*bsXQ<cl!mUukcp>g3z2i
za3!Kr_m5rRzgc-w-ANMXXqw46f$>``qLdPM=LxqmQS<^S-NZheaVlQZOhig+@~0}h
zyUX<;m`4nf{ILXO#PbqdjV4e%mmN)XV~=2Hpx`nD>GHZ%+_%B6noVETOzwePnudf!
z#QGnjvi4aTSlepOhm`ij*c$QH^rsJ}-9Q6Xq`<Lt=PqUu7O4M4bCOF&M%ofTca}8#
z%|be8d7VB<eE=KVOLi~pUDIZ%bToR#)FF)uwN!AP!dVsS>wIEX#sv_s@{{CRJdEhr
zRc=cR5d-RSMR0WgYIgL&2@B+oPju*zK>L0;#&UpLTt=8sBEJ^QEwMUL%B{>Z$niF2
zp@PjLg$x+h3xKS6kK2z<6*upFQ+<=;^a~G0ZzDad(v;eXGo|~?MVqMj&Xk=^+3&E$
z=6TcCs!djQSCOBLnLpbW6Ba3okEqAy_pe7OiM1In0&Zpi>oJ2FsaTBOEwWzldhEVV
za51NLHgN{k)Un8?wdB^PD_K@7EHB354v8UbJJ77u-^6gXT{sQ~xhnaF*xD~;hhUW?
zA}U9TTrXf-K2H@%Ts01V-E$i^AMLiJL=hUYDn^W>)tPMaQogbn9rS!C`{XElM(p2n
zQR2I7l_k69KUSzYsq@iz{k_;Q$9&zJC5$rh5YHWl@pQX8-+EENwCjTd7}y_dil}`u
z3FLIF`_o3npWeJ_lOw0*l<12P4N3h$l0=F`f=>*j*~CZ8XK-e@s_LJNXr+sHy7S?K
zhM!uO%>#F_?GJy07I;g`$vwwO8%SdCWDaU+^Yi3d8Wd|L$6NHux6#w2ewH9?hPcCL
z)huE&4~zV*$#64D_M9C!rgxhub5Fx{<01%;w3GZob9CD|)x6~q!A-l~Po7>NQL#O-
zzwceBz8`|Xa(45yvWbPB;lP(q6qC0Q%BQ@?iMK46o9i#Mgo~HH@`a7m8#>J6bNh>C
zY`1Ro5`J#$Rt`T8h_L4mP`altDr62l+G(RGqxnh>VD>{M4Opg`m1dch@TeD}vEyot
z&OcV*fBe4pvhy~7GKAd4k?`c63iO7OCRI9kSZj7LR)@nOIrzuNl;yD_Pg1T}sh#mc
zNq%Zlg~D<9<<%l~3n$X1qnR#$Rf`$7Sc=gN9MRH(pEjHk?1zpmmk9Q|4oJ36<xFYl
z5+rcwH!0PMw4jb0hEL-u;XC21e%iq7JWtLH_?U?t)!e0v6<S^7(iZK%pZe3qx*R#X
z%=|=KQ80#S7tUZgd${I{98|)~po`r8z}G|%sR@P(XREYQI}e5+lsduizj@R9i(z)O
zkq*R(o4Nbh&CT68*Bj1eZvXz#iFvNs*zco+L|KM*@ka%TvO*Gw8aws_>PX%Liz|8K
zlu~nbE2^0X8|Q)*vyGDnt<8Sen0mK1mVri%iJQsqX^u2A5Y|sv3`ih4mFeB^owzFn
z!cH4@z}pR<2)`N3PAPjdB?*^PZn^{hZI$+bc)rrQzVh31t2wdfND$|kS{V_grECZG
z@CVt(9WKH-iBlTM*l1mr9C>j}j{gtbTt=L5@N}ZZv*+wK{H}Y2bvl?0V(o#cDPn(!
ztmplekI@cOMY}3q3EV^hkG{;q^>;sS`>kbI$F@PGYQ@_>@6&tJau3il5`%uh+vjdr
zk|@Ou--kI$p>zmT$N@hsoS;48X;_g0q!rB2o9+^THe8KMcmKCy<u5ih7cNGN@<m#S
zNb8LZ5K^d=>boo_Q?qT_6N)^jN;+c&gD51cXiA6V(O^~p2uw`_IK%FmyWn@QQ5y>$
zY61<4KZ9j~)Xb@X{V>rE><w8r^*YnMmhtaejTovGx&7jxGI*HHET^9Xp`Tf{q#C<q
zzpw%!ce_b6-Js)!ycVYLhe{3N7E!mSKx+z<aSz#FcRu10Tkpv#Ff};i1ac9BTKX*t
z8L;h!1=_-X?^Tn08ea@8_NFLwa}Y+a1x|UpiCk9KyHFifG^z`gSh(PKF<J&K^oa1F
z3TaSD6Ti7OGG0rQq}DI17;`veUdR-nUazBAFiQbko7>qmScv$ZxTVicS8;<83mE40
zPpaAkWj=bAcVPbw6K#kU#TTNp-&YgpB?ZRX5>L`|SYE;b0+q8?3)=Vs{bs;tH@X?$
z6ClE)Hw&m#vnSpVm>$5*t!yIACBz55&BlNC#~$1<6&ZL!=?@_siY69A#JD9EaGzX#
zi^z%seo=cNF~n_fDB|UrP{5~2dYP7xwf9m{@O2!iToQ)ttgq_S12AN*lUhE}dWJEV
zIa$ZSk`1gZ@q-5V;Fl0Ra3dSXKoEU!b9)$aWzfYY&}{J+y-n?pCJbAy$k+IujRHLN
znh{g2q7>tY;!ssGubh7cLr254ZRul+J&|pG4rIFlM)34eIKE5POITG@&hz4VAbGl9
z9^6!<L!93X?p|9}g75;RQPUJzuc&0I#v^wA(#-*TR?;MIfQ(yffex$~0!#?Px-#nl
zf(lgvu^<sX@GX1XD<DOGK#8KldI$ZOHi0fY3XlOX#;84b$Tb>F=Zy;o{5a8=U<G+H
zhDJ8imxSVNF(RbbMF!9;z&0KJLaZd^xbC!1a7yUk5hom<1>g)mOkzNnUKkxErb3=j
z(34A2=~P{6VQC6FZln+J=GV+2CRYLKz`ntru`3DJpaHlm4MG@SV-g@^c6RVkc@wT$
zMS{eqSynWIc|IL6kHr5vVgEF-4Mk)ipdjyaJyZHXbUT(SuR<!uiV=T89rsThT!B>w
z_pg04_Okd(M>0(kv9R$`sz}ra{AT<ka3}d~o^OSpP(F=B_`F+#nY*cD5&7r;Y&sw~
zm<s5<A}5rw7&9y?=9d+3X@Sl}2p^z_9LP}?L-$1LeDsvlm>OZc1AZMrdQk8y@WVE=
zpMcIfU`YA{m0gLRXrVAgA9GhQ`{9Ja-vS7V%lpI&`5!kOr&+K%SiE5zbgXvXhDX=(
z^F95hP=w$WXOHeOB?kMVQNsF*hctjZua1lABUt=F3yBd>8qg40W-&d>!bd0k1Q-(J
zN-bc8xh5*q<lF=}HXu412H`Be$<ezTNp>+bn5h&V&Kw#DA_H$$qXY7^M;Ws8LdN6*
zI6Cak4R7e07{%-~$~Y+%N~I#UsuH2VZZIHBjHyuH{Y8k3>=jzRS@LrO0-}YK)u6<(
zN(fJ%+O*`85V7$=CIp%b1Zr#2u+<lS>kACUry3*h>!V}X^KV(oOvhGzO7^D?ccC35
zY~q$|<ct(j=zdcZUO(A(pl9#zX@g%Q-1NRd|NO$efKrWlOzt08oCV<J`Wp2%u8|IU
z)n8EsNmzsA^2_69(>c`KSHj0&<AOouZkDyb7uC?Y<WlRu+MF~Beo8~E;yu|s8r7;K
zPzF4MsGs!@$;=<%AzL|hOymknfd76}6HWC=K>MODJ6m1OZ|i7=lLcrwn-6e%%PCCq
z8TF?7kEiKvwleijJPiGmH=gc~X^ek0zxTv1Z6=*1OSyHdRhsw<!wg0OIT*+1k1+C^
zsw0H&@qnY(@hNgp(zSW*Jgk0?8x|Kj(;$y~<P-eY<VH>*@Y~udjf7(Sg|Z21!84QY
zlA>DCz~8=(P8QoNccbPtzPP!;WyII`&P+lu_~bm^pTA1WD+9$wVFus(E4l65ti$A`
z%jl>Q(fK;3*U10uQ$?J%*9jmd9l<%y#K>XWvCQ(S=S8?rE%fhA6qStzP)eKB@N=L(
zh-E%jk{)FiEdW^)e*mzt7HSJGln`=TTXW+4S<8H?A8k0ozaImJ*rWC#yKk>zF9<s9
zH||@^$0^xaYq-2B5N(F}gM^B5RT!de?YuCjZ#mln_rstd6QfaH^;Ty{^#{H|N<<Oj
z79zHjt@qvWie6)5{s-;^KLj&5D9&!|{oZk)-W8$lrR38d13q$ORHgz26eDo7<N95L
zehuwaX<rYy)D$h_eeuX1+FbdAgv6%7x0z?5Yi?y!rJtjx^TAdP_YWH@dQqcC&b`M0
z%VIqix4J}HUghF&s+7~j$|s9j&~G8&Seh)Y)l(#S&+gI{A4D-TXIL@-;W%ly>ql=n
zr~qMo$ck)10tKRu?N1+_cPPs%3H{}HMe<m1Vkt1Xgb235yFz?qxa#65p7@}oKQH3V
zvGBw}WOh||V)ejEbTQ9GMlx;o*#>gYhj9U?jPEEuy~1(^gVh}rdf9*q=tSH1OU!TB
zXabhPWZOk+NlC6-n1q@J{sPs|LR5!qKv`%+|6~I$sJ^5)xIM;5e)8sP%&kL1Md!Wg
z<uA`Tjma%06(WY>Af%1`cBtI)O_`9(>DP714K?Vui<cyhpQ<@3U)t<06hZ&QqEWXr
zEP@`Qo+WTAvdX>Y!^1N&jLNWhU&waY9FTpUHZUPC9Ud-SI!R&<gi?`;3O7J3&f8lW
zU+B>Cf=X;8XC+D;q2E7;?`y0h<S=`xzWub0<7M8VZ(Uf*=iTUGT=;F)TY@9ZI<cNQ
z9!2nQ!K^#P6WUSS#s3cu7RC|-UqmcDjTusA6;uS~ek3}!gT+Xy8bjfq%aMgj4uz^t
zq7tnHLjSp`yBgy&pH*D}Th250ckOfcbO*$vIP9g2)k|C?mPV5A!5WnQ@yoq~<JZBI
z`@b}F#12c8A%Bn6UTjNOqH_0=*oIskv^=;65AyQ{hMT@{|Jr-LJ~!mkz!)a=Pc(dS
zJxtu{FFuihI$qIL20do9rQOclg!&6|pYCQAXU}PItw;*Fp0qf68TT%zJbR?oZ&J?V
zwSwAu7o{3zN}pYoTo$=!n%lgtAa8!?S6r?KZI{9syQt+zUw4{6tXdzp<i1c#JzQ+K
zez~gJqOe3Enr~qht9m#qEgyD{`+6uu=XF*;ls%-CSnKJ$eEoIR)M7G?%VyVANC*m|
z2NOD=L^>Xs-zk3JGzG||w4-gn{v}RCKayUTbagj(e5R&tecD$^xQ;(V=AS!b)4cdY
z>~`~^^g1$KRQWQ{+tTH?agPL4h<M8l9Y>f)NeJ^*fOkn_SweVS%@-^b@0-zX=ep4n
zEBVHioe}Y*quj6e1F1GDax>_$2TwWS3F(@LJewr`{u1Tx&&hlumopQ6aGFe7{GwF>
zmv<0ZYoe_ppSsLgoKej`&+z2vgJB<=J(6Pn^uf7fSajr{UyE>5Kh8PdaR~X7G$<&S
zGS)5h;uLt=62dy(u&bv?3^={mT&9g8JnhJ;h4fa55g(9_*g6$EEkc%pnE@)CiI>pg
zMBvdfOfOxiWv2X_@@^Kj_)|MX(o<P#>>12G<<4sHq&;fOcMZ~d>9t!OI%Cxo()?4~
zT&9*0u;ZHL$Fd!fY=IlwnEJ$PS?k*9JRY4V$r>fhOE}Jl%SYQ9kM8iG&zB-{VYYd=
zJ#7<u8bJeV6Gun=MN-AZY}K{Nz=PV0s%mm%Y;#;xbPCD?`;>+e=Z}wgO6hx~rTPbh
zXc+FAMP-$>$NT{l?OdOFPOjg~UcL?g#tdy}81uK-w=yla5^&94oeV{Z$Z)5ZQrYzD
za*5ITF$3mkou>+>6y7F}p{MKdP#Wzx?(_Sinxj1j#if`ZOufl1hA>WCnd8Bl_I8uC
z#xRQt9h(Z?P0x5(xGKfM(qR}GMY30FkjKLpDh((Ies+ti&QFUiW!5s$TQGcWysew8
znxE#KyS6h*XRI53Ez$?gUMrvab9hv?96%i26&9B9r9^3RE`Etk*`c;A&n3U{UZp;)
zAFHcgpF19`Yq9CX?Wj#yFPtUISkT^Qq01T%Dg3)x8kKG))!m?HRWGcqJ7E3vkO{YH
zG`Volc7@kwwbxRSb2PAToI!VKOo1uB3-~-K+u}s7lESW4Wd*HR<A939bo!1{JD@Yc
zV^@XGhoRziy~%>$<+^B`5askDdNt$a%tn<mdSF;$VcxS?|FVadK~den$9eG_LUfGP
z6|Hbz=(l^U_ErT&()hXNpme*8>5${Li{l124-=~xxyyeS{^!s7m5mWe(2}FY1&cxl
z*M0d>1Q|l&!b-S{LHt3qw9yuuHG2A9W$^Pw6kYw7SP2QWiAL!1Y~Q#HYtO%m?%xIT
zClZ=IN(VWQ@atc5l+9f{wI+0&zV?VgNn$&?>cJ8ih={R(8x`nlC-!^_*&HHiSZ`Qk
zTl6MJUk<N{8pUd5wz2C6*RPpMs5#tx-Cdl7o_AI_1LF9!sO{Q<O^35WN=#(357L6`
z4b$0n!bR1I9iB8b9tC<+9#v=6B_CuwrtEQZKX6qaGbdAAs!vxzW^kEdE6$|})K=)U
z(*sRFL+Y4#HQ9<zh2Ip$OC+Gy3Gjin?ab!)?C1D7KcWL|T!ljN_+FPjN-VgW-%gL8
z8|GT8OiATrD|^9+Dtf}BNS;5ZsZ&8))^Yrxg!0bxUH@8r<?hT|OTP5t#ro<N5jK|&
zi|-dAvrs!93MtCA{zB3v)L|rIp&VFkP4!1HT`tSxQmspG-rx$GugnbC?dY(5D_5E2
zjD81h6>f*ln#3Rzw5E%;q$~7txqtsE0GO&4+fne%+&|4tu*TH9NAA@==l7DHVm)WP
zOnwCE7cV<f>AR<8)$Lr)4IA;Z<9a(cI~Uw;Oq`}ha<kfBMfEOz=@)2o65pZUCfaG4
zW+L?3l0OVv^Iac?(4U|1$0UtN&Noxsg?;T)qM!5FSmv~p(ix;ybU??%Wg1@X7RX9>
zxvIrTAkBp(-h`tz2kxlAOB!G<)0nk7!89wST;5FQBXlaPKZ=^08mSn@fT=?$A!JP%
zZNp`U<awuI3O>+-E|NWubA6s@EQ=BbH!@NpUx9@WRQyhT)0d0ufZe&z{NcB1gt947
z(h=ElY9ELE4x86fg<sQpAsQ7&nyIm;oTHXfwuXMK>Y>5H&;#QV0<+!q^VDZLJdP%h
z{mpmwa}52=`4b037V}tyDk?M)e;m;2PwF?K5A^FTnQ}%jIO~{yG>DZwur5k?m6l!A
z>MhO_(LELs4QVN?B!|Im@zzzhl=9Zq4H~}-MFzqSh_({ZpEw0YJPIJ^QU+e`#0Vc?
z#IQ&$Vh?+nvrq)HQmgfMmw6EJOY-)a(bRkOu_#3Lj{JY!(Zul*(tflKEYlUmc|Hxb
zzBue{gf}q|@#ctl%{=AKY7F1JkNoCHhD)%H!_T~8q-VR#)>H|iIj%#{<aYKPdc3Z7
zF<Y3R>AgL!7*?N8M;^VO<6;xaeH0BD<>O-aH2-P;l>~nn_Y?R$hr__>4Mzv#N>^4X
zzQRCY)950bGc8Re!_upX4;H>Av^)j#pLj7iZpT)5j^K8P!P@2~#mL^27X8(B8gvN?
zYwq_vF}}|E>8iDp#Hbs|wWl<_?MD2#C{cpz4MQ?$oHJ~{96sPg`vN(N483*yeDZ^$
zkm$sU6ds50mDi#~G)9)<Z@tE}MdfRAg}}PWS($yDh+krA6UK8oKlvLv@7eixzy-4N
zC@(fgOf7#~P^>hcrzH+nAEeBBbKB5wq;g6Y0*{$SQ_!#`4k+Kvinvy6G(WhC%3<A%
z>~hTgtXa>S5}k$F3Sk_r=*>q^nst7}O0_>Z48iqIp*=iU_!3vU&=T)|2pQv9zPbk5
z7x+XSi5T<=%zi<sCoL<jSU<eJq)GaUr+i!XEq}pogNMJ5#vbp}mcO&Xx9@5Dt8>TW
z^AZS2FFRbVjco1$jAWHp3fOeY^+r#RQr0U^_Z^A9XRuUDB0nt8GBGlqq}EPp;)Ckf
z_ZHKobe)HnxXC6SM0iKW)b6ZlcQx!Mh;LTvP1AjLo|%a^B8{z^%k5#EwrYuHXu~oZ
zU7B~8{HM{}6_;nS+%|2phnMGit!{TvrR$OfrNy3)@XD9Z>qrnPA;%CIf2cAnrAV@$
zsOy4?O+!$!@&nj|bjZGIttD&u*y=h~GS}rD+c+L6QOE<RaPawj9}2$SdyR@T%+{>K
z7fUQmGel-h;fG~YM*Pz4z2Z&wjj1k^-Em5^ea4OY{LqmjJ4oKf{;=elEmzIkL&2`|
zLk%M{HIWeJb}ofyb>#{E<>bl&UE~v|<5o=Ml3$u#ujA^ffnpIXD=nR5-*sQ}5gOjf
zOQX9Q+#+pCS^C`82=P722Y2@}yNhVti&@QOUdiN<!Movmo{wqGd4=QRFwpBkBBvUD
zt*(1WU1}`9;_t__Mv(`gr|-k(YCqdX|Cl~t9i86u0~PDun+`5my5aD2zUKj4Bwzp0
zBy#lRqsSICXfS{Z-^Xdol2&dT##g9r*xeymMd|do?&aiH<%^<N@$Fq_bL602O0o6=
zXx3GROj&dweCqBveAehlBXr*VyB6Ph@l8qYHr@-jSI<T5aLHQVoQ+sQ_^kR_#_)~y
zWu?RXb4*$Gw8PG|8=OGrXF44A!+DyabuoE8GNYax!6v|;3vT$fDI@V@W?hHul(upt
zhsW8LjZI1wi32!vh&IKsx&A-i#nq6e;DXIJ6qJyD_oy*1nZMm^2N^rH)(+$dLGnQB
zoY&FnjhCh>U||~goIE2oPV@PepC2@>$+<AbEuZh=PiPY^F$qfOXgwaz%2%uoRcFkO
z8_Pz*zPo!&@g`|F8&!g+TNPUycM+nzUv#wuS`ye7Qo>;PJol>}WNKZsul1g#YkCg@
zKCC9WsfATOmPc869hFQx*hdfL-FFPGz8oyJu|7;rSRIjEA0HhR`)^Bm^PLvgF7_wX
zNG}66qBF=hU+7IZUip{NrlnsEygXh?ptrEVl>Bt7K-x?9AYft}pJpu*pMLF#+;~@8
z?9Pscp-3TDe2~Phm}7xUo}g^0%(&}j(&s2mYM91v`o+2IMYe<Qmg6JE>}ELnnG3Rl
z=!3tB<bc-k-4Kjn5Q_Y)!Ni@4t(!p+xx|p^OGJB?88xL9?M-e==3(s)+~Q9uR_#p`
z5-97Yl0iNC0yAuBDtEQtZSiikvvUUW_uBj*L-x1)LvEI?-(`-j79l_1Y!nCr*~b;M
zX|^^e!i-V9=6Z*ZexcvLfM-#;)#d&G?S%+KETdf)^3B*u$}NGAmwt$Mt{&mrDKuN<
zTTBtg4*rezOOz4MqBVB59?Y-dlTLSyEAV$Y@}{3LUNS*I#L$!cHXMI<{|%-#P*x%F
zKctQ2q8|WfQ-7&b2g<;=@(0@={aoDakO#7&z!{;JmkdP}vi19{K1BA{41A<@bXvLZ
zPlQ`I2?+J4|H1+G7_7p+k^7g3MT6s+{xNVij~>F93VtX4$arG@GX{@Xt_xq9lU@h&
zA<E@Pun)~OncZ2ZRi?NP8lgO}&+2$h!LFKyM`Yl+#v)dF&Ah<u+<5uHK32oh>j9Z_
zok!ej)PaGV+L0A`tk*8m1Q~><jPa7e_>shnR_yUtf-j;eoI=%hJw?;Y5BtL4v9^mm
z&{?&R^9>uq9IKvU{v)Xpru`V#NGLWjGf(xZY&|k3`n>y|5)thlnpe;K_#e`-W)t)}
zSXzi6%K)Dsf4f)Mk}evJ%}gAdg@XSLcwNUJgF2tbzbDYJPNq)9#cu9gwn4}v`XIdn
z>j2agX;2KJ6x6(&FO@ssN8VQ)Enz4*b`z%G=FPXAS$?|Y(&N{r16HkmBA&SGx5<=m
zBWaQ{n0pK2GLPTWbmy*|0(GAUCc}UUoSO#k0;dh;Wx(0E|Bs{HLm+>@8_;d^e7n|$
zBi$Psau7>Y<)H!X8JRdX%4PSebS>5Rhn}IkISg9{#t669^38#i+c9f~i%O}Ss9=c7
z&eD3Xp6E9SVc1JxFU$!nwsA~zIzwJZ+f4pr>q|qmpR~S(#wBW6=rQ5z0H8}H!4}jY
zqOqbI%o_wzzbCaIgU})j)nIksvpeR(m(ZgFlEhr&mS4M}^f{xYPUImX-SF3FDVB7S
zP-c8xwc{`%YZDOmWEL4sf8)%mVVdaIe`A19izCiOv{THRNGg2=q!Rr7Z~KhS{fz(>
z@G*gY&Tzf)=C%K1!z01NVBS>(NdpA@4Td9~{8m8D0G!4ZZc*x&=kNPdw1A!h2Pih5
zv!Oi)FS}E;*lu<f-9O=fN#4<ZL?P9f<l+aG3&JE0v11Sxfuh9xxh8A7=%*^!q3tQv
zbOnk3yz|u=x4ig*E4Us0^=&vR;3WFM7srC^mp##-!<NAmFYUKECMCCbL&p9;n$WpN
zbkF|xmO15wybsCbZs~R^C_)Dh!nV(-&dXuVu&J|=++H(}81f(QoBnVzaH@H>>{8xk
zR|4X-bVqQbFQgx!ir=8y12CqwCw?~vpzaE_s+gTH%DdscTSa#j(cpx8INm?${`D%|
z-@h5dS2B~)?tj2!7z0?3(#ZX*SG&zrl7_y&%m4yuRjmq}f#VT?xU)b}*Bg)vR2;yr
z(NHor8u}mIVz|6XmMieuhoflYx;M#{O9%I<vbetcCo?^l32y@?o`+V_3}1;Cdi<);
zf@I>J2Xd;|Xx(DO|LifetNMY**`&s>kwK@@UlYi!)G1g21?0$m7$o&A9v}Ns{C`|k
zJ!`Y`aq4Vs5E8teE4wy~f}cPw#hNz$g#0xd?^AKq7No_J>^MJxbup;|`CFB&QVS5R
zO*V-3u)7L54bqkpG3Ub9fYYbK?4`Rj;G9XLb}jNl)Q4IF=uG@(`qbZp;Jp3|z3G(U
zCkAkFuD%k;ZT67H2|o)Q_xk@>+<A(`L&cu!Pz%Qw#0*4nYT(szVS;d}Bj<HfNop==
zZKzbk!0QnB3hW29E&CxJuFJe0)3BH(1BaJ%`(V_Xh{|*V;RwA2;aeA>`$WHVEr?Uv
z#BbWDtM^9@xM@ILlYZNR>qbaPvm%(R)%4m=K+Y@P!EJH4gkBdrj{?N`t`|&ucDh&z
z4s6`M&C3=u5duxW^hSyWbVuATECKv2#_O?%nOLZtJf;jAu$`vr6NCWjmk#1KqE~My
zR+w{q9?-J-!T&=Ku#t)w!HuiObYaa;B<=ciF>@=_IvB;UzY}2tYv^D!F=XsGyT-h-
z6%&9t52&C6bxNKZVFDx)tu*_cA7#5Uu2J-R)bDWuv13OX9ao6JMK}mdd|gCyWl8_#
zy(P845iKCbzk)Vs2%@vhpTq<-yHGy-5sfBEhd*RIkJK1kh$%T4f8V>o^mr%~8wYT-
z_#Fg@HgGLPlDK~DD<h!We<=>91)lNDv(Dc=N&R7XYF8d0Z@Nk*$?qA74l_i>=^07k
zFntf-MbO`T<?CAdRT!<CZ+_sM&_8QTN{A0EmAI$vUpPw%ef7(n-1%9!MR8+Nx)3B%
z57)n?%|)?u?-9TL7ufXPcf)}b31CA9)d;`bbneVKEQCoI+hE>S04{=XhJJSwAJ3)t
zH^~w0GH;Q-2+U{bpM}seamRE&wu+dcV`~D|h;bgJm&q-Jf2bnHlBDqxx!T6CqY#`S
zNo&3N<z4aC%1O|qwb4{NKu^;Q2dwI{CH$mo?z)=%e}J~RJ%kunY9Ba^2_@dt6WT}t
zduZTjM@?YhEfg_t1U5lBsJ5JGfNXXfFi1<O(i6FU`lrX*!wd-cl*s8-Eg_4|z*Rm>
zA*UTnivPHiK=T{8dt<D4ngY;v>2?9b>#y8&in7egj}F^R{C%1*UjPbRKvw`8>7ykD
zg%E*a!(ucjmjn$~NYSJJUxhRl1(XW;`^p#_-rc0Whwwo<30jImi$K%+{x=Zo2B7!>
z*WK>fNEoSOJ}X)?rzsY^4y`}b{Nftz@oVk`EA8_;zY(Cs1?HE$jN;!rJQPv^Q%0)N
zpm&l%Y7=z1PXAoo_>YSL(VC2fLIADzuPXnlf+YW=f<(1>?u-CS4jg<RzEy0%1t6vX
z5;TQiegPk(U5oB@h1@pAy+e0&z{gl~z<7NyW4kBxj#}8Ts`Yo-r`HFC{#*Zi30~_3
zR3F$(JofyOPaqSAB3OJCCZZvL#NCYfpfZPPJHH%E`2oBYcC4emK<8si{Wpb~oL72B
ze|Fhhh0I6T1e$vLKl;FMfiT?XyZ)r9#76cff<=j<*VXu>h<$gH&r9gP#CziW_)8Wr
zDPT5?U({m3M&ku{`=kvsK49Y!w`DWO$9Y=WZIm3_sAu<x^S>X|u}up9U@K>-a-CHL
zsIIc#mOHHlMO5f#U2MZ%h`^*|9Q=WneGAtt18>9O`31L{OcGxde;<f&VMDXwemkY~
zl5eF9wW^j-P`6?H<Q5-k_z}6~D3g=Or9%;6y>Th05}eXslKC@Ns!qjN_5GjuV*XD&
zjXV6U_is5>hxi|#qmelXBp9U##lMPrM5Ipc38F&iwDXQqaVVCiQlYJ~P@s(ih=RWn
z_oI=Tt@y<v<oG_MVx7l{v7|_LL6#BdDR(D)+&MGk%F>N{jeDO(bR-qAyRR=+QD@P0
zZhRobBL>u&T>VsicI*JH?L(k!%=~2NO&h>}*Bi5!X@L$<0jj5xD_Gsvv|H2=p(5H)
ze>}=0QLBm}3^I1q#&JD3&DAE(>f+P7AWh@k2xQnEw5v<K9W$%b>Gko^A4@ziUQVD%
zi<G^LxL&d!B><>`b#Ojv8k$6#!b#^dQSk=@iw500__O?^Ri|n*$Ql!wHRwXHL7ecf
zcu;Fw-PibsqyC#+<{oR67JMl5$o5yX2#&M{JA^>TdiBnrbl3-m-IJrwfV1fHM?H?3
zC%1r0qf{dnc~GA*P!El<H1q4S>clTo3x*WS-Nd&qzjv8+5vhcT`$7@L?{MA=&J=jN
zYrhe8`}4z&vw!gcnmAJ+<!1js-qk-PNkjuh8KDPYNmAcKeuh07uiff`M0zy&+v52Z
zxW@t(J)^s1X4l+77&{;a2_1Qj>7&5lcGy;vxS`UpTT8c@^V-y#M<ZwUD|B<EDqa;i
zmC|vM3y!7CJ4d5x7jx^@EG!vpzjgM@pULqbOveT4LZc9xHI%edv02UlqaY7um?NZf
z5ppx#h<*5p{%>#c<Vt~x%%9jsw;;gsLHV#yRA4AZKvJ43lS#}wsQ(HbR$;NCrLJ4R
z4W@rDZtIt~G1Z;Ih*H(P>A>t20$fWE&=jrPX(?+L%@89DKcz;Yo$#u-G6(lJxUZjx
z==4yTlT(!q_wqt?Uapq1YR|~hqLC1yg1wajV*gVh&I#R5xvAS~`99#%)<-8~L9kZh
z)QfSKTNnNJEa~P8oq}%jd)NzFHdHa||ES@V9d7Es!r|?KJA-x%zp(-pzEcCtg<$)p
z4g|qMSHlmKsDsFQn9Te55(-%A061XN;7#DjcE>rC)eo0P9_@SIk}oP39X6cWn%=#K
z5PXv(|I?)$`b$DowN4_={R=8kse;j)W!{G!tQQD=enF2x>?k&w!7Wp!6~jD4(*STU
zG?>${0{wc<7F1=e&f*I$<}MLE!KoO53Z*@mrvQfXpNbCvXAHVno3-)<W$0i}X!H5O
zRHW@cK(U0LTy-|i4Ia84s1b~(gT+NG_ZuJKHa-cI)_)_K+|Ttmt~Yap`7#8P?UA7H
z-?kuq?Y9e2?YdFQClFp>?1v^WEjs2$l<q7{t>65{+ypAaal-KRsUD6fUY~>3qw~$M
zt;OizX)&jpQirXccH0M?f*j?xr;}oE{4v@$RXZ`b!9U>kvzE%8kK5VRxX>e#*sk5P
zg(PBkJ3x1$wo0sb-V|vhR0s9p?6Nns2fE^`U1#`%$YW-pKHRH2FEp724g=+})JLXf
zq{86wR0*zxQASX8T&6zvF6+?aXaJ0}v)f0*FW<#;m#cOuKBk3a!@X)!w6D3c$|MrJ
zgH+3Ml%H)T%xl-tmFXpiHcz}v+2MlZin{e~ffSQJdwH@Q$isyau*$#zS_@60L3+{9
zpqa$}hY(5q`A+_8HKxvKD{pNf*y#mZoNYO~l68}Jm=p!T{mk#t)5JBS$qD=q$>ZQQ
zs&P>AbW<R7a=1XYw+4H>q-ma`S!9K%sjVG46}x{tl<7C<BS)l>-a<gt-`PelN)$_c
zuIYwLR*cqxpa3KlX$fSUqxZ5`_Tl2TbWdZeVJRqz8K2M!TYqy$WQEwjB;)Gqdp;vn
zTuXgJ7*Q*=n^7Dq^BH_PkEg2(GGTZ3yiqTA0pEfvI{jllZ;zTUy_jIr{zhlqL;^*x
zR;z45PCbos005BKV0!^Kt9OO$hzex!`}vxbmX{iZRjPp$HQOX~G3C&<m^eA<i8_6k
zQucasnXIKy#y&BaDwTWMuh%<rA6LKdTowV*bqqMVHcG`YXIO{X^h;c>t6iS3_(@e`
zxUlXuR~tLHD~i=2K-Amr4AnV}(*dP--<-QW;<8uNg~>@^r1{5~>qRYGBppoDtBQjM
zgo0ibWewulQ7AS590>Co1+>yRNJ=dQuX}TAKp3LN5b-!-ML|*O;_NIm-d9jQe3i%s
z+Hp@c?FVuy8fl*?#o-7y>Gyd4rvc6`J$@&J(D}v1<8FeNca{|mr_Xao=jrrHw(&-c
z3(K4@Ey>eT`rN;sU3n2f;g9c6)%dEpU^+E$RDJw`X%~!Yi|cp(srX`~zWqBDK($1c
z!zU#rd23=Xj-FUD8;Pqg6q_=<+?_;Q*AUq=&5P|kb>qfU)!vyL7#Y(dGkvnv7@%@%
z^SJ$TnLE+f#&?B8@VC~rZ3{9WG9i^saOnp2aje9VY9(E@ey`LX@{ELW4*hlkC${&G
zN!5nI2f)Yp+O!qI(55_<s=zw6A0el)MEe#%E77ZQo8RN*g*7%C*7Ye#g&I}G^&ke=
z<mh_P4nJWzyK%cqq?-y_VmnVPj&0x{Uh9scw_J%^^7dqfFko3mbxUQkw5n*aaGN{+
zSNtNa$xGsOOwK*?v9W=>XE-?mPTd^DM-?%Qfc%^D285h(H<lLhAwzi@;k-WYHjS9G
zcQGNWibb48iuAK?O}?~SHg28BtIHWUsy7mWhF4Gm2;U+*2OQZJK){d8Naon@a&&L$
zPWRZPI%`=wq}p$3uVCM?z{s%}%yU!QdU<y&<AQ_KskLk}q1{k~rD>Yj&+CD~X!W01
zX@&mbk8Km5ik8cL%lSxCazK!Ez4ad4dfAlo#NWjvrSg;@&YjN(FIj24!wPM#i=|u~
z{oKigT2ySC#pqxsadAMff1T2>EbBOVid~K#fN0g9=R4~^s*tQ_BnUgUP7L5Z<op_2
z^uQ?o`vM{vm6UKJC;Q9a;%E@iJ#WZFQS)&Nt?WlHxp|7di=I+uYxTwBEGqQXaX_0f
zlW(sT<MgN=vpP9J+U)VnwCB^tXndQ7yF{<7^|8Mx^^V-duF0(EvZxoch&lQAMjI4Z
zHmwBz940X!_5Ll=G&~1Y#Udg4$T?%DqhPd<3t@I!0*ym;{fdr|v=pNzxbMooO3Pmo
z7Sf0J>dfH^|1YAxGN`S!3%7Er6lp1LEm|Ck`@xGBcW7~U*S1)T7T1uX!Ge1T6xR?S
zKyZgZAh<)|=6rYN-v4`MPu}cV>s`-!B&MaTv2=RFFzd|Cu+HUA1h;7lvwKJNBz7nH
zM@;Z&=#j^4cRHl4oA^al+&b?sXl<E|d05zCL<iu#+Q!7MMH924YvHQV0+-30!hx3Y
zTOKnapc0c$lpY9bkat&H)`5O@vY+ubnWFDSe&V@>5Ne9`#aF+1Rq>F4EWd~8F0OWV
z*EK^t1J09BZ*aD3t+<Hz$DhhS3-a=g4Du?l!N@uPlY$u=ZMQf5-fIKvRq=eU1B~?x
zU|!A2AgUkGNhiMGphI&4kY#Vr8EfLKuoZ-dx=e8C&D@dL1BsvNpSYD=gEr<$-7~*Z
zo6X0M=U2eRjISJv^58ezHMOKgoXRgr>WqBV0To-5ok@?4)(ZWz)8Rlq0ns#dj=4x}
z5PCzx_Gi3FHbobcPq6)2qcV5ut8AhsCmAn`PoIfunzr$W*se+ohJHvo2>m9#V&(|K
z{OE~K!jFv3Pg(>~LYjW|4rY=9-pJM2WuK_Cssz%`93?T$t{%z6|LiU#hQPl3PCZA`
zld{<QOzfS1o9Ae?aX#xEf|51({e1IZ?{n5;R>4u46|Kt}N^62hJwK=qJ^b>4k(o)N
zNt~&6S2H`4sKH}e&boIK6jFKEbdyL%e9<>WkyQC)bSU;fG@bJ}8A^v=%@^T|4Ih0L
z2AoOu?YpP11W295&rH`bXvY@h^K2~Z_fCk-(}JrO7LA0SqMC;5(wL&o$=m631$Z}?
z(uxT?#9CU$9gil(^hR45;+6#plNwIqmR*PWj9$DzYXX^Y8+_S4Zzfk@i*U2^#%ybF
zAjWlBlr=&Yb1SST9;)2uOm-r;U+xYL-lC~79VMCldUWrp7W7U~&%9mY=(|zD@1Jjw
zY7j?$+w##oR*<@Led3o3pQDL0T3XKotfzhyUu#<?N9IS@V^i;A;9c;cwkuU!o!`0C
zG*h3KsnzvbPWA=$<zFkN@F}RIHb|nGs|@q5S|c$OTc02JeZrA!F?Au0E4ye!>h2nV
zvNa^W3%c20t*FFPRWS|@)Aioimg>+Rcr;E=XJDLBAwrcdmBDl<o`qmxsxs3LbwgA(
ziBky#Zr%I2a&uZQ?k$h}im~+OmT=UXLsO0DE@ML0ex_eXfDJE^n@KK0oY|7>H%~3G
zOGsa+aa5?x!~gg=Uq0b5ZjfArAx*r*CT?@H675N38%pd>l*H#f1kv|~7cbBgPe`^h
z@Z~wH)8Tpb6o$CV;)HIFPdN3=hfY7bTjsx6pwW7IHE%r~uY1Gq{nnz<*JNclUg=`Q
zZ$9G`>}`2_)Oh6xwQA#%&bwHY4157Kb~7?x3OW1XLQ%C_C~|r2XaifjSWxXTrL%A+
zmsp~k5kxzgI$d8}XzU~%wSHn&y|rtv4rXO>|HmTXnEQeqhtsy7;sN?*nzva+z4?9N
z!hpxwIg6f4F4}t0V>G1M<I};B;isrOU(Ji#YQDgfia;C8`d10PR=d^g#gAslQC6e&
z(}d&}z?3_;MGYjt)mgBvDKf)EtZH$ViPF=yBiJ<|w&}5~xa;_G`pM%ldLTWuIQBwF
za#~>BJNQlG;e{jJW+2CuN}xQ|bbIrRO5bi*G-MSmRBN}mTOBmJEAi3nq70=*K)W&;
zcJB_A{%=%inHrp^YdFIJ8u~0MVHB5h51VUQ)+9&XtKZDGjn>s%qtSFHj>_JQd;+V*
zrEBL1*#Nseb23)f<>i%mHW6;Hk{3{an!T(;gfrQ2mfgmR;}@pQq%L1(O2+6V`0?I*
zl{kFw(2kTG_9~j5D_|UvXtGk1*z-7C*)nv@_x#bq&z!!>sGyJmVItN3Di5>IRImjx
z_4jkA61PacUXWzBI#^|fzhhL(2(pzywKS;Bd$rZ)9(ET8!hisYMp9UvitFe3k_K;n
zX5^9{b4_Z>bV@db2&?AT%3H&aOebZsQK9b(+l)8~x9Vp~G*^Gb-<TT5l6Y@5K6w#F
z&PDM!h~51EZRb@XVxs&Zv2+fSoDOKc(KZpnK1gWA*ha3Pw`t;GXOw5{LRFSSJceDe
zgeU3pvZ4HA-8?a0)|b)82#vPif%SZ@ssY!ON~jMnAnnH{-mh1!tonQF+6s>)J_E1~
zO`fdM3kV@*pZ!)FSdnGX6G0n4<-K}8=^LH^bsMiC7q!7JMjKMeCb{}~4N8`vKF`Je
zqplkkn!wGj9m#nr%C-1JU2V2uXh|8*Tt-qx3*Vb$Q(#X;rZDHdldZ^J$#Ugu$n@H0
z8cD3SH1x;da8_oGFuN-MHul+ngGwq9%ZmDQLZ|`q)KMNbH#w6hj>}DXKy9wU8RHA2
z+n^D-p2>_)^lvFEzBl_F;;Er$Fh4x-LJ4s$v~gXl=RU%S{JXd{(T3kqpKvp_b@xZI
z#cveI+z51@f2p1uY$|YAY#lAK-$8XcP0Lwy1U-9e4?Z{3NU+nHm^hq6ff^}{%^8m}
zy^Yowrvt*`h=!~BkIkA_5~t0y+H!!m!*Q$_+uEd!?SXE!23p6CU<DLSBbpoKJTprh
zuGRHC`0AWOB*>v7dvNFI?n~oq08utz%)V!u@%~2tDhlj;u+JHaS!y8`6FqIlo5+~w
zG9~~HN(B`+9CJ%m(}SGx&?H3{SpM*CpLXHn$i^aG0f)8Pz1{x)@oBin=^rb}C|E%g
zvrkg>AcBvS<zzuBX+}Ng&b2A$BGNEsp__&a;D5fDe;aR}Lg{&7-hMQ@>*Cz;SkG^$
zHGF}Agoe*U&qi%$gl$;rYPSX`YDNtVS=QCCIog)2AKo!8@2VQ!G0}G(yx|5c%pR>h
zT2)-QURk3mATKmTZg9E%%Xyq&>bs9Mzw>E<nqPBd3ZcTNh{Zl?D52*L@mjot8=6XI
z?dUrriBU`Ut<5K&#S9zHAQLy{Mm4Px2XRSU8$g(Mg8SA{c%>9RB-GyxQjCH3wcVC+
z8+I|_^^Oof&3XABn4@W@7F{o}1J>cti2gcm_x$<a0g~i4o(hJCe5@}QXH(b}Q=myy
zC-5C-R3`s;gZ;?Da9T)=YV-y~U6@0R&{rw*oR+mCbS<UvRb1}d!S`(4-`HJhApL7u
z(blGpy+3c)@>mm>()oh#&+?jd&44`xp%8|3X9->|L}PswDnjF;ZKZ)xe#ghU>nzdP
z*}6_H+>05y7QVATKFry04?!?IpDqSTN}etN@V26~qMWH3Y}&k2eM1d564M-kBA>DA
zbMebemWA&(YUj=p`-vRph|DN_+x*}yLj^d%co8pD<7o4!#!_JCj0lJwI9WW{uy(v(
z&0m=Wc2r`nsml*<Y?~D%vvrr-jZ=`D+TCdyC#YL42>3oHtBr@NBEO6?n|LV+<_WWu
zY}%R8OV9;eyX_J;nPbuX4;SwXk+UaE<{k%D$Qgpb^`&(6{q0&*S%=?FcDJGCT9VP(
zAEOQZzH*UkB+&dwJUpH{P=tM}==9tvHMRDg_$5SC_5iFMiap^p59Bn$-gRF0vd<|-
zR{M*uO$NAh`^2|?R7?Lilj;KF(sQ5pw_5-=6j5u1|FC**?pi}rR))Ff@BH-waj3Q)
zSJ%PH8h(KOREw?cC_}vhd9Y`j79GvL0ufrj$!4c%up4!Rgg7+J-<Fz5v3mfQ#vmbY
zdsfaV-}tnKW57fn7O0(r*gM@nddss}r0|l+mI7tA(e!j5^Xko64Gp~+Xn!7Ri)uvB
z1F$26#SZ6ZRjSwzzugyg^RH;%<@tk|NPh}IXA>U~$++a6RR5;prHya+t`Ms}f?Arb
zy$f4gb5V&nDy%%JFAihr*fw;!>RsG#iy3}}F<|R~zD*><mh89doY_isa9lKcp3l$Z
z^2z-YZE_m#g)OC(jaI<~@GIp*!jWcINAN1UR`}<(j!A6YTQN2co4<<EVR<LTQ#SAf
zyUMRDlb4XRWjNibKG3{>?lK$)fRQufL<z3Bt<4~l8ez#S^<Hk52Mm*^eiuc-r!93_
znMM=dpR2<T3uJC3Za@$Mf17Qu_ad9K3GX)|UP-gJ5?VDHsfC|rUCgyX?KQC|50wt~
zwI?-*2x{K+PWm1v`V))UTu`rzri`y`>84yMk_whZ$;tS>1jxwtWq!ay^HMYFy;Qo-
z|6d_wNb&g(l3)L_uin!Oo0Wc^I<*;3UeUwxb*5-;ADXqpyN~yCMw?lpXUtb@yy=Y#
zL5CuK#3Ns+s#Te@C$0^j1?PW;Py6B+YddkI5fG%VUe$in)nms*Iw#ShCV2<`xbPmo
zASvPCg=P*tr&m&3j2Zn-5(5J%cHfcmXq_vCq@^Vk4TssuEyk1E^mY@~1@qv$d-A_}
zz{M`}VbjyfsGB#_LE2<;ESLGg-){-r3A=>!$;p)jJ^sO^wX<?v^Apn70_7ilLJT(o
z`j~F6%C#1nJWDHu@V5?aed61!ZJ4V8WU2sQ!v|zX&AddPkz<{$M%MKfy?Y0EyFt?H
zZe0a|Vu_68DQ*9Ug*SFR(^BjirLOIt#@FC`XB2f`)waada@vQJC!JmMTUgR(rJ$sR
zdw)<{CFXnOiL2@5q1D%$Rr-vM)7fF5RMN7uO9Fsb+~vw*S?@J5<$j{oP2CiBIYI21
zkAZc1a7)TpTs>V6tWNLNFCRiC?VWpx(tiv~G1&(}>wh6f+xQeI;TK0PfH`i-!R+I)
z9xJ4y^mVu~f@NW@P_%V@qd3{<@p;gHh(t74=~wW9>UibHtG1c{w5~X8dY}SHsF%G6
zdX!r9jGwi9?pl0%^@vXV4K6*XHTNGi<=2n{a#jARg8@ufscOs6zZPQA>z*?`_UP*t
zQ^&J`AI^`ZKnc+r!S{2Fl~Q@<%}W`hMS<M2BA3+NRn8k4G&!=FXG2uOO}Td|L>|Z=
zINr|^x#;90B34n($sL_03GUr7#SW8J>qW6TF|MlA1s+fPrC(5*lLoA!_w`BWl#Q!>
zRbKC^B)Ne!zilX4u<#-VUE3Oxwrf!4`5H1caX}wur+J@UyqtxQS)dv4i5Z=3<b~j3
zvu8$d^7p%XCx5qlM~q20k;hIKPKZwj3BAt$oZwEGx7N~BnGIQ4Gv)>^?h@dHWw?G~
z<6xDML0tH;eq(wP7W?a89D%@RbXU0MVuKJ?h<r9-r1nggc{zHGC1C4k{fZKCt3;f_
zeW$N{R;K_aM@Dw;U`g&%*W~FFMGfjx5d!QI=WGG%xNCq-8|iw4hc_-oe%wiDe&9AD
zdEugTKP0op5hw0>czzmC4*%Pf=xKF&I=45SaNgnMYd~gg24EfRBvcLN0^K<*6KYqP
zFNb70+6{TnvYc-Pav@p@F}F481lymg2>Tk@xdn6})5t47NAPSErO3^VVPQhi!U3<J
zjLalf;p-R9W0sz!0%tw@7xwT#f$LV<)b{{5$C}Aavu&TFb2H;`m5YuY2WdG#6%e8!
z3hAfc2L=wA1vc(AoEF|US#<rdl18ORTXhu>+D8ke>RN74{WqVcet26eYIr>uv4)c}
z;<T?RUquQYJ-GKeZ#3dc7XtD>{WDOWuAx-qpR&A3upb{Fe`YL2N>fkZ!#OuJ_cAa@
zCUCWW41x-$7N4d*+^|~Rxa2`D!$=~0UBGCM2p5mJLMv{Hl4(2l#mAXk8{<F6mlPMO
z1_WERLd3D|YT89O4ns&^u;spi$J2J|Zrf+ORr}5Fc@jC`pN6?xH{5x=NA*jE4fp3Y
zv=ow;ygS8W=tkZ%2PYUBUE8NGI$A<Gdfn)5>#K$1PqaF^{hm$IhY~QABVli8`Cy9(
z$GVK`=3*fUXaABLo^<D{`S0-7TNk3h;DDSz2iR4uSxEZO(k5P2tN)Uz;g5LNgwxWC
z@N^0NsC&ilVS$4W{#IwP{vvu><KWeYH}v$snJ*{uLIP0X=G$CrYwP*QLw4K_=_#3N
zo4^(1*k52y6Y7m>2|W<MNZ?d-^$$ei8%GluJoZ3MY*oD!g(Hb=HpJQXv9uudgReV6
zCcnP!Q%TqhI)f)J^_(>0ZuW8|MdFZk24+DrdcUjAC=LbDP2SL8bmr1Ohw-&IB1hm!
zD2~W+!4VLxWE<CH32Sl_fdsdraIrqNN0~waOP3c(U5it!896yy>*%uGg%%Pqh7cJz
zg@GoO)=cnzPVBp{@FMz3jlCKjKxc5c2MePry2<6gs{Pug-`cZn<Tnj)JGaPr(yJ;;
z!>dFLl^W`~!j}wG%n6*}zdK&iNT5c}h}Q)#Kq+uQTC5&U<$zP95!DvxkRB30QO$6R
z=J8{K8Tq;@JE-@mN|9PG3$m=c4zq~Yy>qci7kf~*ZhFBLKk{jryt`6I;NDBAWi!bm
zUB=k*tf{_zZ_a%C2-elMnpPC5d@gycZ(VnBjztL7-Jnfkiu#FXWYRvhRaMGOuSm|O
zt1Dw@b=?<ik<2y0mGEy(<14;?f<j+P?PGcm2RrqjyPFAYPG25Qx?d3^{tW@}Ox^hn
zLa)L=m9Z<$Qd+xZi;J;GhE&AC#sDM34O(`h)b_j^t*-g|S|^d=z~F_&#zB?jwN~y1
z4>h-hqT{k-3_zf*!4{wCC_&P*`v%!UDNs#%MQk<RfON`&p|2BH&Pm4i#o&X4>AC^)
z{QI_*E$v+ut$q9D86vVAhwU^w)Q#zErEFPew5*hOp#R$3xKlfS65K+Ee;2o@dX4{W
zRTBT;Vz&2qV$r*@EuP{oA&?oo;hP;*;~{hW%$MtKE9;;3{+r!2c+eMx+aE7#R6sU&
zmf%Ut5bY2^QNiDL<EFc0_qS5zgm-XsSVHB(#rCl&At!dTJpyO~#AY+bBR~_^xs<u{
z+Uw&cTV)L~O-thTmRg_qp_a5EM7E2ClR_xP0DP5>8nn2tMWj^wzZ)Sq8UEcy-fX5c
zl%ukA(AO5qKYr3=t?CtA&%+Xk<ji~zbz(j^qu3Ta&huWfbz63bqUFMZGp4VZ0Z^WJ
zGKrs{uc5(C+686+qQv%qV@_cn?3MG{MEl2<z_1OgD$K1Mc7I$!U;jPBdK2=_n;tY@
z`<gg8<m9pyo@?mkYD?=l<2?&Fn@FN-MZfs9s~gaVsJab3-+wF~{qg?sWwsM+-^RCY
z#~Sg8f$`xR8ItFZNTZ*8en>LB9KX4tGby&aQ(A)BF&vle`9b@fUM=hNjk3)1o%c}Z
zA?EBom&Swp!L->$D{Up2p^KA|WLrM51F?fPzB;~+Oy4C#KbDSUkBt<RHqf%T!zX#*
zSjRziG6qmwwwNd@zqj>9Yx9d*v&80Z10O2T^EOC0SHD3H3L&SfKOQ|ecm?>v8u4Wn
zPiR>wVt45HaeAwvqF0$8JF%^%rC~l3H^;s4KlAoUzZG4VZ||Vl`U&P1NO!@nzxJW5
zJ=ZE<#?~<MSUfmxLjGboW`c@%;xf!RYy4SoTIJ`<hvW{D!hV~PBx&cdUebSu37?8{
zCZ{Ut8c~`uu3Y2Iq}4t9Zu^0?*lc%av<9FvAV6@T89EcpyLmLIS?`0&s`Oe@2EFdG
zI`+${Uis52zyx<1W!>g#sw6QP_4}@7EQB~0rOz|}8yxd^pFs|Ud@Tb)RF-)Vm{$ff
zj=Xsr&ZNFXH|N%9w7R)Yvb*Sg0G@MI*H_D?`*IhBp}7XyhM|in8^3iI;4g`y*{6Cb
zVV687n1J&Ml9R0={<cQLZ9!MzLwVpeoeH~Tu94qQC0nWA$8F^?yU*iH<FH9e{XrhJ
z@%#yperjEhM_cMeu9D99ubKc$LYAvCo|lTA8bN-8>eedq*4hQRTg+}3$L?fCf<k8X
zLcB)V@}GRnRI`O8#wi1eDxkh)I?T2wZNmI0B#h(XPO8Kvg19&B$PACnCAXjg$=h}r
z%cYg-q@og2H(Cx3_ubcF-z5p>T$>5-u;6uWK!TXioU0?5_ZhhhjED*O{qx{`0gTi|
zc@1JW%9)GQX7PO^e<M*s;9c0CdG4v{=Mt9m@@4?0nx|jdGQyvksTpYM>zfqVZ60-Z
z%Qf=lA9tCmH9OB&x#40sE5E0VJrEM*Y?9WSoLSo=Pk8t^OUPVXAfo&tZPu5foA|j!
zcToSc!01xDXrhTZhh_Ln#QtrDx~I6I4bPtOhjM;Vz|vSvFfW|@v|D~{-*a~~^Td9*
z38m*amjFWK$hUb6!VO0u2zSa}JyIL2nPuSeYvT<OM-v{BBQ!sNnY%fsMiSXR%-Whh
zKzf8mm&NP)AXErZWk}5?mHut3wR_s81DOF`yUd76SD^wc<sCv6NnBKH_Hi|#6p@%o
z;G)X`8VE;fQspggxIx;-ip<-5_rkFOtIf~kqt`7T(chL{R&z!@y&;%!sA{|J$Pa+L
zc1rw<dsmS?ic#Iz25@z*Q`W7}TO;2VV2|EmV=4XbIZNpXEIDcau;d!kby9K>lYWcb
zqN(zqkXVPp>8(Nj#xdTtDS+7dMc7N*(w|^rd&6GE(CJwpg0y$^BFCa^Yg3K@#`kiw
z7F6)_b6P68EIT^f3vZ-GKlO)v`-2PLlMn<5vT*!8wQ{9!c>IOuY(?hV_@~&bmUPtp
z#qG`zDWFbO2Rsy8v49dXtE}I(M%|_=IQ-~SNHvX&ZA#O4YpkyB9pyutdp~<80=(r#
zB}?#0&DCv}TT<~kp$@u;mjBVt1)!$Y24S++O5*?^=H}4Py<KP1>$Ml9uDkneAEw_X
zBzU!Yy78?8pkcP!tjcRiPyhaIOO~pRQdF)aYcKY6ej0=nWAyzne-W4v<kzTNd62<l
zIZ8+dH)OB?XsHyTKj)SkDK2sX!20$8ql3>y`10i!q(TksEee{Y{O79^6TbbDI%URd
zJ;}Clro#EYf|kpXTci}a2)5vIggt?&q@_??w|)cD!L9dv|4tGm2%mAX>^kSivg9>}
zCet#qUwxtJNXsb7rFZj53svsynQ~eKiw!_|pQoxzXj#0`yv#cb4Wr)`c$aEfh5*rF
zw<lLy;~~Y3rSC4HghEY%&87%dCVT%h*f4V5Z%)DBfV9abcwNyKBwa#~zSdsGkkJ=}
zN+fvQxDf+&4V0FS!u$95#{xIf<w^_9_-g|Kb){kF))d?6fwHpf4mUs~_0GK4J9#Vl
zxD+M{9+=>9Qesr1Mjg%~^=(<6wu*@tob=r3OcK7zg1csXEKBp<QVa~p{2{oMPKWLk
z6yTl|8rGFh+PWPL0IoI1`4%)ZHg3%W#cmYn%tZQv4r(s<c63>&#ENe^#9W)D#>9U)
ztcTv;x_Y8o&LZqzFHmrZs(}&Zx@GjG`7z>8sEfJZzbR6fc5amJvZJlTyfv&8)*5%p
zHT(PJy7je^C)uFdiwLKDz7pMN1}4~K%6DoZ8UECG>s@L8fSUN$;W*NPuRa{IkSs<y
zZbuzjdppePOX0h{N9dEpJ%AMOI<}v*t@Tu;J-h$;5>n5#HTrk%{(3hvrLJ>w<*vOj
zrD<?-a{8;xqn=#%ko~3N>?)iaNu^PiS@~{2pLh+Nu?E62?pU7{JyPa{UW@-tqx=bU
za(t9Sxw^kd+9X=I_F~q|wD8R_vqHcV-sj_J<3bpdz7^BrlnPzifXI8X6eYA{>5rQJ
zc#>Zd85}Z3jAy8-0RUMB?0l7YqM}UKThE*NqEY#?p4GGw+jq5#)jFlR1^+j6aN$q_
zpo-<|{51Zpi$70Wk7mmjg)11gpM1;POs`;%iY{k<79hpX+ZS@XF3YfuPx`!Twh{xU
zI(r(SXS|r7LawkvsYF2)X#6f+;inPnqT7~o<5wN0DD&9bKNqc`7>W%M1vS+rp8f^X
z7zb?V-NFV3d7#}a{>8KGA7;Vb>~tX7zJv7v4~saa?bHFbrKP(<1yF$35<9W1ZR}o=
z8Pxuvp2VwKxE@BDNjM_D0+)f)@>O+gd;)!<>)j+5ap|@gski!%G`HFgT8S@0>)h67
zoVL{FA25;6j}Ecn(@8u4%X8}_ClW!eZ9<)7qb@UXDK|v!kk7Kl&je#yCyDptCG47(
z3@M#|mffrX3ZN`Xk<cE+HtbX)$%A*eV_1cS;ZT_k&wSO|1La=mi$vKnC&@OKA8uTV
z)qWE_FXU(MIh-h%&oH74;#SeWm42=&8}7tk$FH`2pK-xZHV7M2)v&CEaULjxgW9WF
zLdowlj##=F>Y^D%8a!I!w{`~;UYxxko(b?138m*IWVa8W1Kerbo3?`>2gf@J1%eMA
zNaJqpq(X78fo$nHmA^USU%sx*M4}59UO>{I4ym<zmi$<gR<~J>cGIfbwZm=eq<T(R
zaj9vfUy3a74Vu{ADAo<RwTxyYKr|Cs@bf#o3b)^1`CQJw!u;fv*!n7+q7xoRW^iZi
z(bpRu7vui%>(yqx?MGwHuMAr5&ihL-*`_8Y2S6-!b|(c5UL2j~gMV;5=y2T0C*4Bf
zx>J}AEq7G&VV-~PXPw_C?x~AZ$_o<rq9QU*S679mWUDPqU>Ad&rIL!oHf2{gAD7~J
zuj7RSr$#Z@XVBOmd%n{GOq4pQzJjC>J4Y<)wnvr=e1rY8Wy>_PufW`(euf#pz*p$T
zf8i^4-J6h9q|hYt;5m+{%Ha}E1S~HT)6;{#(;C4s@;)8fnLll1U;N}G?dqt~ZETwX
zG$C^e(U%+s9LUZs3ese~Qec}DAQvs)P)IF%O||)4inv3t4w|9urlJ1*TesRPFR~Nz
zeXpm<>ilfpy2&5QYkAGXNcfVlS=Zj+ijSY)`V1G|e5QKvUrSN|ypt5_)b~=Nm(Vfs
zcl6pm{7|1?GeShY<M8X3AwXk$mn&ln^?Ja%wL#6vNw*Y_k`k_Ll+`-Y-c&l{v{do0
zW}(<TEOpz<s8lA?5b?Wm%&=pV@j1B>r&x+JcE@*Joz(t0to}g&hN)D`8fKeCmcHv?
zm(-N=dv3p_re(iQTtwBP#L4cHetpDf;?fc#C_++|OeQvt6K1dLy>+B69}vhKcq}pF
z71=pfhR_pim~_CfOq-V_uRmS>l;F>ql@cw5_tpTfAX=<o!(h}!H16ojkWni;O(50(
zzWZ@I>F~=Wxq#-v>por!T>l+Wc75kyTM!}BVapu-m?L3ej`1fp{Y|jX*Uo-@D{NwP
z$b5K>&`o3Lzb*=LCHbYt(8-rC-q%@JrU!^kvc(+ShYsH9j9lG}pFq^4Q57@f!zqz{
zwU4i?r&(>}6|?)xnIcb}x?>RZT{AvRr=)5GgkU!AAO=WYFhAF+Y4%&n6~7Xtw>M(y
z*6a^n@q?GXDSzvf^R&nJ5Su&~lX5QtK@yb9<p2*J{J`NOaa|IZ<GDZDp3Wvtlx6q*
zj9O_pI+tI283vXH8cmxZzFz<S!y8J4n(Foc;v_e6v(E%KO);@9lOH5@Ar<LR{Y!GK
zW}&~Ql^M91e`fWt;VINw8KCLhG&$*)<~z`_#nU|M@tjDpt$3=<><cozfyZt!<zGEQ
zE-Jfmdp*~xdh$3hB~+$B*TL&$gWlKc`SaQ9DlM?%&Gaeq)OCKDG*!=*r+I9)<X>`r
z>a6hn{Fg`3*VtR%4&c3u7pM4RDX0H^;(H6!??)Yf({PmKP*Dm`=9_?}QRIOh!0fj&
zy&nDoG`Zis2@M?SAZVT0DQ^*&(Zh6H>h!v0un69=&}mWimT$E2O-?lD%l3$9Km4Hr
z5fU<MHs&ZPhDmN|J3J>gG*!8Qi$hopA+X=O!OCp9f@w?tIBD9#Icn3_`%MJ;c^>3r
z^HU4u67I@MZp8Q+7AiN*p!BcbcFK<to#D_u7M3}Bvr?Vfo9M#6zd5wDqTTa)jURvM
z$WO{&k3%-g>#y~|mMaeNI#Z(BMCAha<@NG)#T}HU%QlLh-s9`2r%w}OVwNJvb&QIr
zlEvoIpi?tHa9ns?x%;M8(WF@NTYrL4O4E#Qa%c<1(w<!o%9fGytqjDaMn!MF^)3SH
zHSB)xKid!g7u}@$o-=gaB=yX2q4j=~WI2N;q_l$xjfY3}j4a#u^#xPvH(3%n$c4_9
z&q;C;CwS-0i+NKXc=uU&Da=jJ&|Pm{+yt1Dob%;nE(tz~tEtm=_d{9RcV7DW$;c3&
zV&P$Gach07wU10ikzJ0Jc|cju+k}?DUZclnTj(w_?o3;x(eE^O)_{feysh*`^hwp*
zTUWUpQZ&8-n}V(p!_@ir^0|HW3h%<DvGhRcN-i#^r*XqpZU$al$)nc1&8nhNgDZ{L
z`SbMd`k<9s@?y9%ogc^3uq{zc_u^&ecWFdEz8=*-Zw!0eE}4dl#c`J<Y|Vko->3g8
zMSnYCuRG0i;7h(m7N7W^*Rh8W%-c3LA%5|9*yzox*#&4;L6&uL8+Pi7?-o5%F)shy
zyR)*Z{!XzAd7i#s^Edr2LJkSiF_KxV<10BPA#m-tbGDzGowy=wFE6jbf=$v1;+o_7
z2&!mZV*GHLR!LI_eeuoo*(8lO!JK5z#ak49#S#?&zgYr+=bu8L!PE8;x)Wm`H0KqK
zs=qlq**8or5gB0Ev>CplzB@vgd(!OZ?sRvr{T`>vgEGUk;#u7q`f8^CPKi%&xd}XL
zq9$D|l4^AR_TA>)>Bmj{@mq638yAh|xs;ngo`AH)pT49IaeeuLo$U{x?Kd${nY<GD
z$X6|^t+HK^It+3f@(V09_etZu_f5mLvV2^!fqE0Pnb$oh>I(QygDY|1IxTvvBDtki
zDuw#86|?20(d;GEXQ*RsceUqbSg@`S_V%`ZYJMU+Y9fe8(sPxVM+|!E#vATn=3}U;
z{F{vpSOUHKMLdZuiWF?O5&mqj-Uv%PWi>j3V&f_uen)&0|BA-nQ$Z|g{T5>;rqu`C
z1I3e9TseLhqj=eZwAhtYCiB=XpV9yO&yiYk(h2qTMg+M@`hFr&HDXb69M25pqIi|o
z@r`W6R=XotbKG5{p%|-ZVZR}+ASY2>1~u7jQ(4QpR(Nv61H8!0m$O{fn)+nj$V-28
zBXn&=rt899$~h$p`#ZDQv8Aa1BA)5ntv^$L<B;XWn6Z=b`{lsStzUf2-}ZjJ`0PKB
zq*wcqD6NGtRqSrxA2EJHo2(zYdXT{CSRX_+1^-^x13-)at!etuXv*}atK`{i7%TT5
zng4qwG2=Cp@vNJ39BqFtFtsVA->^FzebG~2^GjSMCz*(h(j?j3A^!f=w0;-44hb00
z;~5Z^MKlZ}kexq1=Oodb3^Lt5N`KB4Zq*wZ)gKw9LgnCNolc|eb>@-f!ai%~rS}S{
z-sd~eV#KUZOo+fZ&fOlpXVFMC0#4aTifcMS!OkTP<0dClG#qQ#%g~b&(%wq0=ChY$
za1&E|!s3eHz#DX3M5y}n%rtUAV($sMg)j-0XMz7b&s2>wE6~wPRf4Cnjs=qb*Dt<F
z-hOVKP%|Z)?*|^kH8$GBxw#f@8I*nL;LJrBmofCDE?~1nQ;_24z!dc$x{7&fvnwa`
zgREZvb<AGKsC;fnxy^Oa+s*dO<;k!=Az!%2`1N!3qP0~C-y^Q>To+doO~+?G8CX7+
zYs_b3>JT)&LH2#Lt6Vj^)vT8$j6=O*i4N|EKpI2)8%D<#%Ggo{j`<m?*A6=D`+{p8
ziN0qugh7@;G!wCvbXw-pD>T}4S4TZ8QaP7au?NBr75QJDH+&x!y=^q)Hj`qmSQlHJ
z<a`1ls9Id+c_Xtd!Z*Dvsi5j;ToQY~AH=H?b@KY(F}w%D+T?p52Q(&2yRk43b?oPd
z$tBbpE15YTC`;Yvps&MrtuYPJ3clVM`rKNEDSD~xpLHq8h??Ima7Xw5>3bH{Uy%sH
zmMbIHH@$bNFZ(3Hl%9GyUv7|{t+ke6SQ-{>7fIte(Rw}6SETMEs9}cIIf&Jn7=cZY
z(WK~)v`itUxiDu3nl57hIB$jm&TXdfkyt~wnhhPj-m8Q7;LB4r*MuJovwv?Z#!=Av
z+g_f?6K=!(L>1R7Tvq!K9TTj(?QZQ&SC>JtX(U%Hfi)HP@DbugFe;i2r=smFjj8M0
zXVI~Ro(;JbDb%(Z02r;tN{3GZ0C@?e)E!V0!<zleE0p?{Y0^CX)=b71si;Tu+}aT9
z_v*}UW)$k>`B;8!%Yv>_Yll!F<U8k*Yg`DcEEs@#@Yr7jl|?wd)aLR??mKc(1$8If
zKP=4uav|({qs7*x>weBH43Io6*CLYruj{T0U|XQursdl9PeoQk<nWfQzKL)WR?Mfb
z-Ovl0eEzyBzf>5qr{$j<452;WOR<gU4v{v-_)q{y*=Z|}Enb)5TuK&{uErqayY+I9
zjFD=QHgQFW<!Bq68#gdvjQOAA5SXury(sHs_Dgd|%B;9P$s!NE!akdw{(b~)P9}f%
z?IT8n*cX7R{Wj0UObp3*vTXo#k%LeiX*TG`YEChPy7J6;CG(a|S2Lzob9q?x>x$sM
zV^?KnzVedRlmm}5wsK!wC@&i`!WcABnC@}jp<O^n&@hyd{ejl@8s~gG6`Enb;GVf_
zjNbzbV*P~3rztXS`^L~3*r&Zq2!&$yB5S_?YX9+W&jH$Pmiy+DDW8LA^TFV~u92uJ
z(NoEjxwsfktfbo-$ZfEFHP}&Sykz3FN>!pq1ZmWZ|2|WmY80D#uGrZjp7FMlsgpiQ
zqe1atMPgb1;mElCzXmh)?3ti>Kn-78+B46J`wtq~S#%eiBzY~ZJNPt|-_S3!8~8>t
zTSTRlzSK|zOB(Oa8b0G2Sp0}b=e`YVh)~ShV1mt55S@r71|CmZQZ%Nu%iRBgA7qic
zwB*Zmxa*CHb)*o)1UNVB*Rer64VfJ8?qK^>;o`25?Io2B_HSLk{rbuWl7~qFLceq;
zdOMPZZlIOBE2r^H%+Y92zSA!E^v`sn34^Pn(OQw=FxY!u0PkK{PgdN#%Uoy7IwG|}
zjRCFs?0>(M5k6$P)1$j9U6+%2G@E0ew{|7nSefwFTNhM{oDfvuVhaOhtZVBp$+FhO
z@`>mu4$;a01qNU1+;hToQ~O6=+$2SxsVZwd(#|&hd~ErZ#(QmjYrjeO8=7Xeq_JET
zO$OvTG)0@8ImUc%u3Zz~PQ{Pz3&9J7#ZgRMeRNI|skeEVrlw+K!<rzi7#k2K4MGUj
z^Mlg$?#!b2GRNbkZQoen1!R}8>)rj8JYF$4VOVw2fRk%wiM>keuEb>-A&e25YR4WG
z9y0qm*9K;~3ihUVm8W#BN=6ZuhVU!@gqkPN+Gw9Tw;=DiPtCk@AK*?s1dsn4=1?!f
z-m&@IhFXf<&%HAat(@Ku#T1W}6|z1|1L_yAv>R%tm!~A@Jx<tVwwMDXk!Rfp*6g!s
zQK%8~Suijx$G@s_cB-_WI%>Xk`wkX^*B+NC7+aUs70nq)o!}cx|JY@Qe|Bi0CIxK!
zQmbR34YII{jXH#8@jZdr2=HU5i611M!Ne42QEh6^AsRMp^DOXY#IfQUg@f6gP6}M;
zghcP1^vr|hrF?Sy%WYddMB&VR%f71yz|VvPAZOb#_H1MBbmem^q%V!ZxnSz&6?f$2
zH(^)WDD`>=Z>WxsD8$1T+IP(v>hdRBx%XFfqeu}1FD<hC`C|Qp)@fOJnA&uZ3mwD8
z_!<5gz7z?o(^a7VBFR3^&Lls7Xz|SX;Yn=CN_Ii;tq8x&>6tfvKUZpIznB06o1NLu
zbTZ)R7?sf!I$zm7b*SAvdh2t@`e#VZUf;%&)J;pmdD~@wFL1p#xjV?hjbk^rw7fGb
zTIJ;%N0;tBV$qo3{bmgGVQy&4FNdoN&f(OH)=x53^4i#23OU52->Tn~HHx?<85P;*
zVW(<6+$(F1+G}4?e9EnQicfab=BFw;zPoWeaa)@lRW>RD5Hl9$X8M4$QcES7M_9qa
zg$^o@g#I~M_}Si4v3452V9F*l)0WBi!$O$f(Lc5^U^`0W%^QlK<=C+oOD<kAUU3t9
z^|$6D>&`D|%)X$2<c%MssdrYl6i&gM$7L2b{vw&Va!)HVs7X=X>)9dpM&;-UxY?}f
zJ6BpmBoonpvQ(`2*BR<(lbqf_;0WA{dUo#fEnKy4_-JX8n<EA@lj=TS>UM&R;^V7g
zOU;v4`B_Br0oLE$-QVqQWY`<{<mFA-Lb86#4EH1@G32H2R)Q#ae5R((Ce4&%V%87T
z|7IBNSu~q&3PvDIOrN_KiA8prBzz-d{s4r2=(LdvsABUrWt7kqz0u={zeGCoS~Q(w
z-I@ENq^dvp9*sow$17bkkepPh*e>w)K}`iuiu)vtO%FFNL|VcZyb-a_WT;1rXCznQ
z+;||Lo^Ms>w~Eu}JC)hRM#YoJtq?zEm}sx5XEJPE*P?qLdaZWBqHwgQOui-+sI`F2
z-0|IWrc1@^T{Y*9Bj%&Q+07BUUX)9de6cfCJ=Msd>+rfS;#{T;zCUOR9WAV@wmrAw
zwjWI0R%*UV`h9bD@sI0eaeb?E`H3<1TxRz%G(5!`TlMEc?0S+_gY%5fbws7PW8pfg
zuG)XFt;yw_SFM01yJ5#fZT{iA{Yt~oOp56vtkR2LH8JVs(Lx6y+i(J;saCNhgufc|
zA@3Ziv#Ho2=*rwcCjUf3EA3#8m4%hhj9)|cG<ir+QdfDZTNloYX}&I+r^QFjVOsQ1
zG};dyNI{h3q;<Z(&|XW|-wZyP`vs#3&3waB5y-<j=<cJWZ9&G3oTe7JMkf^M&#4`d
z9jAkAxX;~#et%#SW@XyRo2zv`pu4PiF@N$$_5k*PhmjD23@E9rEJyF**FFt1D&Q(l
zby8Ccyfp|ZUQ#+vs=c+1-?AOZX;>5}If=o$Z{<Fz-(u6Y?-2FgaZ{_CTd`W(V{d`y
zd~kNw=ked9vQMq`LF^haO^Y`?ecTYs1_=kmfg<&GFGRsR_6iK^<2$?$(->m!dRk5r
z4#Z9dmJkVh5?3qx$>oHF&YUiV0UzI7`b82X##?{=d3a1^>Adi#%z1mJQG}P*emHt2
zzJ}fVwN2L~YbXbP$F+_II0yNDtzqRV3Pxsd9!<-Y7K#3*SD-IS1t*~KJseabc+E{>
z+pn?U0Zm^osp$aYRw<&`&WTr3A+ESOO!DDnd8)39NFA+lZ~|d$|NTIt>0QRho6S#a
zAz1I_Tiew)=tL@S%;*I-^+9(PK+Z_bc?FSgD?0eSooBG0@`Hnst}_Gn!2I|~oaKnc
z`}GH>-ibKk9So{++}K64?lyg1v5i^nrRn_XB)?p__AGj}3`O32yX;%#^V1LfD4#CC
z@>smU5@0TM)8CJKR_|3uivPV4hgXK=$NhrP5l6Zst)L(2ZFQ$O1#r+)ntj|h^|N@r
zdEj+q2S&nQfa1fZ_r0X1qTQ*TByxG;zA&%VFcCC%YRslKT#T2Q{j*#ONHtJ=Vpq91
zn%~;G!98W27v%D<@z&Fp{IjIR9>ocSR^7IYOcuZB%>bnCXp(gr6r~=Mfh#>)ajXwk
zZdeQtgi|(cu}#yf_<hQ1yZd`iUBK26#O+fB^$IH1zB*`u<FmSPG(Um7Mf6>uEytAV
zJ3dfr@ONy1R3@2S6WIr@E!{#D9eR!N4f2|fZ0mYA)rI06{Brf7>l|frT3_GaXZ%Xb
zwHIznzu2A_Z}@9#8PrOpmyL8;D78>E>1yB|zCN@|8;Nn4>N(Anemgrv;H~aiOX?Lq
zX0MvNbjmrtiPAUG<YSh97GT_MbX0g*6h(0F5T|tKY0Rjr;!VPot&yIFo)OYW$R?9n
z#4yO~ml-SVZsyHlI01crJev`rcju=^3ybTr#eKh$Ys$~$=4Rp17EAa-{W`HTWOoNb
zH?Bq2QXM`kyVH^+tehu>^gb>Ak6-?58(i-?@_{3+BVcDWAzbX2uv0MTS;KpW>|+(U
zm7b~QwqfVmEP|dM{5}7&u_x2sHo-S1aNY*KGJYm1<0$rnzlncb*vy2oD)?XJx=`Xy
zU3g;Ox{(jRuI5q8M|_s|;!0L=nvd7$G6K$5D0WY2S9kT~E@$I?)MopSxpoXB$McfA
zBewcB+M53~FK1VPJE6Jed`z8<qHGBV!;(6cjFw%l=5M8zOvGE=wh)LdXwdb$HM9S~
zYK922bmWI&su`U~^i~?$>@?2lgT2d61^1>N>Z>zoy;Qu2()e(XyKvNCk|o+G2N<ZC
z@PX23`1Mh|g|wxm`N<@&s&VYU8x(Z(f7I<Zi)J;f>*s~-syL$yM@=_K{0VO}Y@rxk
z<#Ui!z|F$5d=5?X$&H#{EHLoTo1`Ilcbw?>;_>1tAuM4+f7r*C<OMeP0KXV(<B;d~
zK>+*NVV>je<h095`=!7C(h!&=Sol3+=LJB)YoZp}4)GgklftmxdYtXONJH^2YWMMm
z!afGCa_8D{nres4tzoui2i4P*)R%c~Qh#j)j<t>?qn1i(Cmf}4h;`ucV`FWktLXW7
zxk|@I%@qSXn&!`+tv#s=>=i73SBzR1oj`rlPpIz%hc&mzUsf+pwq9qY<vKs(lrncR
zl<Zz;{!S&NZFjrKiw4jTJ7Z_OKJHeqz5=`IlswlDIbHxQ+tAgi1#E)oJb+lIaf@N&
z7c!`TLR!Z?4uZV#Uu7%V=5!K>M#u?vA=rwuZ<PJS0cAHko;4cOy1qHQ<wMlZQe}8l
zxv6QtCM2?OFhb%PcU6T8el7wJ-Rw=RPw%Ruenl-az?YiEze#6h?Ut@BdM=8iLCj)j
zy`k2*0->W-VvRyKciz6>j~7Zi#83|O<o6)z2TxyT2|fP%TU_%mPsOt`ZpW@Hq0#d7
zfAYSsHeDJ}Rio>@k&X2Q&KH!A!X&DGuVg28qtu+Jk}&jUGDbps98SRL&=Kmyv&o`{
z<3YNUA!?ROkDDXy$&+$bU|=8bP0rBy$FI;e3qbd2+*Fc=Ryj3ajiaq9=&aD%zW>OX
zWvxMexTD?gvurTCabR_i&-Zsd-QtqW|Nd*LKNsBG6~YA}Kp~A!VxF&1L@iOe+PW-Y
z5<`w);0Zd7px6R)c+WB)YmIE%HOg~&$)4`$&3Sawf!Y1gM8b0GZoh!$Yl-2OD|p+Z
zUE8I}niHD!-XbN*G#VYG+t%;R0G+cYV>0p6MF7DO$PlI&<p2gJ&E9fV*-kpmImG+T
z4}C^x&1^=Dl>;p2<EIZ%F~;E6RXd_xC{Voi{pAybN2Xr~w0}*f5w}K4IlZn@vV>jh
zcLZ*nNSQ5vc7PVIT<q~{3|E#`US&r6fp_iJg?6Te{1Ei0xTF5!Sln*pL1Ycuh;>Q;
zoFIHXpPghzN!Q3R;wxQ!U*?|ZI+A%j_YW>8yd-un$w6nfgdk9C{roh!$th#?EAX(@
zjgF0NR09TcPpS*RiWuAKe0@(FtS(<-hP2C1-;aND;gT01@Q}n*#r#IFZh==<+sr?)
z!7yA5=uDvjB~e18Om7I2fu+Ujr@e(77kYa)?@~~+s_FrQgNUnV)uBFNO>3}clIlmq
zjdTZ+R|*|InzX}QBc8io&x_J8R7%YcepB7<wI9di$U;T9r7n=U8+xGXGfwb+P;I@G
zxkxb@h~62eKI;X#gqEA?Yn4reSx-ZwhwH?p=r%T0)P4hwpml985o}J@7wgZA*-gE-
z!;@L2PWaA0gDd=0xt&JCUcj<(Qo`J%?!)55TZ4`JPr_#}$(la6e@Ur(QLgMMS@yZ0
zOW}k-&s@u+R^72(W9W<LkKl+^$;M~beWpSuh+?dUS>wXxFT8wO16|;7s+eh7PM5=#
zxcNyRYJZfzPqKK36`V>uOm6mhWV+!!8GI8gr5zwW?u>u&=UD)V*lsday{}2o?o)HN
z<*3fTqo+A&dG*OzwSCM!??Q-%guf(vvD6hN(Yqt(ihf9{6R}7JF=7E>V(#4gyC3SX
zQ^>3FA>h|MY_50taYJC05YP`1!St+ZO~YkG*LDc>^G)apEd5vWdNJN$)$}Xd{?8q2
zn~ENl!mzP82?dT<7aNBZ;dWDuBA;f0BkcGZNlbaq8zgV|`?msE%yk9qm3Uz9zy3~W
zIy3eUarSPCpk^V+a|qR%z>6hao5@a1Yi2odZw&kIGkf;q^}3EvfVbu0<C^cCE~A9}
z(WE=yha)-!-mTGK*gQP%{Kso{1&X%|Zi>~K8sDh4zn9!Op2geB&~?SI8CHEj?-kJ*
zg+&wL455RUwG&(OebsL|B;s1B)&<}DccBJ}(d^10bG}#oqQRlq%Go|dK$fn$Zb9^A
z*4mqEOw78EYjchz=SJ(iMMYBszI8+Qmira!Yos}cNT!*G$ND(oG=~XlZS9Awp<>g*
zC61%yEP6MWhIAmuo#lPjgJI!S@Nme)U7JT$DZ>+~`<m$(iL&EJbS~M}Czt6UAfw6#
zY11}x9ipmVn8vJ$?^~{6VwX<E;C~ex#T*s+g8Y=L@`<;eNZ<ZN909@vMs2?+fwX|O
z7ET$`d5ooP)}Q{_-!mG!4qxP_V-fpr%yuIuKZ-6NG6S=DTDHmuxZ$dKJUc*|!3bQe
zH*T~rqXkyPk6SZcaG|pu%M2HO3Z=yP^aHO$U=<&VnV#mpXC#|@Ql2&4HR_?yuX5Hi
z{FZ9_G_qXD`*3@(ueezV6@OG4Xn|z^aJqAkXX>>-5-pn)a-X=nlVGm9^wn?#R(w~s
zEomu;bI#!{FV1yi?zx<gW@Eqeopqao?Pp$%kSM83t8Lw>7`sr1vJlu{T9&2yfes+^
z_K@Ln0q4${QT@h%_KrPC$8O6G>*0IF#-*C3#KO-)HfK}%Mgf{hS#4O|x}=fs^%~nI
z9qVF_&vCOzgL8XM%W%lJ_>bv!_Y<{+^0K(4#ci4t5mEI&$g=C3NjZ@In%6Vj*6Bek
z5jbmFC23lv<Y!_4{gPDgjsfEiT`yR3d`kgZQSz_QM4{K^QXH^{XG=8rF{l2gvm^!E
zR#B^^8y&KbAIWU2MycrPJd|gI6EW`x4>wB9rw&5-WoM;8lTh^!<lS{titFm1$k&Ap
zhtsq9vLmN_2Ca|uVFBle;Huk@)f-`Vk3Sz?-IPg4)qt&?oTmxNO1LS7Bs%~XRrBdD
zE7GmgIOjQpljR#yKkER|*4&*p8oJ^Q90T<cH3iyj1Mo8|Jw_546deb+R&Ch$sy6tx
z6b$DY{9s@5yHtQuLMivojRVrbZgGhE^!)1~^m*xy+NZ$PZ7Bl5WOegGaCxhV`A;w<
z?jTlDs2?a?(YiXaeiHa-HC~eFa`#9E4Jh)o{LBIsZD@y=?$L}4NMt0kpA(uu>?~>p
z(lUO(p-Og_^!*#!l(W(1)VcAYjI+Mo6k;c9Du2`wIw)R36Ot{HBP_z?bA%(zP{mgS
z?o4ZJ^FNKZew&)#Ny=-VMg6hr{&W_TmEdCjH7*dlrBT%SX9A$IJUyR8J*sL>%8xir
z1n{3n!uqd>!P(G?({I?_)3yFZxy6TRsV*V@_W@DNXL-Iok$8bx?<G2>#DqQhLx=x7
zV#@LrPOut+u;*jaWz9@{I5Z-FzW?0{O=+Bo&L6!><~w@Go&>_^#pjjky5<&FSEhz%
zxi(FJ;>i-jCgNX^sgY{0I$9e_@SgMPDsJkgYE)JQ-Xvyk`L`OWyTt)=E(s3fmL9Zk
z-h!_eR_6(Z-aM9Dr5B7LGK`@wFw(1}=*vbW4H8=G7wam=go@o`YFQJ$m={zVHgf&m
zR+YXP*`1bx$e(o5=!n_k$)ClmLH*+(hK^U2L70&b2!Usbes}c-kpel>MYc{2zy<@k
zvuwS)?)EdVL?yl@A=}Un|6|d9AL36N9=pkRZI?d>#U|iBfNc+d)|&8zuAYobiJeU0
zl~*)=)~|8CiYBr+G?k=3u7mxjg9-<{R`O8@4<xeHmtI}fV7dnLPXE^65eXGPUF+l=
zN+(>2Sd<ma3|X4HF=D~0*DTIHHnDX1YL+&1ln@ij(E~0~ZawpuN=PsFNhLV=*#w^6
zsd%Ps=H^{Q(|T(hB0w8kZLc+Np$~LV7}>vkliwEycb=VfUq9a-UKpk2tZDqYq8{Ma
z=QnIS)%(~x{w6FJjV>+93udk>e)9YVO8e9$p62oYY3r-QqWYqC2NeWC5dmos6_D;6
zN>Uo68$>DTW+<hVQeZ%7q+7Zf8tEPyq`PB)f#DvI-~FENK6n2)u+Khc@3q%n>s{|&
zYaUyYC}+$$NU4gorsZC$3B17ru|kO2S@Fq>?;fdGrwVD;2)RBFK!sRoIc7s2!F&Y4
zeOw%RZF-sRZ*cY?XK-y@q97H8!6h0=T`qW;x?~Q=CK1>Cp;AxP)D*N!gPgqPxcKD;
z#R~;@bP@COgs$yjx~IDy>nlZXTiIP`h%o$%f-9uRAb!m!(CwL*PRZX^@5rBVPOpUF
z1t+pNo0ZR}*9a5}FVlQE9+~Xj{0!yY@mQ}IiM5gQnvaQMWfD9ei^mn>_Bs68nw91*
zZmdt{afCtFA=(Uv%(p_N(u{$lLWle7e0}Bd+<;CA%^dvw%x1%Kd<=Tq6m1GpNzQH2
zk0xX)s1C*Jb=sJzdwJmjt{wiF1wJV-^!VW8jHEB{V0dwY$~ny0w<>}biR?)cOc@Vv
zZ;kf0Q`XH+@DT$Q@pf`6O;w0`h>LynQQl9c1lT<ezLN361Q4`Tg~_-b9Rim@9i&c=
zh7Tuun?>j$y>q+MsGTsSlA>kZ0qGZ9wiusMUE@Bj^n&g}D21a#<2hHi68d29vHC;R
z;+d-rpB=^odo%y#qqbh;dXSB~#H`-vJ3AtW`k5_Cv{AXatMaXDyw*e;Ka<9bmpQdY
zM`tjf(9KQiY<#DP@Wa!X2?o`l1{q0g?Vfkx+u_Kky|Fp&cX$N`QKxN%5o8{9Peb)1
zN?IexDw1na6Pv##XDl<ThrTc-N?}H|E!ww;``H$-=v`Wy_OVUBv89Avyw$b@;OG7<
zVV`HqXT}~(n{yW|)`HOe{t0+|A<n)pdNR{MSH=T=<GoWogRIpfoc)5zi!$XoC(k+j
z_^v#NKxGX@b#B`tg%0x{C9^HtO665Y)*_srmMHRsy#u1AZWj0(!MYA9cL;y=EW8b<
zpQu+z*2=F#RJs>Ca=2)wR3f2pA(N*`p~ARv%}<`JiF7oXfFCPKBl>hoen=|74uXGG
z*OjQ`rd4B|2@89se&DW-IOdLgLLPyE2jCt+5;^72I`xNyKFGCJw2tp#0Y8-}vIi!@
zj1N!eQDQ;k>M!&W>w|YSp|Y%{<@lzjNF=TF{R!#1n#qd}-0m&xleXjlml=4~1N0S5
zwN0+R1uaFo(FLcvsM!Q!d1QM5LzqSu@0vgv{_2^uXC0T5R#H;U5T(3!cY3KRHf{zX
zRmUR}jYMRR8A-ShJEz<8)Z{<x3+h}6sIkCkZY1jWoaj22@w2%+UPng3Xw4nAU;NgF
zL9U}bF&BrT1=cNQSZSH9J?RTdiTbvd`F*Rc8?P<2?utkU9t!x_m=uZ&{jM@N!DAgH
zqGB%D465E~;eA9=3X|zu4oMSh@4%2mSH`oQQq*N_!3ewA-54By>)OQ?H6-25Y`=|?
zsTCAR(qzn`9E{+aX1L}FFJX2&TfRm8$$8H3J@vUbXiUSF-wRzr&7ZQD7wu^tg>Td$
zbohpX9xhG1v1C((`7FDGbJ+2d3Ps;FUoqonJ7ZXqKwYrFWJ=r`YC4Bp&WzSC{m)Qd
zQEf3(b(4Nms2|5oee1iIJeT#>QCHR5ff2Z#@wBPU9>!XtLyxr<y~gFe&_4t040+76
z{Ux9Gvh*jqrH5ve?(e3uNXm}|=;DiWE6y)s=BOYJn{vS2QLTd=Llm<4i>0pFvOr>q
z=3!UQX&QWIZc9!#@tAb<V)z7HxwPVHj?LT;yI3EN7cS4|udkT7S0U{P_8gtzy35_6
zVJSV#CrILRv{t`M$*-2Tp}ahqdC@&rf6;q4i+nSqW^c5^`+QH#IjnC&4<HL~O&iW#
zo{dUb_4;+A_hJHR*QybfFS73ubIr}gTvKolL@L7)0*rqbVn6n;ir*a|mZd4dckV(d
zN|KR13;oM^<}{Yn>l;Jx3}=xOYp|Pi={`B#QwMF7zZ&Dn_H4@3gd0P@5lhHp=bz0^
z8kCP+H&Cw*nbUp^OUF{#to``A??EK;jEB`|vod-2azO`HjM(ayJCE5D@wpJnTezr7
zkx`mCg8!Zpq&l@c5jADSJWl%CPYC2L4J*_xe8rH@Xg3Khjz>moDH&!eIut)IFGo1;
z8^sH}9aHDts9}tl8+lAQ=aiGR^)Bp#fz1&pmCV92XM^t}ugj-z-)Pm<i83}E;=Gg7
z8vVA#-&m)0o#|`gBIwy0n&ic}YD{zTUlkRnNN>LRYq9}WVvd!|-B8VQzFRx$g<Ml1
zhbK030!9OfP0(VLPWGb`junHKqPn=EdcWhCVYX$}K<AM{qJOzUVqk5IU^hX<Uw5mN
zM#u@g=1M}Gjy@arpgGgHE>N>q=WwN>qlBT4s&n?|lUg@7Lhnzzj6jafh^Gi6fe(%w
zXDbc;l9LRZ7VfzSXQqg^x*1zb(7`Tvu1%4OGrt0?zc@Ys%9E93sknVCklV(3{OwvP
z1DQ#`;I*((m8nT=x@`LX4aWx(T~s|EIrsb}rO(c!%d3mY_p8=SP!e;(<plObo5sZu
zv_>|4oypGaZFmO@2$ay>8T*Im-VgB5J{9k})q8o3*9sNsl7)`P$4}b*kpbA&Ge-$v
zK_(B|idbSEN1c%iz@z3NcemYMhyWND<3NUCz;wj9`az@)`Xq!SkoM|aka;UWB6Zo2
z)^@ro*nTlVUdx-_*}2By9@D=5FlB#U*Q<i=xkS|PLb^vWP&^JGVt_!Rf54H}6CT-B
z&ri-aUo^Wb8kUMJJ>&*I%Dk#o=I4ls?fpCbFRE3=ySNIN%s-TR_QuBsEU~VK7bP-#
zkbM{Qi6x=Gr-6}k%5O7?>1pr`$&M2(!0OC}cS%WVScq4+q%+$_`!rviW*}Lu&ra5m
zKCnwi+Qa~Bpd>f|INOE0tDu9%n^NwRE;Az?b)-6*XCuoP@)kTZ%G|nt9=jSM2;jn?
zBXWo>@YTfDDy$edE$K*8uoLNGQ#6rzVd!FYEo*FYOA^ivk8gE}m<?xgf)9ifD5^4U
z-HE^#<Do#lNwsM&#<`0@J_vz~;;4f_;s^3TR`Ha2cY8S&XcB#P=f(VKM?{~;2Xm@U
zJEptD4A*y10^j`!LA%Qmg8P_g%Y+!i)oy+6;rRcCIRZ}_R($UY7)6N#@$*mFM0bD9
zHm$U$n-rK3_cWDt9@8l$^&hO~VNnAf{;$^t;P(jbr!5aV92;qQB~siuK{&nHK*dCf
zZ{OnSEX0Hwjcgn)lU7H9t4V0GhU64&1P@Sg0h@YyfeQ~=sJuljyr_9V#61d+paUd;
zJYDCq&Gbk9dSb+;K*9Meep-k#dA|0`14|Cv_cj!9*pzgN?OxtEe*CQ=bc^?|-7XwD
zJ{M!r;T1FvSF6Vwf`Lusg%Mb!!OTCq7mDJJfpoeZXA01v1?s9Inuwo%Y5X{uGA3`8
zu>gP8XxwPnf={89jZb{4ou?}ln?QkVF~m8-*e=k*#IpnaD~7pAI6hrgD+`>g3~)Jy
z*UQDN4apY4JO+a3x@dt-Bnxe-?Wf^w{4cLbqdcBfHhiAO#0gDh#bE;dvuvMxQ|7ZU
zFfJa@`kib&#K&-YDn~1_^a3+K`stO6_u}LH&(JZUP9Tm((<o{(q|I35@#J^_eR8u1
z7On&OL@oJCd1bU)8J7*yGytO&2??*rIQd)RD>MW60!DW~eO?s%qbrw58*q-Z^(CKC
zsE2Mvt?Bcx$%I!|Zc2#M<39p(y)8<_o)vp+HD>;3Gh>vn`Q!1%FWC<O=Y=f+NE|N`
zVkhv(Z+;eQ@Hy|vE4@8C+t*dRR}!zw`#zQ0ulZPEY!}d|uC{%>Z)_k|NJRazh~Y(%
zfQi&mm=meV8+FNVub7kxAjM13vnefsXqVx@oi6*9kE+iLrPZ&Q@5WL~?JVZsWR_&r
z&~dW&y$vec{lg~Pa1l?_TkmObbup$pSm6vrWJ;bE)S8@&&MlS`&4YIw+RoxKOQb$r
zwhoYp7l`>v7|YC;T7G~=X3APo4n#Z=%M|oFoB}u@*<Wp{iNmT|u|eXB0rIGWCbcld
zh}=(T&FRA^ep}gZpMrA#<VI8WuqZL-(#Lk`mJ8I;kwJU}WccKzzTJ0~t(vRT+HXA8
z8%6(;e*lz@;1(Et;dljby&TQq!-O&?6U{BL-Tgp8<!omL-zdAa<A#-H)2g!vcNSYa
zqh73p?uSW2ZwUU|^3_XoSsZ)V=yx5zDxx^NF4Wj(wXoS7$h#*lhdZzACN4B%J?Z}j
zD^r6+Umu!+h=2Lst9ypVTc4|F%ay{DLRD;DmX+D2M2Jjf07*qY-pyW^oR+n)cQ&D?
zw7&=Pr~;C<i|bU&qb}R(LN_34=tm*@C<BCI@Y6JP&c-({Rb}U<PkxAv)_I<9M31*+
zKL%aXL0Y7Y{kG)Ydue3+Bns8mkK9%sLmX3HSPAsTcF61<Cor6Z>=C5Mf~0<zEi3|5
zo8I)&D;K0$gRiI><~4&G(50X7J^-FYKLI)Bez;EU=?hIwQigjVJwUesjqM#%GO8+g
zRaGYz2EL>l0rDN+b_%#rvF=1fWA!8=u_Mc={|J9TpU_t-sQtsmSW=wxhthBhCre<c
z8?6ltwH%YK{`uE$9qncK$Z__=n@a*fmlqFiqE6^>j6liEo$lpTU>2{TD4<D2|8H$5
z6>$3sAYzr3Hs&u!A+|P9l-&Y8@NWN&()K@)?*5ly^!qCR^E870QERBW*9IXE(5|Nu
zXl%da!+wC-odRXkrrnmj(guO+nE&@r)0~GS4e{A2w1=WA8Y38u=8L}F`Zv86^jSm@
z>kXQKI*+wQzEmdQIYrMlf7PI8-;hEp$55pHQC;iB0CJ!=K+^1soA7t<#RXqlboani
z*=rsG(8jF*!sOsqA}}klHkba2CODIxG$ma^L_VEG4Wa6w%U4|@HckGup6H)%_Vjfh
zYc7Ff-KpeksCr+BVU1N+iy!B3&w-<EpJ&?WN9C)5U1F{kfd2QDh%SwxM#|&zcueKr
zHy$W?>aQCMvu8bcL@>G{aYpA>gM1{`iX86je<(%4K!xGkl-<uHvdf6sdO2H|`hxL#
zG-774`ZOYWEekGEfSW3sq5f#C?qx|AvsnYrleF)TafmUj*lhR@%Oy0v)RO_K8fc>N
z==d=~ZqhQ#k-b^MWCJAIigWpcc|ZmN;<eXln!Ac~q58M~`f2@W40;0GDxghF)UkV2
zaM}v$&8VA~^>(Ix&Vl2{xuj_p!%JYiyMvQ-eQtn04F+_qIBL~(FPk~$u_}iB=r*l7
zMlva`?6T&timJDy%(GR0_qNZzpLKl@0!=D$!h|RcJ1$+B8;Ecfef|SW!C-ym`x<H(
z^+on}g^fxr;iN3_$?Enf^OHjF{<?;Y_K^TW3>-hMk;*r}Re`Hc*N^&*XD8v>@Amca
z!RCbRdHtboxncmF0y~bStA=$v`M;e3`W*#m7t1BjFyK<kcV`SgO2RuWvBk_s=&Nhl
zH!DtD9;4(V2s?i(qd%Jz5>G@kJ8X*=jE9HWsQejxZ9Ib5AjKpJNHY-t7Oq)cSc$Bu
zLv<5Bv|-0V{kJOwf)65+sbC{U1Ganz4N_Uat_Wo_pS`Ym14^4=mdOKbM@`v~%FIGT
zi;bBLhIur(PpNlZZ?=?CS=ul$Zf2dR;|D-_c~xJK1N;5ox-|wrdG(-{`Ak^sPo@bK
z{YbB7B2^M=n-Q577q4*l!~P=!#@j&#Obw=o<T7&5xqku-)1|>`A6VBnqDHHOvi_Y1
zhgQgRLBx(N6+BhyR?PMASy)&QS%Oy)))5&*$$a{Db~;OS)RZR>_a4PZKcAgEE+)pQ
zuTU@i7Ojz_R3!hgy{1lpleS2gnDPys#EiokQe?IhR-?(sNeurHGT+{K{P-H31KQX(
zx_+~>GP-NtbnI}lG$7Omo`L2)-HBJNbsimiTHz*(?b1E<0og@Mmj`o)wS&0MhLWM!
zSjImo?=E9L_tSo9Xs<poVT(=LT6e2-3JY?+A9BblWyK~;yB+y)W6wI;+pnu8a6W4a
z%XsBmiz+s!&l=9rGWoR|r}}#nmt==QAWx`SMo~+svpeyzzC9{Xgn?laJ)0#RN}P=+
zz|4N~k(;#mEe4f9$&C!y+4c9Bo!b2J-H^M_;Y0Bjo%2|Oplkzz4A#_FvETXAR!{d}
z_~X(@ufE1_rN_REX64eM5?u7O{LLb?%_2MJKS(G%d^sZFQqaD*H-#gqD}h2cW))Db
zxbgqO#anz=GgT?+qAFU+&-2lm9_a8SJ#a725ZQIx$=3&)HXwcxs_uM!x^|e`RZIy^
z<58yK>VvCYUbz&U`jM?Gs*du;<{|m_+RKAKg8qpPXfgz!uAj8_jqQL9Ux{MY*rP<>
z>e~IB{($u>;Z5Qw?y>E2%dvXa%Y#-tbhxpZ4%R3;)L?i*qW7zB7b@IsUZ*MX+I|Ot
z=wcMy`6?Oe-|EyYHoMBviHzix-E|Z1he!P`Y>?b=l48)$qPG1px6t}mi#5qZmpH|?
zAO{w$abQ;btNZa!LJ_*>G}GO&3NHl0%IX-7l?P?GshLor>uB!;24kRC-_3eI_a}dU
zPdYJe$5Y<dIkO@PYbzFG8DxCW?ea9QDA!*NEjJHUH9gJqTI=snXVbpHHL*+~teQmA
zC9Xp^bG1F<@L(w0s|~KZ%%!*{0bBQP=Tf)w(oo2p<XBE$`h>|FS5dwG)>7eh5Wi!~
zouUW?*70A+!4IcjW1WP{nhbqO*_~URnV)s*Z)Ga*9&O^WQ5W{6OCkS{YGwmLnl6zB
zUX@+Z03MgKp^ke>w4**JY^j<9;t@Sn6Y=wG9M8=cobh+^@=h;@U(E4)4GxCVQ?Z2L
zlvCm3b9kZBp<tAo!Oy)v%+z{f!IQ~#r>h!=m0x5c=AJa#Am9Jc+8>@{m;&y^SoeA1
zWwvK-Jfnr?+R`MLCdN)`9xjmda$-|rE{{v{qU@vI#i|o|R<azsb{|EJ&04aM5n_ax
z&VgNdoC}QlPik>r^{AqjHL~PhN7Q|oG#M5b)839{v=KY%l^dneMO+=H=ZC0Eg$BD9
zp#u=$_43-WjDR6c{<pBDn$CwJv0*e)u2wjF66|oCW<q!TQhjZF>?&qRIH_zI->Ew<
z<Z5G-T3xm)&;_puy#)bpBirGUtjqN1mtyz=H>&r@8_i9GN_=2$@XBp~ev2bwG@6*|
zrUdvxIe&N$8i<~lX@PTw5O!_OB?Jbp%n;Y`sNmSIsp7HBsF&f20xsbnsIDGOSD@W4
z2E@9_f0=q#<!$v6kOcxkI|Hmj2_uYfQs^Cd)X`8p?<KcE<2nO$ZJ#d6sxpTxPtjWF
zps&MYcH(=YOgU>NR;CDtMW7dh_vxJN!zao)N6a*<rHCX}=3?5^;RMzquY^HsJr{MF
zGK_xzDhK|Uh8#7T3?NpJ&O538-Dhj&OXhlBZfR}?4b3rBb|Xw;GVp<njHK6;vc6A_
z+RX7=H_RM2O-4p0fvUxk^WF|#N>+xlkC_xz&y3&3%2`W{gHH$QNUkS3MYXW7?7k|z
zq2BBGr>`GYlPZ2o;imlnSd_(D!2O^%nq=h4vuORQ(hcHxIWc1flYzpW<$pU_J|Ik4
zF>u@NPLNyaZyPENI^1#FK7mqZ`qZb{V1a$=#yO}FF{fWDUX__Il4ZGQq4*Oo2AYW$
zrq_kE3`|^2LN!&yKC;ON&x)ECO0a@!y?4RxnY=icadCU&k-YDi`ghK01Kl#9@a5&Y
zLj%RBUqVhDxb~TQ>hu$5P2VYm44^}_8DH%P%!Xj+{8!}1&X()phS-$eyY1*J!Q+O4
zA;~TkXQWod17`-I09zkuYraL#?AD`$FsVh7E+jIePuP|ys$OgCHzBhWD>m<a-Q{*~
z^F6kmJ$m~*d6TUk(Gn#Y15#(6IG60B39=>B<@pe!^};WBf_gS){}*(B2cAE=o}4&o
zEcXkHI1Yy;#jYo?*j7USzV~f_4hZi!(?Hh0^SU%d<)u9_I6v8vnIb)s``tY}%tHL`
zr0kou_{>ToW764jYVX<VBwY^j+`PUk@>8#Chu0082zcR^?ycgGc65ce*>SuSSJ`It
z7+f{%8U*!M%`v$Ba-?s923$v($*3aaDkP(K-hGVM8yy^>5-?SdE>Esu?lggDy~WB=
zOMkvvS#y#j3!AGh6q^$<ww;q*`vz@T$`iVZ;m>@lPM#tAXZb6HB<KX9&hyGjlX}E_
zR<QNjLkF5DLgN-n>guQHo#h6*43@KgJ3mF-55J0l87)U=;Cl*f_wSj|)?GwHwtXHP
zU{bSxZi4S`eOc7xT-v?^VxFhj2QM){QAW5El6)ZPGc_DervWm3HF;NR-X%XS@PDLj
zp_?wmYMlIF0xyJ+yb*W*ZxtuzU~XqCsA+Tw<G%xy9ZlK^5t;?3f*{9A=UVS69ueMi
zA>n!`pE69X3AI85Ue;pSu6$j`l`qBidXbqg!vtCKO(Nd@CA|CX%&{i^bM+D$r9p-E
zMy*$W_;($Hw2&5a6Wq7gFL3zl*vdf4Y;h{Xk-ebKMQ;9`=GoDHrGhrM40)H?K<$rs
zuSb23ey{y+5HJu(DDuWY6a3FWYlvTSVv8^8im;u%`dc%?Rjbm#+3e>rwWgkB%S|^3
z#F|WM2O*Drz>Mj|t2q$E=7yfU5Q{0V2^O=>x_c<4hT4VR&6XE6U)>TLeq@If4b-^*
zomc#42P-irqa?hYP5;N#)8FIDW;igs(%Cy}VSl~iqiy5uA3b?Llt`)GH9de0{P5LB
zhm4UUwe&o54a9rxP_JBn7yP4Ke@*_PS-Pd!2Gd#3N8CD7-*l6-X$zt)xO9YM<h`E-
z&O`~~#XtXNrV=-g<MdsYCL_)JCxwW3JHAAT*QW)ky;ptlu{d#!Q`@>NWmZOxfCieQ
zsV3AVofC=Q%-PZB`G{RWW^}UmU?VZyH6~YyX3IK}tmrugw-P?S&|Hf^X0>OxNVw;&
z^3Y!zcIrD4_2T#ddFooqj>MvUE`Upi7=GR@-x1sb&a0`t+{7c0sd#?^2Q2FL=`M~<
z%%;=eHOduL4e9#bSRo!K*$%96l&fS3Sl0kQFh1LaA$h6=oW7xp1uk3Lt|%@l-G>Yv
zCfpVBhqDvH$S2Y_OpFISIBy+1>Ros`@{Xi&dCAv;97Eil1aXboqwQ|T<Oi;^2AuED
z{M|OF5;LKX-%h=iYFjq8#|=bS@sxA(%Wj%oW!Fw-J()pv)txb|$`sI~N%l|xshvl|
zu=y!gN3))|gqnMB_1O!CFmAc`>1_D=M)m#5l0PAI0h>&(NZz`Nf&&(zKm;?Tnj?I9
zt47bePut3b{M%=+b6sVuPvgV>!toN*b`vb$#~X~cgOQ_^m*DWP^v5(oh0$eR>c}>k
z9Bv(xP_Od}-X~StCC4(vhvRX&4L;n<S6U-)+8P8IHa<uQ7TYO&JI?!2>~m!_pBTRT
zxMVp-HAyKf>?i6x?0F%L`pBfBDj@w7vp!R)HRJ-{N~o0d*;BrlGI7LyqkOU`@_Ft<
zUxia>@pWDzk`f6y11%H+(<Lily)faD$Eah=DM*s=dhtL<8Y!=~;4%nnFtQiXAp&t;
zkJWqX!?t4iv99M4w&AJWsn{<~`f>qMzYd2j6aK@c`~2|!DaGCRV!PC{cBWpT{dt<&
zXl6pFLYM%l?%K}wB1_rJWl0abx5hFrC*0Ys7mg$jMVJN?$`OiDT}xI;L_u(Mw2wtw
z^vOxDDlS&b&FA~a7&o19ScTK_)Vvssll|%<69n8t$1hq-_SyqP<Uf9QoH9+FEVrJZ
z`l{&Ip!LH!$nEP{soV4+6~>H9j>hw&&@Tk;&O##h(#FVdVq(BowE?&-K$hdkI10i~
zcEWVRSTRkqJZw0i4I&vFCwD;xS3bKeD3bkZokJ?@vpbH342t02dHsFiv>5F>N=;A(
zQ-!x|yf?364|)Y1nWA|Ya<3v-Q#)2Q%ekg53nm1$G$el1ah>ueCo3;5A$>y?Yul{f
zNkjig5&kE&%+KZU*CBJMvW1K$e|Oi_u~}llE<?;Sj@v{wU{-n8<*iDzTj%KL1rbp!
zQZ>nplYhO9u8)A+Dfl;r6`$Mbjw`Nh|93dpL*cg~gUqFUhyGQc;0y~ieJPQ9N?0*=
zgwLz`w~l(U>hDWKSpD$QkBzlrB#(itYK2@{tXU@P>SA_m`_SQs=6t*@IgrSxBK}L=
zGuM{k+F7O{EtF=?lBTxUlz0UsaZ+19C;w7xlZ!N{=)ED#W6RF5mfvD?Lrs5`>B@Q6
zYYO{;z>z*J>z~Vm0A@p3Fui=`vIMWj-G)3;dzl+CN+|-~8*rBH@<bhWMX~}6v6b(6
zyv)R>{K&6DEnYMA^0wupTouuTyh}%aokawT%><#csrdylnPVF@atoX?TRW0R%wJ<l
z1iSUbsA>-n-G=@0yK<?LL{3#Z-U}xUKcB9jG)*h4(o3{qCR1uWd7f^N8ENQ+HDP5a
zOx?>MLx(+kTqn=Du!p6>7$@iOo<^u|LX8^8XL3N#$TFt$G)w`pT>7T8StfF)YeGa=
zxBNcx#jQR;Ax2XxheDJj&gt^C_mV<Fu&I~6DDXxc1w8VYCOu-9^qr+DCed?DE}9IH
zwR-??^R7R+uuEu6sMi&rP|Z-l?QBd%8?7qjzw=Itg(U=L2`>ez53a)UqGPlc8}UTS
zLOj@Ua=7o@o{8s534HGGaY=!C7eH+~LAF1g41$TpZN&{<1b!6x56nU<1MnST-$q%B
z{c!BLhsc2S*CX8(4Z*?T&bZ=ax9;Q`6^*ky;8<9In=&`0FPgvJ-3cW+2sTJEJco%9
zYM{yyZC=P5H2nI-G2n$|R4y5pf14Zmd7iCNFbojid{D|9(~sZJOE0V}R9SNG7zH=p
zgt9=OOIOIYtOwf%<_)y}aDqQejF=KMARLPS;DG)C+eeK9xA+Bs(%b}Nfw7FLuE9{C
zeNqWha5Wo#6IJtpkJZ1n=-GH85Niq4k$h-Ch;ut57~KQj>4+Nb_+5&{(GwZ>l~F^3
z?>MmYhcJ^JM*61LE={2uLlXcJArPT}d%JgV#CpK#X$BZAr08sXOdJH@nHul*vC0U%
zVN8|E_nxJW)-io3xhtg2rBz_I{<9X!MPAeJz3+XhO=|&Ww-6)iu`-jIo|qPKh*Z|_
znx3%}@f3T_L+@7M1@*ioORuT>0==~%YHV@;o$5Yh!nRDLk*u1QG!QZ3m44v%PpU9o
zK!<8_sS@mTH{sB~N*beYf!@-TwW_>S?Z3&8GVTF$$>o`=tz%A+=a#*4I-vnT-oM(y
zxhz(F{I=)0G%yEVYHzH>%sV&O##R4py(qUAjT<@4>W;q*4j!A}OXZ}BS6V9beA1qw
zTH9(KE+1W`EL=w@3WRD$3&N)+7B#j*XGWd#ktH6c<OnshnsUxflQI{LEOACY$XAVm
z6t%}{%9zi^{3S4D6i@d4Bx~N032o{?I64S+Uis`^&UyH)?k%FIK%eNiPK5pSM|XwU
zR=+XmGs84(v#qvEH+6G&v>4@GSw@f7FP@c)-WVKv?@eceepjMVTyc@e#X%5kDVsAw
zdA-Gh@L3-Mw7ZsQ{?|VLU-l=7Klxb2V>vipN{s91hL&sQDR~<>wmbiEh9s}!(eEE}
zyufc!o6k)VH7+F%8~xBXxvr4f^^@IHM^-CZj5dF}SHZQ2NRyJ^Y`{ZxXjWxqswNA#
zS0g?=fzM3@wxp+v7_(GZ`H>YJ{xVIzC%lvJz=#?js30BE{xqP(r{MEjDb_Ao;w<$X
zu<y5m@{^OZVM6Ib;zYqBt<=s$y!gkjCO_Ix@|l_A<xcO4AoUU$B*x?NY?B0zdb6yP
zA}Cr8H3@1HMLu0ib3@q$%<ODiYsmKS=VB$wZm9L|XBsBw4s0GzEp?K=<9mq{NTN0i
z0arrUW!`9Iy35jHgUEKT1r}D5V#J7x^=?-8!Nca_r;HQ=kftr`{tn59`Np{-6=zWe
z&h;E;-)Qa%#Cv(~WMNCW&iYJf)L0<8EU8yybrU$qb*p_gL($41{Yq*V@p+m3V;BNC
z`i5SA&64Zq$|ab3Bb_81j|&x+O=dfL*UOE}g)ZXxgTi*g|A=^>wRoa7`P*KETz|sc
zeHB=JO-~$3#AlIf{{i)#G}9#;AC+O)R{|^wvb5NSquo+{8Ut$o=f;D+Gd~@u&npfW
zuxk@Cfn0HK!9sGQhl7cc`n?rX<lWg%L#jO%Q<P@Zr2+)<Xq-8mr%zlR6X}4?JUbb^
z*t1_YkTWs`=+dlP@or$Tiq4Iu;ES`UPoMefubtRE_l@I^7$iTBN+mmm2D`0ZF*7k%
zU|4j#d{1LpBbH^Uw^h}mK;!4a?U`N{X~`a~>7}iw>)!vAAS<4ohQIB*zq3!$PshS+
z<y?v<$deG2aLbCCMM)&A5Qax^zBN)@O)?q0Ltuf(e@|7iL#d&WSft)sF|qW0ULYes
zFY4i_t4?b&7IhR^Aw_=aRL7(2QRamL?uo?=r7&)~aOlb$+n3gK*pFo4hL7eB3Vpfs
z)tic5@ZD1l>ZQ*N)x8Yw9jNnrnK33RpGR{*-`e1kY>x*2+ZcbptVwF)XW*89*tO*=
zh9O(DQ@=p0dlLfKKQ|9dv#Eg}DrEAcawQ3tF8U26X)!V&yxrW7k%ud;0WCkWOf9}X
zco?3Tx?$I$ySR(HLa{bEXm_r+tSH%9H!;{mrz6O+uI9ep{xCxf^Eg-2Cw7noqHyjf
za@k_wO1n0iphtfa{q(ZsByxDhnumb0zU=*GGnK!y_ei)q|M7^~vf^Q+Jj-O^)@0>E
zcSX3cL{1q-+Q8b*M~P=QkqP2BTJH~;i?fPwu(Zb$TRQQXzK0$MoYEc9a2$6-n#=ph
z2IS{qA?K?&;F?zLWS_c1fPfk%{n1H+lU1$0&6;XX!TcY(dcO)J&3(??)dHsR0#cQ{
zEPvN5%2JdG{l37$#*Y4~<80iwfe9m7n|DrU>RI^em04gw7UMQUdmG5`DEWY@ihzK4
z_9TxTwM$9Jb}-lPeKh;P&<{^AclO%P0!vq0NA9RI_0-2l^0P!iX_$PB%T2#AD$cO*
zyW!qnuO8M>+2E7rv*{M8cKlo1#d6Uwp(060K3`T|nq0(cRa=B-udvJZF5@~<;0DSW
zGkhm46O!D100D}_>|FLdCt4O4i*!o;GB4|#^7v`#5Ar;F9myh?(Gv@O#@?>Av#FTn
z=%-YZhIOluk5?Eotm~$@V3+ZrLT(MB-SNyHT>Ax$PX`e;8c*5Zl}uW^7$WNqxC@2e
zjo6*I6_Gdf2EWJ2%_ekm<#e|YQ|j*U=A7=5WLf3=T{F2uWrwk=boKnyZTs3$F~xk$
zCiZElRyw*$LLP2F1NwD3nP?U{>Xu_Un&AqOxwzskT?YTvOt=XH2tKuXdU(mr#&Obd
zReWS<nw5?d=Z0iF+bB<dy2Pt<ZiN%lr+smd8a8oQ&&aQruhFkzL50Hm487LXjflY2
zJKdtr{*NI=9(i55hzq@pve6-BNd4(~BAILN_BIdu*;My10n2^2`iq(J<=xDzLU-ef
z9yNvjDECDU_(k@5y!A4nL9~;Z@OF|HHoO!)9yQM7t0y%(85#IM38uO$bx-WtM;6_1
zX5Em54RuS=<Vwhd1l2owzM%s1_7rd1jLtfDxSVJod2X-{Kj8*4-=<r6$V{AkPkYLg
zRLvwlO!0lVL{lR6vVOKD(uJ3<6j2CTE#Km?h}ze2HrCG5oSSbHzIQaSh8myuxt)||
z8<Mc_%G{$RrlP6AfH0@J`h8<q>*kltDdg35rOzdu-=eHZJg#gF;b;A=DyOureT}Fy
zr}Mj)wSlfL?xfy`$<pbr3bz>}EGCI4{&Z_$1Py&5%DZkES|9RDiqESiY^J!lel<{w
z`QatmPaYa6&fl^fa>8Qs>6+!xu<%zbj?u2L^bktO(ZKVZ7?e$UwVAhp((wj~s7n^Q
zAQ>Jm^YfeS7oo6Jj@-|fbCrPv+89kv229a1kk=ASgsgbh*~9e;IArriWYy2xq&lRk
zNB#_Lt)@4B4c461K95ohh++V)bj>YOpo+%A^AnF`)rwf?pKJ8Q_61aEPYE)vmjvB)
zozUL}kPV-rz3}q>i|ljt(4TxO(zZUkuDUDvYB&>dye5cY?857M#ejl-lXu!1YZ#Wy
zo`Xbq9H+xBo9s(-nqCqZcO@Yi0`B1sc4@}S9V}!y)kl!Qc2@UUee^YX9n^Fz-CCcO
z3naVe$#(0Z`hWVU8~Ui**6}{);%Bcj)2^!75{(L?@!CoAy^22>x-hcvJ_wJGvmN2$
z*7-n9TCR}?cMAH2*DvwFerDS=$o^I_DLwX@n4D38861=;PCm&GUBh><9(zzI01NxA
zmcM?!A~JGy!f%a;wE0cTtsBJv6cO_m{Ov<Xy~ACJJ*Y39I+|%w?nPQT%U`qHV}UDI
zvt?YA&0i*5ZUaL^T?~YkX3CI3E%kGC#$bxNE&8rHhi+TMrvK#UG-0N{$k$i1lDxmj
zw)A8-xJ}vHzQ0CaPO<!iV5i!1%#<C$<>0#8w0pU`e1L|F<c=|%F5b`tlP0R=oV7JQ
zG*r{)#H}-l>Q#mqcjS7->k6^(zro!W3)aX%*Y$wO*y(o{7HM6bOYbI*j%g=)>8?}+
z4s|0dJ?k1e>*008cw<M8(m-G{pKIVnxO{1KIB0Xg_(e%_cqsOhUo-G&^D0R7`q$xi
z)YR;x)Hns@eLF+>ior=wiqx>FCly9lk0td#h}|*adSGmBZ$JYX@gK}phB9ekI?g28
z>QRhcMizQxKf?Y(&^$7(vpHWyo<_z^Euc1aR%TE-=#-b4`o<P7sXvm)#OJpQ%%IUn
z=w=V^{RlD;EXJ%aw1eEhj-USFB#mi-wKoWna665EV~gL7Y5{?6b`c14(-Q>3#k$q;
zVR8_(e@{C#HF(Rq_xtbBwd2s(aYq6a1@d0Pybkv@F4FU{`OAPr9B{D|Z$b^8@_FGo
zmf1#>fBUKUX+PhLeA>x3vy>ai*mL)SRzJ4wWmWU~HTmsRgo0vR=gY=qTAE&?Cwj&=
zT7R!N{g2s@^90SPq&)h?Ga(i*{Q=2J6QRb9Z!p&#e=hY+@P~V&-_w`4eevfQhit*1
z+C;P9YmccxQ*FO3JZ<yLnhck%4{N`DT4ACyK~d?(Dh{MWc%wo@P^z`%VL6}r(K`tl
z9j@5HXJC2sn;9VH;Uz2bI?B3YDqJ`>IDISS=`uS4EjGsb@9*BwN?9dMPy7A$p8_ok
zr5RAH&#QyXqIkqlNCY4WHVzsGJ?VeAGX4Kew%d?B;>OYXf9Chb_6C7&W)C2){quPK
zo$&wvAOD#ypxZUPecJ!`$N!e_e-ClHuK)YT|2^*Q(^96c?t?&ASGv{?<Raq}H-S+Z
MNkxf5F(cpq1GWG7z5oCK

literal 0
HcmV?d00001

diff --git a/doc/src/week46/DataFiles/ride.csv b/doc/src/week46/DataFiles/ride.csv
new file mode 100644
index 000000000..d03d4ca16
--- /dev/null
+++ b/doc/src/week46/DataFiles/ride.csv
@@ -0,0 +1,15 @@
+Outlook,Temperature,Humidity,Wind,Ride
+0,0,0,0,0
+0,0,0,1,1
+1,0,0,0,1
+2,1,0,0,1
+2,2,1,0,1
+2,2,1,1,0
+1,2,1,1,1
+0,1,0,0,0
+0,2,1,0,1
+2,1,1,0,1
+0,1,1,1,1
+1,1,0,1,1
+1,0,1,0,1
+2,1,0,1,0
diff --git a/doc/src/week46/DataFiles/ride.dot b/doc/src/week46/DataFiles/ride.dot
new file mode 100644
index 000000000..50aaa7638
--- /dev/null
+++ b/doc/src/week46/DataFiles/ride.dot
@@ -0,0 +1,13 @@
+digraph Tree {
+node [shape=box, style="filled, rounded", color="black", fontname=helvetica] ;
+edge [fontname=helvetica] ;
+0 [label="X[7] <= 0.5\ngini = 0.48\nsamples = 15\nvalue = [4, 10, 1]", fillcolor="#39e5818b"] ;
+1 [label="X[1] <= 0.5\ngini = 0.408\nsamples = 14\nvalue = [4, 10, 0]", fillcolor="#39e58199"] ;
+0 -> 1 [labeldistance=2.5, labelangle=45, headlabel="True"] ;
+2 [label="gini = 0.48\nsamples = 10\nvalue = [4, 6, 0]", fillcolor="#39e58155"] ;
+1 -> 2 ;
+3 [label="gini = 0.0\nsamples = 4\nvalue = [0, 4, 0]", fillcolor="#39e581ff"] ;
+1 -> 3 ;
+4 [label="gini = 0.0\nsamples = 1\nvalue = [0, 0, 1]", fillcolor="#8139e5ff"] ;
+0 -> 4 [labeldistance=2.5, labelangle=-45, headlabel="False"] ;
+}
\ No newline at end of file
diff --git a/doc/src/week46/DataFiles/rideclass.csv b/doc/src/week46/DataFiles/rideclass.csv
new file mode 100644
index 000000000..a0831ac7e
--- /dev/null
+++ b/doc/src/week46/DataFiles/rideclass.csv
@@ -0,0 +1,15 @@
+Day,Outlook,Temperature,Humidity,Wind,Ride
+1,Sunny,Hot,High,Weak,0
+2,Sunny,Hot,High,Strong,1
+3,Overcast,Hot,High,Weak,1
+4,Rain,Mild,High,Weak,1
+5,Rain,Cool,Normal,Weak,1
+6,Rain,Cool,Normal,Strong,0
+7,Overcast,Cool,Normal,Strong,1
+8,Sunny,Mild,High,Weak,0
+9,Sunny,Cool,Normal,Weak,1
+10,Rain,Mild,Normal,Weak,1
+11,Sunny,Mild,Normal,Strong,1
+12,Overcast,Mild,High,Strong,1
+13,Overcast,Hot,Normal,Weak,1
+14,Rain,Mild,High,Strong,0
diff --git a/doc/src/week46/DataFiles/zoo.csv b/doc/src/week46/DataFiles/zoo.csv
new file mode 100644
index 000000000..ca71f7d21
--- /dev/null
+++ b/doc/src/week46/DataFiles/zoo.csv
@@ -0,0 +1,101 @@
+aardvark,1,0,0,1,0,0,1,1,1,1,0,0,4,0,0,1,1
+antelope,1,0,0,1,0,0,0,1,1,1,0,0,4,1,0,1,1
+bass,0,0,1,0,0,1,1,1,1,0,0,1,0,1,0,0,4
+bear,1,0,0,1,0,0,1,1,1,1,0,0,4,0,0,1,1
+boar,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,1,1
+buffalo,1,0,0,1,0,0,0,1,1,1,0,0,4,1,0,1,1
+calf,1,0,0,1,0,0,0,1,1,1,0,0,4,1,1,1,1
+carp,0,0,1,0,0,1,0,1,1,0,0,1,0,1,1,0,4
+catfish,0,0,1,0,0,1,1,1,1,0,0,1,0,1,0,0,4
+cavy,1,0,0,1,0,0,0,1,1,1,0,0,4,0,1,0,1
+cheetah,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,1,1
+chicken,0,1,1,0,1,0,0,0,1,1,0,0,2,1,1,0,2
+chub,0,0,1,0,0,1,1,1,1,0,0,1,0,1,0,0,4
+clam,0,0,1,0,0,0,1,0,0,0,0,0,0,0,0,0,7
+crab,0,0,1,0,0,1,1,0,0,0,0,0,4,0,0,0,7
+crayfish,0,0,1,0,0,1,1,0,0,0,0,0,6,0,0,0,7
+crow,0,1,1,0,1,0,1,0,1,1,0,0,2,1,0,0,2
+deer,1,0,0,1,0,0,0,1,1,1,0,0,4,1,0,1,1
+dogfish,0,0,1,0,0,1,1,1,1,0,0,1,0,1,0,1,4
+dolphin,0,0,0,1,0,1,1,1,1,1,0,1,0,1,0,1,1
+dove,0,1,1,0,1,0,0,0,1,1,0,0,2,1,1,0,2
+duck,0,1,1,0,1,1,0,0,1,1,0,0,2,1,0,0,2
+elephant,1,0,0,1,0,0,0,1,1,1,0,0,4,1,0,1,1
+flamingo,0,1,1,0,1,0,0,0,1,1,0,0,2,1,0,1,2
+flea,0,0,1,0,0,0,0,0,0,1,0,0,6,0,0,0,6
+frog,0,0,1,0,0,1,1,1,1,1,0,0,4,0,0,0,5
+frog,0,0,1,0,0,1,1,1,1,1,1,0,4,0,0,0,5
+fruitbat,1,0,0,1,1,0,0,1,1,1,0,0,2,1,0,0,1
+giraffe,1,0,0,1,0,0,0,1,1,1,0,0,4,1,0,1,1
+girl,1,0,0,1,0,0,1,1,1,1,0,0,2,0,1,1,1
+gnat,0,0,1,0,1,0,0,0,0,1,0,0,6,0,0,0,6
+goat,1,0,0,1,0,0,0,1,1,1,0,0,4,1,1,1,1
+gorilla,1,0,0,1,0,0,0,1,1,1,0,0,2,0,0,1,1
+gull,0,1,1,0,1,1,1,0,1,1,0,0,2,1,0,0,2
+haddock,0,0,1,0,0,1,0,1,1,0,0,1,0,1,0,0,4
+hamster,1,0,0,1,0,0,0,1,1,1,0,0,4,1,1,0,1
+hare,1,0,0,1,0,0,0,1,1,1,0,0,4,1,0,0,1
+hawk,0,1,1,0,1,0,1,0,1,1,0,0,2,1,0,0,2
+herring,0,0,1,0,0,1,1,1,1,0,0,1,0,1,0,0,4
+honeybee,1,0,1,0,1,0,0,0,0,1,1,0,6,0,1,0,6
+housefly,1,0,1,0,1,0,0,0,0,1,0,0,6,0,0,0,6
+kiwi,0,1,1,0,0,0,1,0,1,1,0,0,2,1,0,0,2
+ladybird,0,0,1,0,1,0,1,0,0,1,0,0,6,0,0,0,6
+lark,0,1,1,0,1,0,0,0,1,1,0,0,2,1,0,0,2
+leopard,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,1,1
+lion,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,1,1
+lobster,0,0,1,0,0,1,1,0,0,0,0,0,6,0,0,0,7
+lynx,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,1,1
+mink,1,0,0,1,0,1,1,1,1,1,0,0,4,1,0,1,1
+mole,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,0,1
+mongoose,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,1,1
+moth,1,0,1,0,1,0,0,0,0,1,0,0,6,0,0,0,6
+newt,0,0,1,0,0,1,1,1,1,1,0,0,4,1,0,0,5
+octopus,0,0,1,0,0,1,1,0,0,0,0,0,8,0,0,1,7
+opossum,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,0,1
+oryx,1,0,0,1,0,0,0,1,1,1,0,0,4,1,0,1,1
+ostrich,0,1,1,0,0,0,0,0,1,1,0,0,2,1,0,1,2
+parakeet,0,1,1,0,1,0,0,0,1,1,0,0,2,1,1,0,2
+penguin,0,1,1,0,0,1,1,0,1,1,0,0,2,1,0,1,2
+pheasant,0,1,1,0,1,0,0,0,1,1,0,0,2,1,0,0,2
+pike,0,0,1,0,0,1,1,1,1,0,0,1,0,1,0,1,4
+piranha,0,0,1,0,0,1,1,1,1,0,0,1,0,1,0,0,4
+pitviper,0,0,1,0,0,0,1,1,1,1,1,0,0,1,0,0,3
+platypus,1,0,1,1,0,1,1,0,1,1,0,0,4,1,0,1,1
+polecat,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,1,1
+pony,1,0,0,1,0,0,0,1,1,1,0,0,4,1,1,1,1
+porpoise,0,0,0,1,0,1,1,1,1,1,0,1,0,1,0,1,1
+puma,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,1,1
+pussycat,1,0,0,1,0,0,1,1,1,1,0,0,4,1,1,1,1
+raccoon,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,1,1
+reindeer,1,0,0,1,0,0,0,1,1,1,0,0,4,1,1,1,1
+rhea,0,1,1,0,0,0,1,0,1,1,0,0,2,1,0,1,2
+scorpion,0,0,0,0,0,0,1,0,0,1,1,0,8,1,0,0,7
+seahorse,0,0,1,0,0,1,0,1,1,0,0,1,0,1,0,0,4
+seal,1,0,0,1,0,1,1,1,1,1,0,1,0,0,0,1,1
+sealion,1,0,0,1,0,1,1,1,1,1,0,1,2,1,0,1,1
+seasnake,0,0,0,0,0,1,1,1,1,0,1,0,0,1,0,0,3
+seawasp,0,0,1,0,0,1,1,0,0,0,1,0,0,0,0,0,7
+skimmer,0,1,1,0,1,1,1,0,1,1,0,0,2,1,0,0,2
+skua,0,1,1,0,1,1,1,0,1,1,0,0,2,1,0,0,2
+slowworm,0,0,1,0,0,0,1,1,1,1,0,0,0,1,0,0,3
+slug,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0,0,7
+sole,0,0,1,0,0,1,0,1,1,0,0,1,0,1,0,0,4
+sparrow,0,1,1,0,1,0,0,0,1,1,0,0,2,1,0,0,2
+squirrel,1,0,0,1,0,0,0,1,1,1,0,0,2,1,0,0,1
+starfish,0,0,1,0,0,1,1,0,0,0,0,0,5,0,0,0,7
+stingray,0,0,1,0,0,1,1,1,1,0,1,1,0,1,0,1,4
+swan,0,1,1,0,1,1,0,0,1,1,0,0,2,1,0,1,2
+termite,0,0,1,0,0,0,0,0,0,1,0,0,6,0,0,0,6
+toad,0,0,1,0,0,1,0,1,1,1,0,0,4,0,0,0,5
+tortoise,0,0,1,0,0,0,0,0,1,1,0,0,4,1,0,1,3
+tuatara,0,0,1,0,0,0,1,1,1,1,0,0,4,1,0,0,3
+tuna,0,0,1,0,0,1,1,1,1,0,0,1,0,1,0,1,4
+vampire,1,0,0,1,1,0,0,1,1,1,0,0,2,1,0,0,1
+vole,1,0,0,1,0,0,0,1,1,1,0,0,4,1,0,0,1
+vulture,0,1,1,0,1,0,1,0,1,1,0,0,2,1,0,1,2
+wallaby,1,0,0,1,0,0,0,1,1,1,0,0,2,1,0,1,1
+wasp,1,0,1,0,1,0,0,0,0,1,1,0,6,0,0,0,6
+wolf,1,0,0,1,0,0,1,1,1,1,0,0,4,1,0,1,1
+worm,0,0,1,0,0,0,0,0,0,1,0,0,0,0,0,0,7
+wren,0,1,1,0,1,0,0,0,1,1,0,0,2,1,0,0,2
diff --git a/doc/src/week46/backup2022.do.txt b/doc/src/week46/backup2022.do.txt
new file mode 100644
index 000000000..1ea619db0
--- /dev/null
+++ b/doc/src/week46/backup2022.do.txt
@@ -0,0 +1,1171 @@
+TITLE: Week 46: Support Vector Machines and Project 3. 
+AUTHOR: Morten Hjorth-Jensen {copyright, 1999-present|CC BY-NC} at Department of Physics, University of Oslo & Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University
+DATE: today
+
+!split
+===== Overview of week 46 =====
+
+* _Thursday_: Support Vector Machines, discussion of project 3
+  * "Video of lecture":"https://youtu.be/F3CkH-opbdY"
+* _Friday_: Support vector machines and start Principal Component Analysis
+  * "Video of lecture":"https://youtu.be/BbupEEvMXtg"
+
+!bblock Reading
+Reading recommendations:
+  o See lecture notes for week 46 at https://compphysics.github.io/MachineLearning/doc/web/course.html.
+  o Hastie et al chapter 12
+  o Bishop chapter 7.1 and 7.2
+!eblock
+
+
+!bblock Videos
+o "Overview video on Support Vector Machines":"https://www.youtube.com/watch?v=efR1C6CvhmE&ab_channel=StatQuestwithJoshStarmer"
+o See also "this video":"https://www.youtube.com/watch?v=N1vOgolbjSc&ab_channel=AliceZhao".
+!eblock
+
+
+!split
+===== Support Vector Machines, overarching aims  =====
+
+A Support Vector Machine (SVM) is a very powerful and versatile
+Machine Learning method, capable of performing linear or nonlinear
+classification, regression, and even outlier detection. It is one of
+the most popular models in Machine Learning, and anyone interested in
+Machine Learning should have it in their toolbox. SVMs are
+particularly well suited for classification of complex but small-sized or
+medium-sized datasets.  
+
+The case with two well-separated classes only can be understood in an
+intuitive way in terms of lines in a two-dimensional space separating
+the two classes (see figure below).
+
+The basic mathematics behind the SVM is however less familiar to most of us. 
+It relies on the definition of hyperplanes and the
+definition of a _margin_ which separates classes (in case of
+classification problems) of variables. It is also used for regression
+problems.
+
+With SVMs we distinguish between hard margin and soft margins. The
+latter introduces a so-called softening parameter to be discussed
+below.  We distinguish also between linear and non-linear
+approaches. The latter are the most frequent ones since it is rather
+unlikely that we can separate classes easily by say straight lines.
+
+!split
+===== Hyperplanes and all that =====
+
+The theory behind support vector machines (SVM hereafter) is based on
+the mathematical description of so-called hyperplanes. Let us start
+with a two-dimensional case. This will also allow us to introduce our
+first SVM examples. These will be tailored to the case of two specific
+classes, as displayed in the figure here based on the usage of the petal data.
+
+We assume here that our data set can be well separated into two
+domains, where a straight line does the job in the separating the two
+classes. Here the two classes are represented by either squares or
+circles.
+!bc pycod
+from sklearn import datasets
+from sklearn.svm import SVC, LinearSVC
+from sklearn.linear_model import SGDClassifier
+from sklearn.preprocessing import StandardScaler
+import matplotlib
+import matplotlib.pyplot as plt
+plt.rcParams['axes.labelsize'] = 14
+plt.rcParams['xtick.labelsize'] = 12
+plt.rcParams['ytick.labelsize'] = 12
+
+
+iris = datasets.load_iris()
+X = iris["data"][:, (2, 3)]  # petal length, petal width
+y = iris["target"]
+
+setosa_or_versicolor = (y == 0) | (y == 1)
+X = X[setosa_or_versicolor]
+y = y[setosa_or_versicolor]
+
+
+
+C = 5
+alpha = 1 / (C * len(X))
+
+lin_clf = LinearSVC(loss="hinge", C=C, random_state=42)
+svm_clf = SVC(kernel="linear", C=C)
+sgd_clf = SGDClassifier(loss="hinge", learning_rate="constant", eta0=0.001, alpha=alpha,
+                        max_iter=100000, random_state=42)
+
+scaler = StandardScaler()
+X_scaled = scaler.fit_transform(X)
+
+lin_clf.fit(X_scaled, y)
+svm_clf.fit(X_scaled, y)
+sgd_clf.fit(X_scaled, y)
+
+print("LinearSVC:                   ", lin_clf.intercept_, lin_clf.coef_)
+print("SVC:                         ", svm_clf.intercept_, svm_clf.coef_)
+print("SGDClassifier(alpha={:.5f}):".format(sgd_clf.alpha), sgd_clf.intercept_, sgd_clf.coef_)
+
+# Compute the slope and bias of each decision boundary
+w1 = -lin_clf.coef_[0, 0]/lin_clf.coef_[0, 1]
+b1 = -lin_clf.intercept_[0]/lin_clf.coef_[0, 1]
+w2 = -svm_clf.coef_[0, 0]/svm_clf.coef_[0, 1]
+b2 = -svm_clf.intercept_[0]/svm_clf.coef_[0, 1]
+w3 = -sgd_clf.coef_[0, 0]/sgd_clf.coef_[0, 1]
+b3 = -sgd_clf.intercept_[0]/sgd_clf.coef_[0, 1]
+
+# Transform the decision boundary lines back to the original scale
+line1 = scaler.inverse_transform([[-10, -10 * w1 + b1], [10, 10 * w1 + b1]])
+line2 = scaler.inverse_transform([[-10, -10 * w2 + b2], [10, 10 * w2 + b2]])
+line3 = scaler.inverse_transform([[-10, -10 * w3 + b3], [10, 10 * w3 + b3]])
+
+# Plot all three decision boundaries
+plt.figure(figsize=(11, 4))
+plt.plot(line1[:, 0], line1[:, 1], "k:", label="LinearSVC")
+plt.plot(line2[:, 0], line2[:, 1], "b--", linewidth=2, label="SVC")
+plt.plot(line3[:, 0], line3[:, 1], "r-", label="SGDClassifier")
+plt.plot(X[:, 0][y==1], X[:, 1][y==1], "bs") # label="Iris-Versicolor"
+plt.plot(X[:, 0][y==0], X[:, 1][y==0], "yo") # label="Iris-Setosa"
+plt.xlabel("Petal length", fontsize=14)
+plt.ylabel("Petal width", fontsize=14)
+plt.legend(loc="upper center", fontsize=14)
+plt.axis([0, 5.5, 0, 2])
+
+plt.show()
+
+
+
+!ec
+
+
+
+
+!split
+===== What is a hyperplane? =====
+
+The aim of the SVM algorithm is to find a hyperplane in a
+$p$-dimensional space, where $p$ is the number of features that
+distinctly classifies the data points.
+
+In a $p$-dimensional space, a hyperplane is what we call an affine subspace of dimension of $p-1$.
+As an example, in two dimension, a hyperplane is simply as straight line while in three dimensions it is 
+a two-dimensional subspace, or stated simply, a plane. 
+
+In two dimensions, with the variables $x_1$ and $x_2$, the hyperplane is defined as
+!bt
+\[
+b+w_1x_1+w_2x_2=0,
+\]
+!et
+
+where $b$ is the intercept and $w_1$ and $w_2$ define the elements of a vector orthogonal to the line 
+$b+w_1x_1+w_2x_2=0$. 
+In two dimensions we define the vectors $\bm{x} =[x1,x2]$ and $\bm{w}=[w1,w2]$. 
+We can then rewrite the above equation as 
+
+!bt
+\[
+\bm{x}^T\bm{w}+b=0.
+\]
+!et
+
+!split
+===== A $p$-dimensional space of features =====
+
+We limit ourselves to two classes of outputs $y_i$ and assign these classes the values $y_i = \pm 1$. 
+In a $p$-dimensional space of say $p$ features we have a hyperplane defines as 
+!bt
+\[
+b+wx_1+w_2x_2+\dots +w_px_p=0.
+\]
+!et
+If we define a 
+matrix $\bm{X}=\left[\bm{x}_1,\bm{x}_2,\dots, \bm{x}_p\right]$
+of dimension $n\times p$, where $n$ represents the observations for each feature and each vector $x_i$ is a column vector of the matrix $\bm{X}$, 
+!bt
+\[
+\bm{x}_i = \begin{bmatrix} x_{i1} \\ x_{i2} \\ \dots \\ \dots \\ x_{ip} \end{bmatrix}.
+\]
+!et
+If the above condition is not met for a given vector $\bm{x}_i$ we have 
+!bt
+\[
+b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} >0,
+\]
+!et
+if our output $y_i=1$.
+In this case we say that $\bm{x}_i$ lies on one of the sides of the hyperplane and if 
+!bt
+\[
+b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} < 0,
+\]
+!et
+for the class of observations $y_i=-1$, 
+then $\bm{x}_i$ lies on the other side. 
+
+Equivalently, for the two classes of observations we have 
+!bt
+\[
+y_i\left(b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip}\right) > 0. 
+\]
+!et
+
+When we try to separate hyperplanes, if it exists, we can use it to construct a natural classifier: a test observation is assigned a given class depending on which side of the hyperplane it is located.
+
+!split 
+===== The two-dimensional case ===== 
+
+Let us try to develop our intuition about SVMs by limiting ourselves to a two-dimensional
+plane.  To separate the two classes of data points, there are many
+possible lines (hyperplanes if you prefer a more strict naming)  
+that could be chosen. Our objective is to find a
+plane that has the maximum margin, i.e the maximum distance between
+data points of both classes. Maximizing the margin distance provides
+some reinforcement so that future data points can be classified with
+more confidence.
+
+What a linear classifier attempts to accomplish is to split the
+feature space into two half spaces by placing a hyperplane between the
+data points.  This hyperplane will be our decision boundary.  All
+points on one side of the plane will belong to class one and all points
+on the other side of the plane will belong to the second class two.
+
+Unfortunately there are many ways in which we can place a hyperplane
+to divide the data.  Below is an example of two candidate hyperplanes
+for our data sample.
+
+!split
+===== Getting into the details =====
+
+Let us define the function
+!bt
+\[
+f(x) = \bm{w}^T\bm{x}+b = 0,
+\]
+!et
+as the function that determines the line $L$ that separates two classes (our two features), see the figure here. 
+
+
+Any point defined by $\bm{x}_i$ and $\bm{x}_2$ on the line $L$ will satisfy $\bm{w}^T(\bm{x}_1-\bm{x}_2)=0$. 
+
+The signed distance $\delta$ from any point defined by a vector $\bm{x}$ and a point $\bm{x}_0$ on the line $L$ is then
+!bt
+\[
+\delta = \frac{1}{\vert\vert \bm{w}\vert\vert}(\bm{w}^T\bm{x}+b).
+\]
+!et
+
+!split
+===== First attempt at a minimization approach =====
+
+How do we find the parameter $b$ and the vector $\bm{w}$? What we could
+do is to define a cost function which now contains the set of all
+misclassified points $M$ and attempt to minimize this function
+
+!bt
+\[
+C(\bm{w},b) = -\sum_{i\in M} y_i(\bm{w}^T\bm{x}_i+b).
+\]
+!et
+
+We could now for example define all values $y_i =1$ as misclassified in case we have $\bm{w}^T\bm{x}_i+b < 0$ and the opposite if we have $y_i=-1$. Taking the derivatives gives us
+!bt
+\[
+\frac{\partial C}{\partial b} = -\sum_{i\in M} y_i,
+\]
+!et
+and 
+!bt
+\[
+\frac{\partial C}{\partial \bm{w}} = -\sum_{i\in M} y_ix_i.
+\]
+!et
+
+!split
+===== Solving the equations =====
+
+We can now use the Newton-Raphson method or different variants of the gradient descent family (from plain gradient descent to various stochastic gradient descent approaches) to solve the equations
+!bt
+\[
+b \leftarrow b +\eta \frac{\partial C}{\partial b},
+\]
+!et
+and
+!bt
+\[
+\bm{w} \leftarrow \bm{w} +\eta \frac{\partial C}{\partial \bm{w}},
+\]
+!et
+where $\eta$ is our by now well-known learning rate. 
+
+
+!split
+===== Code Example =====
+
+The equations we discussed above can be coded rather easily (the
+framework is similar to what we developed for logistic
+regression). We are going to set up a simple case with two classes only and we want to find a line which separates them the best possible way.
+!bc pycod
+
+!ec
+
+!split
+===== Problems with the Simpler Approach =====
+
+
+There are however problems with this approach, although it looks
+pretty straightforward to implement. When running the above code, we see that we can easily end up with many diffeent lines which separate the two classes.
+
+
+For small
+gaps between the entries, we may also end up needing many iterations
+before the solutions converge and if the data cannot be separated
+properly into two distinct classes, we may not experience a converge
+at all.
+
+!split
+===== A better approach =====
+
+A better approach is rather to try to define a large margin between
+the two classes (if they are well separated from the beginning).
+
+Thus, we wish to find a margin $M$ with $\bm{w}$ normalized to
+$\vert\vert \bm{w}\vert\vert =1$ subject to the condition
+
+!bt
+\[
+y_i(\bm{w}^T\bm{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, p. 
+\]
+!et
+All points are thus at a signed distance from the decision boundary defined by the line $L$. The parameters $b$ and $w_1$ and $w_2$ define this line. 
+
+We seek thus the largest value $M$ defined by
+!bt
+\[
+\frac{1}{\vert \vert \bm{w}\vert\vert}y_i(\bm{w}^T\bm{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, n, 
+\]
+!et
+or just 
+!bt
+\[
+y_i(\bm{w}^T\bm{x}_i+b) \geq M\vert \vert \bm{w}\vert\vert \hspace{0.1cm}\forall i. 
+\]
+!et
+If we scale the equation so that $\vert \vert \bm{w}\vert\vert = 1/M$, we have to find the minimum of 
+$\bm{w}^T\bm{w}=\vert \vert \bm{w}\vert\vert$ (the norm) subject to the condition
+!bt
+\[
+y_i(\bm{w}^T\bm{x}_i+b) \geq 1 \hspace{0.1cm}\forall i.
+\]
+!et
+
+We have thus defined our margin as the invers of the norm of
+$\bm{w}$. We want to minimize the norm in order to have a as large as
+possible margin $M$. Before we proceed, we need to remind ourselves
+about Lagrangian multipliers.
+
+!split
+===== A quick Reminder on Lagrangian Multipliers =====
+
+Consider a function of three independent variables $f(x,y,z)$ . For the function $f$ to be an
+extreme we have
+!bt
+\[
+df=0.
+\]
+!et
+A necessary and sufficient condition is
+!bt
+\[
+\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
+\]
+!et
+due to
+!bt
+\[
+df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz.
+\]
+!et
+In many problems the variables $x,y,z$ are often subject to constraints (such as those above for the margin)
+so that they are no longer all independent. It is possible at least in principle to use each 
+constraint to eliminate one variable
+and to proceed with a new and smaller set of independent varables.
+
+The use of so-called Lagrangian  multipliers is an alternative technique  when the elimination
+of variables is incovenient or undesirable.  Assume that we have an equation of constraint on 
+the variables $x,y,z$
+!bt
+\[
+\phi(x,y,z) = 0,
+\]
+!et
+ resulting in
+!bt
+\[
+d\phi = \frac{\partial \phi}{\partial x}dx+\frac{\partial \phi}{\partial y}dy+\frac{\partial \phi}{\partial z}dz =0.
+\]
+!et
+Now we cannot set anymore
+!bt
+\[
+\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
+\]
+!et
+if $df=0$ is wanted
+because there are now only two independent variables!  Assume $x$ and $y$ are the independent 
+variables.
+Then $dz$ is no longer arbitrary.
+
+!split
+===== Adding the Multiplier =====
+
+However, we can add to
+!bt
+\[
+df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz,
+\]
+!et
+a multiplum of $d\phi$, viz. $\lambda d\phi$, resulting  in
+!bt
+\[
+df+\lambda d\phi = (\frac{\partial f}{\partial z}+\lambda
+\frac{\partial \phi}{\partial x})dx+(\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y})dy+
+(\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z})dz =0.
+\]
+!et
+Our multiplier is chosen so that
+!bt
+\[
+\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z} =0.
+\]
+!et
+
+We need to remember that we took $dx$ and $dy$ to be arbitrary and thus we must have
+!bt
+\[
+\frac{\partial f}{\partial x}+\lambda\frac{\partial \phi}{\partial x} =0,
+\]
+!et
+and
+!bt
+\[
+\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y} =0.
+\]
+!et
+When all these equations are satisfied, $df=0$.  We have four unknowns, $x,y,z$ and
+$\lambda$. Actually we want only $x,y,z$, $\lambda$ needs not to be determined, 
+it is therefore often called
+Lagrange's undetermined multiplier.
+If we have a set of constraints $\phi_k$ we have the equations
+!bt
+\[
+\frac{\partial f}{\partial x_i}+\sum_k\lambda_k\frac{\partial \phi_k}{\partial x_i} =0.
+\]
+!et
+
+!split
+===== Setting up the Problem =====
+In order to solve the above problem, we define the following Lagrangian function to be minimized 
+!bt
+\[
+{\cal L}(\lambda,b,\bm{w})=\frac{1}{2}\bm{w}^T\bm{w}-\sum_{i=1}^n\lambda_i\left[y_i(\bm{w}^T\bm{x}_i+b)-1\right],
+\]
+!et
+where $\lambda_i$ is a so-called Lagrange multiplier subject to the condition $\lambda_i \geq 0$.
+
+Taking the derivatives  with respect to $b$ and $\bm{w}$ we obtain 
+!bt
+\[
+\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
+\]
+!et
+and 
+!bt
+\[
+\frac{\partial {\cal L}}{\partial \bm{w}} = 0 = \bm{w}-\sum_{i} \lambda_iy_i\bm{x}_i.
+\]
+!et
+Inserting these constraints into the equation for ${\cal L}$ we obtain
+!bt
+\[
+{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\bm{x}_i^T\bm{x}_j,
+\]
+!et
+subject to the constraints $\lambda_i\geq 0$ and $\sum_i\lambda_iy_i=0$. 
+We must in addition satisfy the "Karush-Kuhn-Tucker":"https://en.wikipedia.org/wiki/Karush%E2%80%93Kuhn%E2%80%93Tucker_conditions" (KKT) condition
+!bt
+\[
+\lambda_i\left[y_i(\bm{w}^T\bm{x}_i+b) -1\right] \hspace{0.1cm}\forall i.
+\]
+!et
+o If $\lambda_i > 0$, then $y_i(\bm{w}^T\bm{x}_i+b)=1$ and we say that $x_i$ is on the boundary.
+o If $y_i(\bm{w}^T\bm{x}_i+b)> 1$, we say $x_i$ is not on the boundary and we set $\lambda_i=0$. 
+When $\lambda_i > 0$, the vectors $\bm{x}_i$ are called support vectors. They are the vectors closest to the line (or hyperplane) and define the margin $M$. 
+
+!split
+===== The problem to solve =====
+
+We can rewrite 
+!bt
+\[
+{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\bm{x}_i^T\bm{x}_j,
+\]
+!et
+and its constraints in terms of a matrix-vector problem where we minimize w.r.t. $\lambda$ the following problem
+!bt
+\[
+\frac{1}{2} \bm{\lambda}^T\begin{bmatrix} y_1y_1\bm{x}_1^T\bm{x}_1 & y_1y_2\bm{x}_1^T\bm{x}_2 & \dots & \dots & y_1y_n\bm{x}_1^T\bm{x}_n \\
+y_2y_1\bm{x}_2^T\bm{x}_1 & y_2y_2\bm{x}_2^T\bm{x}_2 & \dots & \dots & y_1y_n\bm{x}_2^T\bm{x}_n \\
+\dots & \dots & \dots & \dots & \dots \\
+\dots & \dots & \dots & \dots & \dots \\
+y_ny_1\bm{x}_n^T\bm{x}_1 & y_ny_2\bm{x}_n^T\bm{x}_2 & \dots & \dots & y_ny_n\bm{x}_n^T\bm{x}_n \\
+\end{bmatrix}\bm{\lambda}-\mathbb{1}\bm{\lambda}, 
+\]
+!et
+subject to $\bm{y}^T\bm{\lambda}=0$. Here we defined the vectors $\bm{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n]$ and 
+$\bm{y}=[y_1,y_2,\dots,y_n]$. 
+
+
+!split
+===== The last steps =====
+
+Solving the above problem, yields the values of $\lambda_i$.
+To find the coefficients of your hyperplane we need simply to compute 
+!bt
+\[
+\bm{w}=\sum_{i} \lambda_iy_i\bm{x}_i.
+\]
+!et
+With our vector $\bm{w}$ we can in turn find the value of the intercept $b$ (here in two dimensions) via 
+!bt
+\[
+y_i(\bm{w}^T\bm{x}_i+b)=1,
+\]
+!et
+resulting in
+!bt
+\[
+b = \frac{1}{y_i}-\bm{w}^T\bm{x}_i,
+\]
+!et
+or if we write it out in terms of the support vectors only, with $N_s$ being their number,  we have
+!bt
+\[
+b = \frac{1}{N_s}\sum_{j\in N_s}\left(y_j-\sum_{i=1}^n\lambda_iy_i\bm{x}_i^T\bm{x}_j\right).
+\]
+!et
+With our hyperplane coefficients we can use our classifier to assign any observation by simply using 
+!bt
+\[
+y_i = \mathrm{sign}(\bm{w}^T\bm{x}_i+b).
+\]
+!et
+Below we discuss how to find the optimal values of $\lambda_i$. Before we proceed however, we discuss now the so-called soft classifier. 
+
+!split
+===== A soft classifier =====
+
+Till now, the margin is strictly defined by the support vectors. This defines what is called a hard classifier, that is the margins are well defined.
+
+Suppose now that classes overlap in feature space, as shown in the
+figure here. One way to deal with this problem before we define the
+so-called _kernel approach_, is to allow a kind of slack in the sense
+that we allow some points to be on the wrong side of the margin.
+
+We introduce thus the so-called _slack_ variables $\bm{\xi} =[\xi_1,x_2,\dots,x_n]$ and 
+modify our previous equation
+!bt
+\[
+y_i(\bm{w}^T\bm{x}_i+b)=1,
+\]
+!et
+to 
+!bt
+\[
+y_i(\bm{w}^T\bm{x}_i+b)=1-\xi_i,
+\]
+!et
+with the requirement $\xi_i\geq 0$. The total violation is now $\sum_i\xi$. 
+The value $\xi_i$ in the constraint the last constraint corresponds to the  amount by which the prediction
+$y_i(\bm{w}^T\bm{x}_i+b)=1$ is on the wrong side of its margin. Hence by bounding the sum $\sum_i \xi_i$,
+we bound the total amount by which predictions fall on the wrong side of their margins.
+
+Misclassifications occur when $\xi_i > 1$. Thus bounding the total sum by some value $C$ bounds in turn the total number of
+misclassifications.
+
+!split
+===== Soft optmization problem =====
+
+
+This has in turn the consequences that we change our optmization problem to finding the minimum of 
+!bt
+\[
+{\cal L}=\frac{1}{2}\bm{w}^T\bm{w}-\sum_{i=1}^n\lambda_i\left[y_i(\bm{w}^T\bm{x}_i+b)-(1-\xi_)\right]+C\sum_{i=1}^n\xi_i-\sum_{i=1}^n\gamma_i\xi_i,
+\]
+!et
+subject to 
+!bt
+\[
+y_i(\bm{w}^T\bm{x}_i+b)=1-\xi_i \hspace{0.1cm}\forall i,
+\]
+!et
+with the requirement $\xi_i\geq 0$.
+
+Taking the derivatives  with respect to $b$ and $\bm{w}$ we obtain 
+!bt
+\[
+\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
+\]
+!et
+and 
+!bt
+\[
+\frac{\partial {\cal L}}{\partial \bm{w}} = 0 = \bm{w}-\sum_{i} \lambda_iy_i\bm{x}_i, 
+\]
+!et
+and
+!bt
+\[
+\lambda_i = C-\gamma_i \hspace{0.1cm}\forall i.
+\]
+!et
+Inserting these constraints into the equation for ${\cal L}$ we obtain the same equation as before
+!bt
+\[
+{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\bm{x}_i^T\bm{x}_j,
+\]
+!et
+but now subject to the constraints $\lambda_i\geq 0$, $\sum_i\lambda_iy_i=0$ and $0\leq\lambda_i \leq C$. 
+We must in addition satisfy the Karush-Kuhn-Tucker condition which now reads
+!bt
+\[
+\lambda_i\left[y_i(\bm{w}^T\bm{x}_i+b) -(1-\xi_)\right]=0 \hspace{0.1cm}\forall i,
+\]
+!et
+!bt
+\[
+\gamma_i\xi_i = 0,
+\]
+!et
+and 
+!bt
+\[
+y_i(\bm{w}^T\bm{x}_i+b) -(1-\xi_) \geq 0 \hspace{0.1cm}\forall i.
+\]
+!et
+
+!split
+===== Kernels and non-linearity =====
+
+The cases we have studied till now, were all characterized by two classes
+with a close to linear separability. The classifiers we have described
+so far find linear boundaries in our input feature space. It is
+possible to make our procedure more flexible by exploring the feature
+space using other basis expansions such as higher-order polynomials,
+wavelets, splines etc.
+
+If our feature space is not easy to separate, as shown in the figure
+here, we can achieve a better separation by introducing more complex
+basis functions. The ideal would be, as shown in the next figure, to, via a specific transformation to 
+obtain a separation between the classes which is almost linear. 
+
+The change of basis, from $x\rightarrow z=\phi(x)$ leads to the same type of equations to be solved, except that
+we need to introduce for example a polynomial transformation to a two-dimensional training set.
+
+!bc pycod
+import numpy as np
+import os
+
+np.random.seed(42)
+
+# To plot pretty figures
+import matplotlib
+import matplotlib.pyplot as plt
+plt.rcParams['axes.labelsize'] = 14
+plt.rcParams['xtick.labelsize'] = 12
+plt.rcParams['ytick.labelsize'] = 12
+
+
+from sklearn.svm import SVC
+from sklearn import datasets
+
+
+
+X1D = np.linspace(-4, 4, 9).reshape(-1, 1)
+X2D = np.c_[X1D, X1D**2]
+y = np.array([0, 0, 1, 1, 1, 1, 1, 0, 0])
+
+plt.figure(figsize=(11, 4))
+
+plt.subplot(121)
+plt.grid(True, which='both')
+plt.axhline(y=0, color='k')
+plt.plot(X1D[:, 0][y==0], np.zeros(4), "bs")
+plt.plot(X1D[:, 0][y==1], np.zeros(5), "g^")
+plt.gca().get_yaxis().set_ticks([])
+plt.xlabel(r"$x_1$", fontsize=20)
+plt.axis([-4.5, 4.5, -0.2, 0.2])
+
+plt.subplot(122)
+plt.grid(True, which='both')
+plt.axhline(y=0, color='k')
+plt.axvline(x=0, color='k')
+plt.plot(X2D[:, 0][y==0], X2D[:, 1][y==0], "bs")
+plt.plot(X2D[:, 0][y==1], X2D[:, 1][y==1], "g^")
+plt.xlabel(r"$x_1$", fontsize=20)
+plt.ylabel(r"$x_2$", fontsize=20, rotation=0)
+plt.gca().get_yaxis().set_ticks([0, 4, 8, 12, 16])
+plt.plot([-4.5, 4.5], [6.5, 6.5], "r--", linewidth=3)
+plt.axis([-4.5, 4.5, -1, 17])
+plt.subplots_adjust(right=1)
+plt.show()
+
+!ec
+
+
+
+!split
+===== The equations =====
+
+Suppose we define a polynomial transformation of degree two only (we continue to live in a plane with $x_i$ and $y_i$ as variables)
+!bt
+\[
+z = \phi(x_i) =\left(x_i^2, y_i^2, \sqrt{2}x_iy_i\right).
+\]
+!et
+
+With our new basis, the equations we solved earlier are basically the same, that is we have now (without the slack option for simplicity)
+!bt
+\[
+{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\bm{z}_i^T\bm{z}_j,
+\]
+!et
+subject to the constraints $\lambda_i\geq 0$, $\sum_i\lambda_iy_i=0$, and for the support vectors
+!bt
+\[
+y_i(\bm{w}^T\bm{z}_i+b)= 1 \hspace{0.1cm}\forall i,
+\]
+!et
+from which we also find $b$.
+To compute $\bm{z}_i^T\bm{z}_j$ we define the kernel $K(\bm{x}_i,\bm{x}_j)$ as
+!bt
+\[
+K(\bm{x}_i,\bm{x}_j)=\bm{z}_i^T\bm{z}_j= \phi(\bm{x}_i)^T\phi(\bm{x}_j).
+\]
+!et
+For the above example, the kernel reads
+!bt
+\[
+K(\bm{x}_i,\bm{x}_j)=[x_i^2, y_i^2, \sqrt{2}x_iy_i]^T\begin{bmatrix} x_j^2 \\ y_j^2 \\ \sqrt{2}x_jy_j \end{bmatrix}=x_i^2x_j^2+2x_ix_jy_iy_j+y_i^2y_j^2.
+\]
+!et
+
+We note that this is nothing but the dot product of the two original
+vectors $(\bm{x}_i^T\bm{x}_j)^2$. Instead of thus computing the
+product in the Lagrangian of $\bm{z}_i^T\bm{z}_j$ we simply compute
+the dot product $(\bm{x}_i^T\bm{x}_j)^2$.
+
+
+This leads to the so-called
+kernel trick and the result leads to the same as if we went through
+the trouble of performing the transformation
+$\phi(\bm{x}_i)^T\phi(\bm{x}_j)$ during the SVM calculations.
+
+
+!split
+===== The problem to solve =====
+Using our definition of the kernel We can rewrite again the Lagrangian
+!bt
+\[
+{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\bm{x}_i^T\bm{z}_j,
+\]
+!et
+subject to the constraints $\lambda_i\geq 0$, $\sum_i\lambda_iy_i=0$ in terms of a convex optimization problem
+!bt
+\[
+\frac{1}{2} \bm{\lambda}^T\begin{bmatrix} y_1y_1K(\bm{x}_1,\bm{x}_1) & y_1y_2K(\bm{x}_1,\bm{x}_2) & \dots & \dots & y_1y_nK(\bm{x}_1,\bm{x}_n) \\
+y_2y_1K(\bm{x}_2,\bm{x}_1) & y_2y_2(\bm{x}_2,\bm{x}_2) & \dots & \dots & y_1y_nK(\bm{x}_2,\bm{x}_n) \\
+\dots & \dots & \dots & \dots & \dots \\
+\dots & \dots & \dots & \dots & \dots \\
+y_ny_1K(\bm{x}_n,\bm{x}_1) & y_ny_2K(\bm{x}_n\bm{x}_2) & \dots & \dots & y_ny_nK(\bm{x}_n,\bm{x}_n) \\
+\end{bmatrix}\bm{\lambda}-\mathbb{1}\bm{\lambda}, 
+\]
+!et
+subject to $\bm{y}^T\bm{\lambda}=0$. Here we defined the vectors $\bm{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n]$ and 
+$\bm{y}=[y_1,y_2,\dots,y_n]$. 
+If we add the slack constants this leads to the additional constraint $0\leq \lambda_i \leq C$.
+
+We can rewrite this (see the solutions below) in terms of a convex optimization problem of the type
+!bt
+\begin{align*}
+    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\bm{\lambda}^T\bm{P}\bm{\lambda}+\bm{q}^T\bm{\lambda},\\ \nonumber
+    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \bm{G}\bm{\lambda} \preceq \bm{h} \hspace{0.2cm} \wedge \bm{A}\bm{\lambda}=f.
+\end{align*}
+!et
+Below we discuss how to solve these equations. Here we note that the matrix $\bm{P}$ has matrix elements $p_{ij}=y_iy_jK(\bm{x}_i,\bm{x}_j)$.
+Given a kernel $K$ and the targets $y_i$ this matrix is easy to set up. The constraint $\bm{y}^T\bm{\lambda}=0$ leads to $f=0$ and $\bm{A}=\bm{y}$. How to set up the matrix $\bm{G}$ is discussed later. Here note that the inequalities $0\leq \lambda_i \leq C$ can be split up into
+$0\leq \lambda_i$ and $\lambda_i \leq C$. These two inequalities define then the matrix $\bm{G}$ and the vector $\bm{h}$.
+
+
+!split
+===== Different kernels and Mercer's theorem =====
+
+There are several popular kernels being used. These are
+o Linear: $K(\bm{x},\bm{y})=\bm{x}^T\bm{y}$,
+o Polynomial: $K(\bm{x},\bm{y})=(\bm{x}^T\bm{y}+\gamma)^d$,
+o Gaussian Radial Basis Function: $K(\bm{x},\bm{y})=\exp{\left(-\gamma\vert\vert\bm{x}-\bm{y}\vert\vert^2\right)}$,
+o Tanh: $K(\bm{x},\bm{y})=\tanh{(\bm{x}^T\bm{y}+\gamma)}$,
+and many other ones.
+
+An important theorem for us is "Mercer's
+theorem":"https://en.wikipedia.org/wiki/Mercer%27s_theorem".  The
+theorem states that if a kernel function $K$ is symmetric, continuous
+and leads to a positive semi-definite matrix $\bm{P}$ then there
+exists a function $\phi$ that maps $\bm{x}_i$ and $\bm{x}_j$ into
+another space (possibly with much higher dimensions) such that
+
+!bt
+\[
+K(\bm{x}_i,\bm{x}_j)=\phi(\bm{x}_i)^T\phi(\bm{x}_j).
+\]
+!et
+
+So you can use $K$ as a kernel since you know $\phi$ exists, even if
+you don’t know what $\phi$ is. 
+
+Note that some frequently used kernels (such as the Sigmoid kernel)
+don’t respect all of Mercer’s conditions, yet they generally work well
+in practice.
+
+
+!split
+===== The moons example =====
+!bc pycod
+from __future__ import division, print_function, unicode_literals
+
+import numpy as np
+np.random.seed(42)
+
+import matplotlib
+import matplotlib.pyplot as plt
+plt.rcParams['axes.labelsize'] = 14
+plt.rcParams['xtick.labelsize'] = 12
+plt.rcParams['ytick.labelsize'] = 12
+
+
+from sklearn.svm import SVC
+from sklearn import datasets
+
+
+
+from sklearn.pipeline import Pipeline
+from sklearn.preprocessing import StandardScaler
+from sklearn.svm import LinearSVC
+
+
+from sklearn.datasets import make_moons
+X, y = make_moons(n_samples=100, noise=0.15, random_state=42)
+
+def plot_dataset(X, y, axes):
+    plt.plot(X[:, 0][y==0], X[:, 1][y==0], "bs")
+    plt.plot(X[:, 0][y==1], X[:, 1][y==1], "g^")
+    plt.axis(axes)
+    plt.grid(True, which='both')
+    plt.xlabel(r"$x_1$", fontsize=20)
+    plt.ylabel(r"$x_2$", fontsize=20, rotation=0)
+
+plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])
+plt.show()
+
+from sklearn.datasets import make_moons
+from sklearn.pipeline import Pipeline
+from sklearn.preprocessing import PolynomialFeatures
+
+polynomial_svm_clf = Pipeline([
+        ("poly_features", PolynomialFeatures(degree=3)),
+        ("scaler", StandardScaler()),
+        ("svm_clf", LinearSVC(C=10, loss="hinge", random_state=42))
+    ])
+
+polynomial_svm_clf.fit(X, y)
+
+def plot_predictions(clf, axes):
+    x0s = np.linspace(axes[0], axes[1], 100)
+    x1s = np.linspace(axes[2], axes[3], 100)
+    x0, x1 = np.meshgrid(x0s, x1s)
+    X = np.c_[x0.ravel(), x1.ravel()]
+    y_pred = clf.predict(X).reshape(x0.shape)
+    y_decision = clf.decision_function(X).reshape(x0.shape)
+    plt.contourf(x0, x1, y_pred, cmap=plt.cm.brg, alpha=0.2)
+    plt.contourf(x0, x1, y_decision, cmap=plt.cm.brg, alpha=0.1)
+
+plot_predictions(polynomial_svm_clf, [-1.5, 2.5, -1, 1.5])
+plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])
+
+plt.show()
+
+
+from sklearn.svm import SVC
+
+poly_kernel_svm_clf = Pipeline([
+        ("scaler", StandardScaler()),
+        ("svm_clf", SVC(kernel="poly", degree=3, coef0=1, C=5))
+    ])
+poly_kernel_svm_clf.fit(X, y)
+
+poly100_kernel_svm_clf = Pipeline([
+        ("scaler", StandardScaler()),
+        ("svm_clf", SVC(kernel="poly", degree=10, coef0=100, C=5))
+    ])
+poly100_kernel_svm_clf.fit(X, y)
+
+plt.figure(figsize=(11, 4))
+
+plt.subplot(121)
+plot_predictions(poly_kernel_svm_clf, [-1.5, 2.5, -1, 1.5])
+plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])
+plt.title(r"$d=3, r=1, C=5$", fontsize=18)
+
+plt.subplot(122)
+plot_predictions(poly100_kernel_svm_clf, [-1.5, 2.5, -1, 1.5])
+plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])
+plt.title(r"$d=10, r=100, C=5$", fontsize=18)
+
+plt.show()
+
+def gaussian_rbf(x, landmark, gamma):
+    return np.exp(-gamma * np.linalg.norm(x - landmark, axis=1)**2)
+
+gamma = 0.3
+
+x1s = np.linspace(-4.5, 4.5, 200).reshape(-1, 1)
+x2s = gaussian_rbf(x1s, -2, gamma)
+x3s = gaussian_rbf(x1s, 1, gamma)
+
+XK = np.c_[gaussian_rbf(X1D, -2, gamma), gaussian_rbf(X1D, 1, gamma)]
+yk = np.array([0, 0, 1, 1, 1, 1, 1, 0, 0])
+
+plt.figure(figsize=(11, 4))
+
+plt.subplot(121)
+plt.grid(True, which='both')
+plt.axhline(y=0, color='k')
+plt.scatter(x=[-2, 1], y=[0, 0], s=150, alpha=0.5, c="red")
+plt.plot(X1D[:, 0][yk==0], np.zeros(4), "bs")
+plt.plot(X1D[:, 0][yk==1], np.zeros(5), "g^")
+plt.plot(x1s, x2s, "g--")
+plt.plot(x1s, x3s, "b:")
+plt.gca().get_yaxis().set_ticks([0, 0.25, 0.5, 0.75, 1])
+plt.xlabel(r"$x_1$", fontsize=20)
+plt.ylabel(r"Similarity", fontsize=14)
+plt.annotate(r'$\mathbf{x}$',
+             xy=(X1D[3, 0], 0),
+             xytext=(-0.5, 0.20),
+             ha="center",
+             arrowprops=dict(facecolor='black', shrink=0.1),
+             fontsize=18,
+            )
+plt.text(-2, 0.9, "$x_2$", ha="center", fontsize=20)
+plt.text(1, 0.9, "$x_3$", ha="center", fontsize=20)
+plt.axis([-4.5, 4.5, -0.1, 1.1])
+
+plt.subplot(122)
+plt.grid(True, which='both')
+plt.axhline(y=0, color='k')
+plt.axvline(x=0, color='k')
+plt.plot(XK[:, 0][yk==0], XK[:, 1][yk==0], "bs")
+plt.plot(XK[:, 0][yk==1], XK[:, 1][yk==1], "g^")
+plt.xlabel(r"$x_2$", fontsize=20)
+plt.ylabel(r"$x_3$  ", fontsize=20, rotation=0)
+plt.annotate(r'$\phi\left(\mathbf{x}\right)$',
+             xy=(XK[3, 0], XK[3, 1]),
+             xytext=(0.65, 0.50),
+             ha="center",
+             arrowprops=dict(facecolor='black', shrink=0.1),
+             fontsize=18,
+            )
+plt.plot([-0.1, 1.1], [0.57, -0.1], "r--", linewidth=3)
+plt.axis([-0.1, 1.1, -0.1, 1.1])
+    
+plt.subplots_adjust(right=1)
+
+plt.show()
+
+
+x1_example = X1D[3, 0]
+for landmark in (-2, 1):
+    k = gaussian_rbf(np.array([[x1_example]]), np.array([[landmark]]), gamma)
+    print("Phi({}, {}) = {}".format(x1_example, landmark, k))
+
+rbf_kernel_svm_clf = Pipeline([
+        ("scaler", StandardScaler()),
+        ("svm_clf", SVC(kernel="rbf", gamma=5, C=0.001))
+    ])
+rbf_kernel_svm_clf.fit(X, y)
+
+
+from sklearn.svm import SVC
+
+gamma1, gamma2 = 0.1, 5
+C1, C2 = 0.001, 1000
+hyperparams = (gamma1, C1), (gamma1, C2), (gamma2, C1), (gamma2, C2)
+
+svm_clfs = []
+for gamma, C in hyperparams:
+    rbf_kernel_svm_clf = Pipeline([
+            ("scaler", StandardScaler()),
+            ("svm_clf", SVC(kernel="rbf", gamma=gamma, C=C))
+        ])
+    rbf_kernel_svm_clf.fit(X, y)
+    svm_clfs.append(rbf_kernel_svm_clf)
+
+plt.figure(figsize=(11, 7))
+
+for i, svm_clf in enumerate(svm_clfs):
+    plt.subplot(221 + i)
+    plot_predictions(svm_clf, [-1.5, 2.5, -1, 1.5])
+    plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])
+    gamma, C = hyperparams[i]
+    plt.title(r"$\gamma = {}, C = {}$".format(gamma, C), fontsize=16)
+
+plt.show()
+
+!ec
+
+
+
+!split
+=====  Mathematical optimization of convex functions =====
+
+A mathematical (quadratic) optimization problem, or just optimization problem, has the form
+!bt
+\begin{align*}
+    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\bm{\lambda}^T\bm{P}\bm{\lambda}+\bm{q}^T\bm{\lambda},\\ \nonumber
+    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \bm{G}\bm{\lambda} \preceq \bm{h} \wedge  \bm{A}\bm{\lambda}=f.
+\end{align*}
+!et
+subject to some constraints for say a selected set $i=1,2,\dots, n$.
+In our case we are optimizing with respect to the Lagrangian multipliers $\lambda_i$, and the
+vector $\bm{\lambda}=[\lambda_1, \lambda_2,\dots, \lambda_n]$ is the optimization variable we are dealing with.
+
+In our case we are particularly interested in a class of optimization problems called convex optmization problems. 
+In our discussion on gradient descent methods we discussed at length the definition of a convex function. 
+
+Convex optimization problems play a central role in applied mathematics and we recommend strongly "Boyd and Vandenberghe's text on the topics":"http://web.stanford.edu/~boyd/cvxbook/".
+
+
+
+!split
+===== How do we solve these problems? =====
+
+If we use Python as programming language and wish to venture beyond
+_scikit-learn_, _tensorflow_ and similar software which makes our
+lives so much easier, we need to dive into the wonderful world of
+quadratic programming. We can, if we wish, solve the minimization
+problem using say standard gradient methods or conjugate gradient
+methods. However, these methods tend to exhibit a rather slow
+converge. So, welcome to the promised land of quadratic programming.
+
+The functions we need are contained in the "quadratic programming library CVXOPT":"https://cvxopt.org/userguide/coneprog.html" and we need to import it together with _numpy_ as
+
+!bc pycod
+import numpy
+import cvxopt
+!ec
+
+This will make our life much easier. You don't need t0 write your own optimizer.
+
+
+!split
+===== A simple example =====
+
+We remind ourselves about the general problem we want to solve
+!bt
+\begin{align*}
+    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}\bm{x}^T\bm{P}\bm{x}+\bm{q}^T\bm{x},\\ \nonumber
+    &\mathrm{subject\hspace{0.1cm} to} \hspace{0.2cm} \bm{G}\bm{x} \preceq \bm{h} \wedge  \bm{A}\bm{x}=f.
+\end{align*}
+!et
+
+Let us show how to perform the optmization using a simple case. Assume we want to optimize the following problem
+!bt
+\begin{align*}
+    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}x^2+5x+3y \\ \nonumber
+    &\mathrm{subject}\hspace{0.5cm} \mathrm{to} \\ \nonumber
+    &x, y \geq 0 \\ \nonumber
+    &x+3y  \geq 15 \\ \nonumber
+    &2x+5y  \leq  100 \\ \nonumber
+    &3x+4y  \leq  80.  \\ \nonumber
+\end{align*}
+!et
+The minimization problem can be rewritten in terms of vectors and matrices as (with $x$ and $y$ being the unknowns)
+!bt
+\[
+\frac{1}{2}\begin{bmatrix} x\\ y \end{bmatrix}^T   \begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}  \begin{bmatrix} x \\ y \end{bmatrix}  + \begin{bmatrix}3\\ 4  \end{bmatrix}^T \begin{bmatrix}x \\ y  \end{bmatrix}.  
+\]
+!et
+Similarly, we can now set up the inequalities (we need to change $\geq$ to $\leq$ by multiplying with $-1$ on bot sides) as the following matrix-vector equation
+!bt
+\[
+\begin{bmatrix} -1 & 0 \\ 0 & -1 \\ -1 & -3 \\ 2 & 5 \\ 3 & 4\end{bmatrix}\begin{bmatrix} x \\ y\end{bmatrix} \preceq \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
+\]
+!et
+We have collapsed all the inequalities into a single matrix $\bm{G}$. We see also that our matrix 
+!bt
+\[
+\bm{P} =\begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}
+\]
+!et
+is clearly positive semi-definite (all eigenvalues larger or equal zero). 
+Finally, the vector $\bm{h}$ is defined as 
+!bt
+\[
+\bm{h} = \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
+\]
+!et
+
+
+Since we don't have any equalities the matrix $\bm{A}$ is set to zero
+The following code solves the equations for us
+!bc pycod
+# Import the necessary packages
+import numpy
+from cvxopt import matrix
+from cvxopt import solvers
+P = matrix(numpy.diag([1,0]), tc='d')
+q = matrix(numpy.array([3,4]), tc='d')
+G = matrix(numpy.array([[-1,0],[0,-1],[-1,-3],[2,5],[3,4]]), tc='d')
+h = matrix(numpy.array([0,0,-15,100,80]), tc='d')
+# Construct the QP, invoke solver
+sol = solvers.qp(P,q,G,h)
+# Extract optimal value and solution
+sol['x'] 
+sol['primal objective']
+print(sol['x'] )
+!ec
+
+!split
+===== Back to the more realistic cases =====
+
+We are now ready to return to our setup of the optmization problem for a more realistic case. Introducing the _slack_ parameter $C$ we have
+!bt
+\[
+\frac{1}{2} \bm{\lambda}^T\begin{bmatrix} y_1y_1K(\bm{x}_1,\bm{x}_1) & y_1y_2K(\bm{x}_1,\bm{x}_2) & \dots & \dots & y_1y_nK(\bm{x}_1,\bm{x}_n) \\
+y_2y_1K(\bm{x}_2,\bm{x}_1) & y_2y_2K(\bm{x}_2,\bm{x}_2) & \dots & \dots & y_1y_nK(\bm{x}_2,\bm{x}_n) \\
+\dots & \dots & \dots & \dots & \dots \\
+\dots & \dots & \dots & \dots & \dots \\
+y_ny_1K(\bm{x}_n,\bm{x}_1) & y_ny_2K(\bm{x}_n\bm{x}_2) & \dots & \dots & y_ny_nK(\bm{x}_n,\bm{x}_n) \\
+\end{bmatrix}\bm{\lambda}-\mathbb{I}\bm{\lambda}, 
+\]
+!et
+subject to $\bm{y}^T\bm{\lambda}=0$. Here we defined the vectors $\bm{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n]$ and 
+$\bm{y}=[y_1,y_2,\dots,y_n]$. 
+With  the slack constants this leads to the additional constraint $0\leq \lambda_i \leq C$.
+
+Using the _CVXOPT_ library, the matrix $P$ would then be defined by the above matrix while the KKT conditions would all be collected by the matrix $G$.
+
+
+
+!split
+===== Support vector machines for regression =====
+
+Material will be added here.
diff --git a/doc/src/week46/week46.do.txt b/doc/src/week46/week46.do.txt
index 1ea619db0..f3f7b5662 100644
--- a/doc/src/week46/week46.do.txt
+++ b/doc/src/week46/week46.do.txt
@@ -1,1171 +1,1916 @@
-TITLE: Week 46: Support Vector Machines and Project 3. 
+TITLE: Week 46: Decision Trees, Ensemble methods  and Random Forests
 AUTHOR: Morten Hjorth-Jensen {copyright, 1999-present|CC BY-NC} at Department of Physics, University of Oslo & Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University
-DATE: today
+DATE: Week 46, November 13-17
+
 
 !split
-===== Overview of week 46 =====
+===== Plan for week 46 =====
 
-* _Thursday_: Support Vector Machines, discussion of project 3
-  * "Video of lecture":"https://youtu.be/F3CkH-opbdY"
-* _Friday_: Support vector machines and start Principal Component Analysis
-  * "Video of lecture":"https://youtu.be/BbupEEvMXtg"
+!bblock  Active learning sessions on Tuesday and Wednesday
+  * Work and Discussion of project 2
+  * Discussion of project 3 as well
+!eblock  
 
-!bblock Reading
-Reading recommendations:
-  o See lecture notes for week 46 at https://compphysics.github.io/MachineLearning/doc/web/course.html.
-  o Hastie et al chapter 12
-  o Bishop chapter 7.1 and 7.2
+!bblock Material for the lecture on Thursday November 16, 2023
+  * Thursday: Basics of decision trees, classification and regression algorithms and ensemble models 
+  * Readings and Videos:
+    * These lecture notes
+    * "Video of lecture to be added":"https://youtu.be/"
+    * "Video on Decision trees":"https://www.youtube.com/watch?v=RmajweUFKvM&ab_channel=Simplilearn"
+    * Decision Trees: Geron's chapter 6 covers decision trees while ensemble models, voting and bagging are discussed in chapter 7. See also lecture from "STK-IN4300, lecture 7":"https://www.uio.no/studier/emner/matnat/math/STK-IN4300/h20/slides/lecture_7.pdf". Chapter 9.2 of Hastie et al contains also a good discussion.
 !eblock
 
 
-!bblock Videos
-o "Overview video on Support Vector Machines":"https://www.youtube.com/watch?v=efR1C6CvhmE&ab_channel=StatQuestwithJoshStarmer"
-o See also "this video":"https://www.youtube.com/watch?v=N1vOgolbjSc&ab_channel=AliceZhao".
-!eblock
+
+!split
+===== Decision trees, overarching aims  =====
+
+
+We start here with the most basic algorithm, the so-called decision
+tree. With this basic algorithm we can in turn build more complex
+networks, spanning from homogeneous and heterogenous forests (bagging,
+random forests and more) to one of the most popular supervised
+algorithms nowadays, the extreme gradient boosting, or just
+XGBoost. But let us start with the simplest possible ingredient.
+
+Decision trees are supervised learning algorithms used for both,
+classification and regression tasks.
+
+
+The main idea of decision trees
+is to find those descriptive features which contain the most
+_information_ regarding the target feature and then split the dataset
+along the values of these features such that the target feature values
+for the resulting underlying datasets are as pure as possible.
+
+The descriptive features which reproduce best the target/output features are normally  said
+to be the most informative ones. The process of finding the _most
+informative_ feature is done until we accomplish a stopping criteria
+where we then finally end up in so called _leaf nodes_. 
+
+!split
+===== Basics of a tree =====
+
+A decision tree is typically divided into a _root node_, the _interior nodes_,
+and the final _leaf nodes_ or just _leaves_. These entities are then connected by so-called _branches_.
+
+The leaf nodes
+contain the predictions we will make for new query instances presented
+to our trained model. This is possible since the model has 
+learned the underlying structure of the training data and hence can,
+given some assumptions, make predictions about the target feature value
+(class) of unseen query instances.
+
+!split
+===== A Sketch of a Tree, Regression problem  =====
+
+"See handwritten notes November 3":"https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf"
+
+#FIGURE: [DataFiles/Regsimpletree.png, width=600 frac=0.8]
+
+!split
+===== A Sketch of a Tree, Classification  problem  =====
+
+"See handwritten notes November 3":"https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf"
+#FIGURE: [DataFiles/Classimpletree.png, width=600 frac=0.8]
+
 
 
 !split
-===== Support Vector Machines, overarching aims  =====
+===== A typical Decision Tree with its pertinent Jargon, Classification Problem =====
 
-A Support Vector Machine (SVM) is a very powerful and versatile
-Machine Learning method, capable of performing linear or nonlinear
-classification, regression, and even outlier detection. It is one of
-the most popular models in Machine Learning, and anyone interested in
-Machine Learning should have it in their toolbox. SVMs are
-particularly well suited for classification of complex but small-sized or
-medium-sized datasets.  
+FIGURE: [DataFiles/cancer.png, width=600 frac=0.8]
 
-The case with two well-separated classes only can be understood in an
-intuitive way in terms of lines in a two-dimensional space separating
-the two classes (see figure below).
+This tree was produced using the Wisconsin cancer data (discussed here as well, see code examples below) using _Scikit-Learn_'s decision tree classifier. Here we have used the so-called _gini_ index (see below) to split the various branches.
 
-The basic mathematics behind the SVM is however less familiar to most of us. 
-It relies on the definition of hyperplanes and the
-definition of a _margin_ which separates classes (in case of
-classification problems) of variables. It is also used for regression
-problems.
 
-With SVMs we distinguish between hard margin and soft margins. The
-latter introduces a so-called softening parameter to be discussed
-below.  We distinguish also between linear and non-linear
-approaches. The latter are the most frequent ones since it is rather
-unlikely that we can separate classes easily by say straight lines.
 
 !split
-===== Hyperplanes and all that =====
+===== General Features =====
 
-The theory behind support vector machines (SVM hereafter) is based on
-the mathematical description of so-called hyperplanes. Let us start
-with a two-dimensional case. This will also allow us to introduce our
-first SVM examples. These will be tailored to the case of two specific
-classes, as displayed in the figure here based on the usage of the petal data.
+The overarching approach to decision trees is a top-down approach.
 
-We assume here that our data set can be well separated into two
-domains, where a straight line does the job in the separating the two
-classes. Here the two classes are represented by either squares or
-circles.
+* A leaf provides the classification of a given instance.
+* A node specifies a test of some attribute of the instance.
+* A branch corresponds to a possible values of an attribute.
+* An instance is classified by starting at the root node of the tree, testing the attribute specified by this node, then moving down the tree branch corresponding to the value of the attribute in the given example.
+
+
+This process is then repeated for the subtree rooted at the new
+node.
+
+
+!split
+===== How do we set it up? =====
+
+
+In simplified terms, the process of training a decision tree and
+predicting the target features of query instances is as follows:
+
+o Present a dataset containing of a number of training instances characterized by a number of descriptive features and a target feature
+
+o Train the decision tree model by continuously splitting the target feature along the values of the descriptive features using a measure of information gain during the training process
+
+o Grow the tree until we accomplish a stopping criteria create leaf nodes which represent the *predictions* we want to make for new query instances
+
+o Show query instances to the tree and run down the tree until we arrive at leaf nodes
+
+Then we are essentially done!
+
+
+
+
+
+!split
+===== Decision trees and Regression  =====
 !bc pycod
-from sklearn import datasets
-from sklearn.svm import SVC, LinearSVC
-from sklearn.linear_model import SGDClassifier
-from sklearn.preprocessing import StandardScaler
-import matplotlib
+import numpy as np
 import matplotlib.pyplot as plt
-plt.rcParams['axes.labelsize'] = 14
-plt.rcParams['xtick.labelsize'] = 12
-plt.rcParams['ytick.labelsize'] = 12
+from sklearn.preprocessing import PolynomialFeatures
+from sklearn.linear_model import LinearRegression
 
+steps=250
 
-iris = datasets.load_iris()
-X = iris["data"][:, (2, 3)]  # petal length, petal width
-y = iris["target"]
+distance=0
+x=0
+distance_list=[]
+steps_list=[]
+while x<steps:
+    distance+=np.random.randint(-1,2)
+    distance_list.append(distance)
+    x+=1
+    steps_list.append(x)
+plt.plot(steps_list,distance_list, color='green', label="Random Walk Data")
 
-setosa_or_versicolor = (y == 0) | (y == 1)
-X = X[setosa_or_versicolor]
-y = y[setosa_or_versicolor]
+steps_list=np.asarray(steps_list)
+distance_list=np.asarray(distance_list)
 
+X=steps_list[:,np.newaxis]
 
+#Polynomial fits
 
-C = 5
-alpha = 1 / (C * len(X))
+#Degree 2
+poly_features=PolynomialFeatures(degree=2, include_bias=False)
+X_poly=poly_features.fit_transform(X)
 
-lin_clf = LinearSVC(loss="hinge", C=C, random_state=42)
-svm_clf = SVC(kernel="linear", C=C)
-sgd_clf = SGDClassifier(loss="hinge", learning_rate="constant", eta0=0.001, alpha=alpha,
-                        max_iter=100000, random_state=42)
+lin_reg=LinearRegression()
+poly_fit=lin_reg.fit(X_poly,distance_list)
+b=lin_reg.coef_
+c=lin_reg.intercept_
+print ("2nd degree coefficients:")
+print ("zero power: ",c)
+print ("first power: ", b[0])
+print ("second power: ",b[1])
 
-scaler = StandardScaler()
-X_scaled = scaler.fit_transform(X)
+z = np.arange(0, steps, .01)
+z_mod=b[1]*z**2+b[0]*z+c
 
-lin_clf.fit(X_scaled, y)
-svm_clf.fit(X_scaled, y)
-sgd_clf.fit(X_scaled, y)
+fit_mod=b[1]*X**2+b[0]*X+c
+plt.plot(z, z_mod, color='r', label="2nd Degree Fit")
+plt.title("Polynomial Regression")
 
-print("LinearSVC:                   ", lin_clf.intercept_, lin_clf.coef_)
-print("SVC:                         ", svm_clf.intercept_, svm_clf.coef_)
-print("SGDClassifier(alpha={:.5f}):".format(sgd_clf.alpha), sgd_clf.intercept_, sgd_clf.coef_)
+plt.xlabel("Steps")
+plt.ylabel("Distance")
 
-# Compute the slope and bias of each decision boundary
-w1 = -lin_clf.coef_[0, 0]/lin_clf.coef_[0, 1]
-b1 = -lin_clf.intercept_[0]/lin_clf.coef_[0, 1]
-w2 = -svm_clf.coef_[0, 0]/svm_clf.coef_[0, 1]
-b2 = -svm_clf.intercept_[0]/svm_clf.coef_[0, 1]
-w3 = -sgd_clf.coef_[0, 0]/sgd_clf.coef_[0, 1]
-b3 = -sgd_clf.intercept_[0]/sgd_clf.coef_[0, 1]
+#Degree 10
+poly_features10=PolynomialFeatures(degree=10, include_bias=False)
+X_poly10=poly_features10.fit_transform(X)
 
-# Transform the decision boundary lines back to the original scale
-line1 = scaler.inverse_transform([[-10, -10 * w1 + b1], [10, 10 * w1 + b1]])
-line2 = scaler.inverse_transform([[-10, -10 * w2 + b2], [10, 10 * w2 + b2]])
-line3 = scaler.inverse_transform([[-10, -10 * w3 + b3], [10, 10 * w3 + b3]])
+poly_fit10=lin_reg.fit(X_poly10,distance_list)
 
-# Plot all three decision boundaries
-plt.figure(figsize=(11, 4))
-plt.plot(line1[:, 0], line1[:, 1], "k:", label="LinearSVC")
-plt.plot(line2[:, 0], line2[:, 1], "b--", linewidth=2, label="SVC")
-plt.plot(line3[:, 0], line3[:, 1], "r-", label="SGDClassifier")
-plt.plot(X[:, 0][y==1], X[:, 1][y==1], "bs") # label="Iris-Versicolor"
-plt.plot(X[:, 0][y==0], X[:, 1][y==0], "yo") # label="Iris-Setosa"
-plt.xlabel("Petal length", fontsize=14)
-plt.ylabel("Petal width", fontsize=14)
-plt.legend(loc="upper center", fontsize=14)
-plt.axis([0, 5.5, 0, 2])
+y_plot=poly_fit10.predict(X_poly10)
+plt.plot(X, y_plot, color='black', label="10th Degree Fit")
 
+plt.legend()
 plt.show()
 
 
+#Decision Tree Regression
+from sklearn.tree import DecisionTreeRegressor
+regr_1=DecisionTreeRegressor(max_depth=2)
+regr_2=DecisionTreeRegressor(max_depth=5)
+regr_3=DecisionTreeRegressor(max_depth=7)
+regr_1.fit(X, distance_list)
+regr_2.fit(X, distance_list)
+regr_3.fit(X, distance_list)
+
+X_test = np.arange(0.0, steps, 0.01)[:, np.newaxis]
+y_1 = regr_1.predict(X_test)
+y_2 = regr_2.predict(X_test)
+y_3=regr_3.predict(X_test)
+
+# Plot the results
+plt.figure()
+plt.scatter(X, distance_list, s=2.5, c="black", label="data")
+plt.plot(X_test, y_1, color="red",
+         label="max_depth=2", linewidth=2)
+plt.plot(X_test, y_2, color="green", label="max_depth=5", linewidth=2)
+plt.plot(X_test, y_3, color="m", label="max_depth=7", linewidth=2)
+
+plt.xlabel("Data")
+plt.ylabel("Darget")
+plt.title("Decision Tree Regression")
+plt.legend()
+plt.show()
 
 !ec
 
 
 
+!split
+===== Building a tree, regression =====
+
+There are mainly two steps
+o We split the predictor space (the set of possible values $x_1,x_2,\dots, x_p$) into $J$ distinct and non-non-overlapping regions, $R_1,R_2,\dots,R_J$.  
+o For every observation that falls into the region $R_j$ , we make the same prediction, which is simply the mean of the response values for the training observations in $R_j$.
+
+How do we construct the regions $R_1,\dots,R_J$?  In theory, the
+regions could have any shape. However, we choose to divide the
+predictor space into high-dimensional rectangles, or boxes, for
+simplicity and for ease of interpretation of the resulting predictive
+model. The goal is to find boxes $R_1,\dots,R_J$ that minimize the
+MSE, given by
+
+!bt
+\[
+\sum_{j=1}^J\sum_{i\in R_j}(y_i-\overline{y}_{R_j})^2,
+\]
+!et
+
+where $\overline{y}_{R_j}$  is the mean response for the training observations 
+within box $j$. 
 
 !split
-===== What is a hyperplane? =====
+===== A top-down approach, recursive binary splitting =====
 
-The aim of the SVM algorithm is to find a hyperplane in a
-$p$-dimensional space, where $p$ is the number of features that
-distinctly classifies the data points.
+Unfortunately, it is computationally infeasible to consider every
+possible partition of the feature space into $J$ boxes.  The common
+strategy is to take a top-down approach
 
-In a $p$-dimensional space, a hyperplane is what we call an affine subspace of dimension of $p-1$.
-As an example, in two dimension, a hyperplane is simply as straight line while in three dimensions it is 
-a two-dimensional subspace, or stated simply, a plane. 
-
-In two dimensions, with the variables $x_1$ and $x_2$, the hyperplane is defined as
-!bt
-\[
-b+w_1x_1+w_2x_2=0,
-\]
-!et
-
-where $b$ is the intercept and $w_1$ and $w_2$ define the elements of a vector orthogonal to the line 
-$b+w_1x_1+w_2x_2=0$. 
-In two dimensions we define the vectors $\bm{x} =[x1,x2]$ and $\bm{w}=[w1,w2]$. 
-We can then rewrite the above equation as 
-
-!bt
-\[
-\bm{x}^T\bm{w}+b=0.
-\]
-!et
+The approach is top-down because it begins at the top of the tree (all
+observations belong to a single region) and then successively splits
+the predictor space; each split is indicated via two new branches
+further down on the tree. It is greedy because at each step of the
+tree-building process, the best split is made at that particular step,
+rather than looking ahead and picking a split that will lead to a
+better tree in some future step.
 
 !split
-===== A $p$-dimensional space of features =====
+===== Making a tree =====
 
-We limit ourselves to two classes of outputs $y_i$ and assign these classes the values $y_i = \pm 1$. 
-In a $p$-dimensional space of say $p$ features we have a hyperplane defines as 
+In order to implement the recursive binary splitting we start by selecting
+the predictor $x_j$ and a cutpoint $s$ that splits the predictor space into two regions $R_1$ and $R_2$
 !bt
 \[
-b+wx_1+w_2x_2+\dots +w_px_p=0.
+\left\{X\vert x_j < s\right\},
 \]
 !et
-If we define a 
-matrix $\bm{X}=\left[\bm{x}_1,\bm{x}_2,\dots, \bm{x}_p\right]$
-of dimension $n\times p$, where $n$ represents the observations for each feature and each vector $x_i$ is a column vector of the matrix $\bm{X}$, 
+and
 !bt
 \[
-\bm{x}_i = \begin{bmatrix} x_{i1} \\ x_{i2} \\ \dots \\ \dots \\ x_{ip} \end{bmatrix}.
+\left\{X\vert x_j \geq s\right\},
 \]
 !et
-If the above condition is not met for a given vector $\bm{x}_i$ we have 
+so that we obtain the lowest MSE, that is
 !bt
 \[
-b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} >0,
-\]
-!et
-if our output $y_i=1$.
-In this case we say that $\bm{x}_i$ lies on one of the sides of the hyperplane and if 
-!bt
-\[
-b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip} < 0,
-\]
-!et
-for the class of observations $y_i=-1$, 
-then $\bm{x}_i$ lies on the other side. 
-
-Equivalently, for the two classes of observations we have 
-!bt
-\[
-y_i\left(b+w_1x_{i1}+w_2x_{i2}+\dots +w_px_{ip}\right) > 0. 
+\sum_{i:x_i\in R_j}(y_i-\overline{y}_{R_1})^2+\sum_{i:x_i\in R_2}(y_i-\overline{y}_{R_2})^2,
 \]
 !et
 
-When we try to separate hyperplanes, if it exists, we can use it to construct a natural classifier: a test observation is assigned a given class depending on which side of the hyperplane it is located.
+which we want to minimize by considering all predictors
+$x_1,x_2,\dots,x_p$.  We consider also all possible values of $s$ for
+each predictor. These values could be determined by randomly assigned
+numbers or by starting at the midpoint and then proceed till we find
+an optimal value.
+
+For any $j$ and $s$, we define the pair of half-planes where
+$\overline{y}_{R_1}$ is the mean response for the training
+observations in $R_1(j,s)$, and $\overline{y}_{R_2}$ is the mean
+response for the training observations in $R_2(j,s)$.
+
+Finding the values of $j$ and $s$ that minimize the above equation can be
+done quite quickly, especially when the number of features $p$ is not
+too large.
+
+Next, we repeat the process, looking
+for the best predictor and best cutpoint in order to split the data
+further so as to minimize the MSE within each of the resulting
+regions. However, this time, instead of splitting the entire predictor
+space, we split one of the two previously identified regions. We now
+have three regions. Again, we look to split one of these three regions
+further, so as to minimize the MSE. The process continues until a
+stopping criterion is reached; for instance, we may continue until no
+region contains more than five observations.
 
 !split 
-===== The two-dimensional case ===== 
+===== Pruning the tree =====
 
-Let us try to develop our intuition about SVMs by limiting ourselves to a two-dimensional
-plane.  To separate the two classes of data points, there are many
-possible lines (hyperplanes if you prefer a more strict naming)  
-that could be chosen. Our objective is to find a
-plane that has the maximum margin, i.e the maximum distance between
-data points of both classes. Maximizing the margin distance provides
-some reinforcement so that future data points can be classified with
-more confidence.
+The above procedure is rather straightforward, but leads often to
+overfitting and unnecessarily large and complicated trees. The basic
+idea is to grow a large tree $T_0$ and then prune it back in order to
+obtain a subtree. A smaller tree with fewer splits (fewer regions) can
+lead to smaller variance and better interpretation at the cost of a
+little more bias.
 
-What a linear classifier attempts to accomplish is to split the
-feature space into two half spaces by placing a hyperplane between the
-data points.  This hyperplane will be our decision boundary.  All
-points on one side of the plane will belong to class one and all points
-on the other side of the plane will belong to the second class two.
+The so-called Cost complexity pruning algorithm gives us a
+way to do just this. Rather than considering every possible subtree,
+we consider a sequence of trees indexed by a nonnegative tuning
+parameter $\alpha$.
 
-Unfortunately there are many ways in which we can place a hyperplane
-to divide the data.  Below is an example of two candidate hyperplanes
-for our data sample.
+Read more at the following "Scikit-Learn link on pruning":"https://scikit-learn.org/stable/auto_examples/tree/plot_cost_complexity_pruning.html#sphx-glr-auto-examples-tree-plot-cost-complexity-pruning-py".
 
 !split
-===== Getting into the details =====
+===== Cost complexity pruning =====
 
-Let us define the function
+For each value of $\alpha$  there corresponds a subtree $T \in T_0$ such that
 !bt
 \[
-f(x) = \bm{w}^T\bm{x}+b = 0,
+\sum_{m=1}^{\overline{T}}\sum_{i:x_i\in R_m}(y_i-\overline{y}_{R_m})^2+\alpha\overline{T},
 \]
 !et
-as the function that determines the line $L$ that separates two classes (our two features), see the figure here. 
+is as small as possible. Here $\overline{T}$ is 
+the number of terminal nodes of the tree $T$ , $R_m$ is the
+rectangle (i.e. the subset of predictor space)  corresponding to the $m$-th terminal node.
+
+The tuning parameter $\alpha$ controls a trade-off between the subtree’s
+complexity and its fit to the training data. When $\alpha = 0$, then the
+subtree $T$ will simply equal $T_0$, 
+because then the above equation just measures the
+training error. 
+However, as $\alpha$ increases, there is a price to pay for
+having a tree with many terminal nodes. The above equation will
+tend to be minimized for a smaller subtree. 
 
 
-Any point defined by $\bm{x}_i$ and $\bm{x}_2$ on the line $L$ will satisfy $\bm{w}^T(\bm{x}_1-\bm{x}_2)=0$. 
-
-The signed distance $\delta$ from any point defined by a vector $\bm{x}$ and a point $\bm{x}_0$ on the line $L$ is then
-!bt
-\[
-\delta = \frac{1}{\vert\vert \bm{w}\vert\vert}(\bm{w}^T\bm{x}+b).
-\]
-!et
-
-!split
-===== First attempt at a minimization approach =====
-
-How do we find the parameter $b$ and the vector $\bm{w}$? What we could
-do is to define a cost function which now contains the set of all
-misclassified points $M$ and attempt to minimize this function
-
-!bt
-\[
-C(\bm{w},b) = -\sum_{i\in M} y_i(\bm{w}^T\bm{x}_i+b).
-\]
-!et
-
-We could now for example define all values $y_i =1$ as misclassified in case we have $\bm{w}^T\bm{x}_i+b < 0$ and the opposite if we have $y_i=-1$. Taking the derivatives gives us
-!bt
-\[
-\frac{\partial C}{\partial b} = -\sum_{i\in M} y_i,
-\]
-!et
-and 
-!bt
-\[
-\frac{\partial C}{\partial \bm{w}} = -\sum_{i\in M} y_ix_i.
-\]
-!et
-
-!split
-===== Solving the equations =====
-
-We can now use the Newton-Raphson method or different variants of the gradient descent family (from plain gradient descent to various stochastic gradient descent approaches) to solve the equations
-!bt
-\[
-b \leftarrow b +\eta \frac{\partial C}{\partial b},
-\]
-!et
-and
-!bt
-\[
-\bm{w} \leftarrow \bm{w} +\eta \frac{\partial C}{\partial \bm{w}},
-\]
-!et
-where $\eta$ is our by now well-known learning rate. 
+It turns out that as we increase $\alpha$ from zero
+branches get pruned from the tree in a nested and predictable fashion,
+so obtaining the whole sequence of subtrees as a function of $\alpha$ is
+easy. We can select a value of $\alpha$ using a validation set or using
+cross-validation. We then return to the full data set and obtain the
+subtree corresponding to $\alpha$. 
 
 
 !split
-===== Code Example =====
+===== Schematic Regression Procedure =====
 
-The equations we discussed above can be coded rather easily (the
-framework is similar to what we developed for logistic
-regression). We are going to set up a simple case with two classes only and we want to find a line which separates them the best possible way.
+!bblock Building a Regression Tree
+
+o Use recursive binary splitting to grow a large tree on the training data, stopping only when each terminal node has fewer than some minimum number of observations.
+o Apply cost complexity pruning to the large tree in order to obtain a sequence of best subtrees, as a function of $\alpha$.
+o Use for example $K$-fold cross-validation to choose $\alpha$. Divide the training observations into $K$ folds. For each $k=1,2,\dots,K$ we: 
+  * repeat steps 1 and 2 on all but the $k$-th fold of the training data. 
+  * Then we valuate the mean squared prediction error on the data in the left-out $k$-th fold, as a function of $\alpha$.
+  * Finally  we average the results for each value of $\alpha$, and pick $\alpha$ to minimize the average error.
+o Return the subtree from Step 2 that corresponds to the chosen value of $\alpha$. 
+!eblock
+
+
+!split
+===== A Classification Tree =====
+
+A classification tree is very similar to a regression tree, except
+that it is used to predict a qualitative response rather than a
+quantitative one. Recall that for a regression tree, the predicted
+response for an observation is given by the mean response of the
+training observations that belong to the same terminal node. In
+contrast, for a classification tree, we predict that each observation
+belongs to the most commonly occurring class of training observations
+in the region to which it belongs. In interpreting the results of a
+classification tree, we are often interested not only in the class
+prediction corresponding to a particular terminal node region, but
+also in the class proportions among the training observations that
+fall into that region.  
+
+!split
+===== Growing a classification tree =====
+
+The task of growing a
+classification tree is quite similar to the task of growing a
+regression tree. Just as in the regression setting, we use recursive
+binary splitting to grow a classification tree. However, in the
+classification setting, the MSE cannot be used as a criterion for making
+the binary splits.  A natural alternative to MSE is the _classification
+error rate_. Since we plan to assign an observation in a given region
+to the most commonly occurring error rate class of training
+observations in that region, the classification error rate is simply
+the fraction of the training observations in that region that do not
+belong to the most common class. 
+
+When building a classification tree, either the Gini index or the
+entropy are typically used to evaluate the quality of a particular
+split, since these two approaches are more sensitive to node purity
+than is the classification error rate. 
+
+
+!split
+===== Classification tree, how to split nodes =====
+
+If our targets are the outcome of a classification process that takes
+for example $k=1,2,\dots,K$ values, the only thing we need to think of
+is to set up the splitting criteria for each node.
+
+We define a PDF $p_{mk}$ that represents the number of observations of
+a class $k$ in a region $R_m$ with $N_m$ observations. We represent
+this likelihood function in terms of the proportion $I(y_i=k)$ of
+observations of this class in the region $R_m$ as
+
+!bt
+\[
+p_{mk} = \frac{1}{N_m}\sum_{x_i\in R_m}I(y_i=k).
+\]
+!et 
+
+We let $p_{mk}$ represent the majority class of observations in region
+$m$. The three most common ways of splitting a node are given by
+
+* Misclassification error 
+!bt
+\[
+p_{mk} = \frac{1}{N_m}\sum_{x_i\in R_m}I(y_i\ne k) = 1-p_{mk}.
+\]
+!et 
+* Gini index $g$
+!bt
+\[
+g = \sum_{k=1}^K p_{mk}(1-p_{mk}).
+\]
+!et 
+* Information entropy or just entropy $s$
+!bt
+\[
+s = -\sum_{k=1}^K p_{mk}\log{p_{mk}}.
+\]
+!et 
+
+
+!split
+===== Visualizing the Tree, Classification =====
 !bc pycod
+import os
+from sklearn.datasets import load_breast_cancer
+from sklearn.tree import DecisionTreeClassifier
+from sklearn.model_selection import train_test_split
+from sklearn.metrics import confusion_matrix
+from sklearn.tree import export_graphviz
 
+from IPython.display import Image 
+from pydot import graph_from_dot_data
+import pandas as pd
+import numpy as np
+
+
+cancer = load_breast_cancer()
+X = pd.DataFrame(cancer.data, columns=cancer.feature_names)
+print(X)
+y = pd.Categorical.from_codes(cancer.target, cancer.target_names)
+y = pd.get_dummies(y)
+print(y)
+X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=1)
+tree_clf = DecisionTreeClassifier(max_depth=5)
+tree_clf.fit(X_train, y_train)
+
+export_graphviz(
+    tree_clf,
+    out_file="DataFiles/cancer.dot",
+    feature_names=cancer.feature_names,
+    class_names=cancer.target_names,
+    rounded=True,
+    filled=True
+)
+cmd = 'dot -Tpng DataFiles/cancer.dot -o DataFiles/cancer.png'
+os.system(cmd)
 !ec
 
 !split
-===== Problems with the Simpler Approach =====
-
-
-There are however problems with this approach, although it looks
-pretty straightforward to implement. When running the above code, we see that we can easily end up with many diffeent lines which separate the two classes.
-
-
-For small
-gaps between the entries, we may also end up needing many iterations
-before the solutions converge and if the data cannot be separated
-properly into two distinct classes, we may not experience a converge
-at all.
-
-!split
-===== A better approach =====
-
-A better approach is rather to try to define a large margin between
-the two classes (if they are well separated from the beginning).
-
-Thus, we wish to find a margin $M$ with $\bm{w}$ normalized to
-$\vert\vert \bm{w}\vert\vert =1$ subject to the condition
-
-!bt
-\[
-y_i(\bm{w}^T\bm{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, p. 
-\]
-!et
-All points are thus at a signed distance from the decision boundary defined by the line $L$. The parameters $b$ and $w_1$ and $w_2$ define this line. 
-
-We seek thus the largest value $M$ defined by
-!bt
-\[
-\frac{1}{\vert \vert \bm{w}\vert\vert}y_i(\bm{w}^T\bm{x}_i+b) \geq M \hspace{0.1cm}\forall i=1,2,\dots, n, 
-\]
-!et
-or just 
-!bt
-\[
-y_i(\bm{w}^T\bm{x}_i+b) \geq M\vert \vert \bm{w}\vert\vert \hspace{0.1cm}\forall i. 
-\]
-!et
-If we scale the equation so that $\vert \vert \bm{w}\vert\vert = 1/M$, we have to find the minimum of 
-$\bm{w}^T\bm{w}=\vert \vert \bm{w}\vert\vert$ (the norm) subject to the condition
-!bt
-\[
-y_i(\bm{w}^T\bm{x}_i+b) \geq 1 \hspace{0.1cm}\forall i.
-\]
-!et
-
-We have thus defined our margin as the invers of the norm of
-$\bm{w}$. We want to minimize the norm in order to have a as large as
-possible margin $M$. Before we proceed, we need to remind ourselves
-about Lagrangian multipliers.
-
-!split
-===== A quick Reminder on Lagrangian Multipliers =====
-
-Consider a function of three independent variables $f(x,y,z)$ . For the function $f$ to be an
-extreme we have
-!bt
-\[
-df=0.
-\]
-!et
-A necessary and sufficient condition is
-!bt
-\[
-\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
-\]
-!et
-due to
-!bt
-\[
-df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz.
-\]
-!et
-In many problems the variables $x,y,z$ are often subject to constraints (such as those above for the margin)
-so that they are no longer all independent. It is possible at least in principle to use each 
-constraint to eliminate one variable
-and to proceed with a new and smaller set of independent varables.
-
-The use of so-called Lagrangian  multipliers is an alternative technique  when the elimination
-of variables is incovenient or undesirable.  Assume that we have an equation of constraint on 
-the variables $x,y,z$
-!bt
-\[
-\phi(x,y,z) = 0,
-\]
-!et
- resulting in
-!bt
-\[
-d\phi = \frac{\partial \phi}{\partial x}dx+\frac{\partial \phi}{\partial y}dy+\frac{\partial \phi}{\partial z}dz =0.
-\]
-!et
-Now we cannot set anymore
-!bt
-\[
-\frac{\partial f}{\partial x} =\frac{\partial f}{\partial y}=\frac{\partial f}{\partial z}=0,
-\]
-!et
-if $df=0$ is wanted
-because there are now only two independent variables!  Assume $x$ and $y$ are the independent 
-variables.
-Then $dz$ is no longer arbitrary.
-
-!split
-===== Adding the Multiplier =====
-
-However, we can add to
-!bt
-\[
-df = \frac{\partial f}{\partial x}dx+\frac{\partial f}{\partial y}dy+\frac{\partial f}{\partial z}dz,
-\]
-!et
-a multiplum of $d\phi$, viz. $\lambda d\phi$, resulting  in
-!bt
-\[
-df+\lambda d\phi = (\frac{\partial f}{\partial z}+\lambda
-\frac{\partial \phi}{\partial x})dx+(\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y})dy+
-(\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z})dz =0.
-\]
-!et
-Our multiplier is chosen so that
-!bt
-\[
-\frac{\partial f}{\partial z}+\lambda\frac{\partial \phi}{\partial z} =0.
-\]
-!et
-
-We need to remember that we took $dx$ and $dy$ to be arbitrary and thus we must have
-!bt
-\[
-\frac{\partial f}{\partial x}+\lambda\frac{\partial \phi}{\partial x} =0,
-\]
-!et
-and
-!bt
-\[
-\frac{\partial f}{\partial y}+\lambda\frac{\partial \phi}{\partial y} =0.
-\]
-!et
-When all these equations are satisfied, $df=0$.  We have four unknowns, $x,y,z$ and
-$\lambda$. Actually we want only $x,y,z$, $\lambda$ needs not to be determined, 
-it is therefore often called
-Lagrange's undetermined multiplier.
-If we have a set of constraints $\phi_k$ we have the equations
-!bt
-\[
-\frac{\partial f}{\partial x_i}+\sum_k\lambda_k\frac{\partial \phi_k}{\partial x_i} =0.
-\]
-!et
-
-!split
-===== Setting up the Problem =====
-In order to solve the above problem, we define the following Lagrangian function to be minimized 
-!bt
-\[
-{\cal L}(\lambda,b,\bm{w})=\frac{1}{2}\bm{w}^T\bm{w}-\sum_{i=1}^n\lambda_i\left[y_i(\bm{w}^T\bm{x}_i+b)-1\right],
-\]
-!et
-where $\lambda_i$ is a so-called Lagrange multiplier subject to the condition $\lambda_i \geq 0$.
-
-Taking the derivatives  with respect to $b$ and $\bm{w}$ we obtain 
-!bt
-\[
-\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
-\]
-!et
-and 
-!bt
-\[
-\frac{\partial {\cal L}}{\partial \bm{w}} = 0 = \bm{w}-\sum_{i} \lambda_iy_i\bm{x}_i.
-\]
-!et
-Inserting these constraints into the equation for ${\cal L}$ we obtain
-!bt
-\[
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\bm{x}_i^T\bm{x}_j,
-\]
-!et
-subject to the constraints $\lambda_i\geq 0$ and $\sum_i\lambda_iy_i=0$. 
-We must in addition satisfy the "Karush-Kuhn-Tucker":"https://en.wikipedia.org/wiki/Karush%E2%80%93Kuhn%E2%80%93Tucker_conditions" (KKT) condition
-!bt
-\[
-\lambda_i\left[y_i(\bm{w}^T\bm{x}_i+b) -1\right] \hspace{0.1cm}\forall i.
-\]
-!et
-o If $\lambda_i > 0$, then $y_i(\bm{w}^T\bm{x}_i+b)=1$ and we say that $x_i$ is on the boundary.
-o If $y_i(\bm{w}^T\bm{x}_i+b)> 1$, we say $x_i$ is not on the boundary and we set $\lambda_i=0$. 
-When $\lambda_i > 0$, the vectors $\bm{x}_i$ are called support vectors. They are the vectors closest to the line (or hyperplane) and define the margin $M$. 
-
-!split
-===== The problem to solve =====
-
-We can rewrite 
-!bt
-\[
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\bm{x}_i^T\bm{x}_j,
-\]
-!et
-and its constraints in terms of a matrix-vector problem where we minimize w.r.t. $\lambda$ the following problem
-!bt
-\[
-\frac{1}{2} \bm{\lambda}^T\begin{bmatrix} y_1y_1\bm{x}_1^T\bm{x}_1 & y_1y_2\bm{x}_1^T\bm{x}_2 & \dots & \dots & y_1y_n\bm{x}_1^T\bm{x}_n \\
-y_2y_1\bm{x}_2^T\bm{x}_1 & y_2y_2\bm{x}_2^T\bm{x}_2 & \dots & \dots & y_1y_n\bm{x}_2^T\bm{x}_n \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1\bm{x}_n^T\bm{x}_1 & y_ny_2\bm{x}_n^T\bm{x}_2 & \dots & \dots & y_ny_n\bm{x}_n^T\bm{x}_n \\
-\end{bmatrix}\bm{\lambda}-\mathbb{1}\bm{\lambda}, 
-\]
-!et
-subject to $\bm{y}^T\bm{\lambda}=0$. Here we defined the vectors $\bm{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n]$ and 
-$\bm{y}=[y_1,y_2,\dots,y_n]$. 
-
-
-!split
-===== The last steps =====
-
-Solving the above problem, yields the values of $\lambda_i$.
-To find the coefficients of your hyperplane we need simply to compute 
-!bt
-\[
-\bm{w}=\sum_{i} \lambda_iy_i\bm{x}_i.
-\]
-!et
-With our vector $\bm{w}$ we can in turn find the value of the intercept $b$ (here in two dimensions) via 
-!bt
-\[
-y_i(\bm{w}^T\bm{x}_i+b)=1,
-\]
-!et
-resulting in
-!bt
-\[
-b = \frac{1}{y_i}-\bm{w}^T\bm{x}_i,
-\]
-!et
-or if we write it out in terms of the support vectors only, with $N_s$ being their number,  we have
-!bt
-\[
-b = \frac{1}{N_s}\sum_{j\in N_s}\left(y_j-\sum_{i=1}^n\lambda_iy_i\bm{x}_i^T\bm{x}_j\right).
-\]
-!et
-With our hyperplane coefficients we can use our classifier to assign any observation by simply using 
-!bt
-\[
-y_i = \mathrm{sign}(\bm{w}^T\bm{x}_i+b).
-\]
-!et
-Below we discuss how to find the optimal values of $\lambda_i$. Before we proceed however, we discuss now the so-called soft classifier. 
-
-!split
-===== A soft classifier =====
-
-Till now, the margin is strictly defined by the support vectors. This defines what is called a hard classifier, that is the margins are well defined.
-
-Suppose now that classes overlap in feature space, as shown in the
-figure here. One way to deal with this problem before we define the
-so-called _kernel approach_, is to allow a kind of slack in the sense
-that we allow some points to be on the wrong side of the margin.
-
-We introduce thus the so-called _slack_ variables $\bm{\xi} =[\xi_1,x_2,\dots,x_n]$ and 
-modify our previous equation
-!bt
-\[
-y_i(\bm{w}^T\bm{x}_i+b)=1,
-\]
-!et
-to 
-!bt
-\[
-y_i(\bm{w}^T\bm{x}_i+b)=1-\xi_i,
-\]
-!et
-with the requirement $\xi_i\geq 0$. The total violation is now $\sum_i\xi$. 
-The value $\xi_i$ in the constraint the last constraint corresponds to the  amount by which the prediction
-$y_i(\bm{w}^T\bm{x}_i+b)=1$ is on the wrong side of its margin. Hence by bounding the sum $\sum_i \xi_i$,
-we bound the total amount by which predictions fall on the wrong side of their margins.
-
-Misclassifications occur when $\xi_i > 1$. Thus bounding the total sum by some value $C$ bounds in turn the total number of
-misclassifications.
-
-!split
-===== Soft optmization problem =====
-
-
-This has in turn the consequences that we change our optmization problem to finding the minimum of 
-!bt
-\[
-{\cal L}=\frac{1}{2}\bm{w}^T\bm{w}-\sum_{i=1}^n\lambda_i\left[y_i(\bm{w}^T\bm{x}_i+b)-(1-\xi_)\right]+C\sum_{i=1}^n\xi_i-\sum_{i=1}^n\gamma_i\xi_i,
-\]
-!et
-subject to 
-!bt
-\[
-y_i(\bm{w}^T\bm{x}_i+b)=1-\xi_i \hspace{0.1cm}\forall i,
-\]
-!et
-with the requirement $\xi_i\geq 0$.
-
-Taking the derivatives  with respect to $b$ and $\bm{w}$ we obtain 
-!bt
-\[
-\frac{\partial {\cal L}}{\partial b} = -\sum_{i} \lambda_iy_i=0,
-\]
-!et
-and 
-!bt
-\[
-\frac{\partial {\cal L}}{\partial \bm{w}} = 0 = \bm{w}-\sum_{i} \lambda_iy_i\bm{x}_i, 
-\]
-!et
-and
-!bt
-\[
-\lambda_i = C-\gamma_i \hspace{0.1cm}\forall i.
-\]
-!et
-Inserting these constraints into the equation for ${\cal L}$ we obtain the same equation as before
-!bt
-\[
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\bm{x}_i^T\bm{x}_j,
-\]
-!et
-but now subject to the constraints $\lambda_i\geq 0$, $\sum_i\lambda_iy_i=0$ and $0\leq\lambda_i \leq C$. 
-We must in addition satisfy the Karush-Kuhn-Tucker condition which now reads
-!bt
-\[
-\lambda_i\left[y_i(\bm{w}^T\bm{x}_i+b) -(1-\xi_)\right]=0 \hspace{0.1cm}\forall i,
-\]
-!et
-!bt
-\[
-\gamma_i\xi_i = 0,
-\]
-!et
-and 
-!bt
-\[
-y_i(\bm{w}^T\bm{x}_i+b) -(1-\xi_) \geq 0 \hspace{0.1cm}\forall i.
-\]
-!et
-
-!split
-===== Kernels and non-linearity =====
-
-The cases we have studied till now, were all characterized by two classes
-with a close to linear separability. The classifiers we have described
-so far find linear boundaries in our input feature space. It is
-possible to make our procedure more flexible by exploring the feature
-space using other basis expansions such as higher-order polynomials,
-wavelets, splines etc.
-
-If our feature space is not easy to separate, as shown in the figure
-here, we can achieve a better separation by introducing more complex
-basis functions. The ideal would be, as shown in the next figure, to, via a specific transformation to 
-obtain a separation between the classes which is almost linear. 
-
-The change of basis, from $x\rightarrow z=\phi(x)$ leads to the same type of equations to be solved, except that
-we need to introduce for example a polynomial transformation to a two-dimensional training set.
-
+===== Visualizing the Tree, The Moons  =====
 !bc pycod
+# Common imports
 import numpy as np
+from sklearn.model_selection import  train_test_split 
+from sklearn.tree import DecisionTreeClassifier
+from sklearn.datasets import make_moons
+from sklearn.tree import export_graphviz
+from pydot import graph_from_dot_data
+import pandas as pd
 import os
 
 np.random.seed(42)
+X, y = make_moons(n_samples=100, noise=0.25, random_state=53)
+X_train, X_test, y_train, y_test = train_test_split(X,y,random_state=0)
+tree_clf = DecisionTreeClassifier(max_depth=5)
+tree_clf.fit(X_train, y_train)
 
-# To plot pretty figures
-import matplotlib
-import matplotlib.pyplot as plt
-plt.rcParams['axes.labelsize'] = 14
-plt.rcParams['xtick.labelsize'] = 12
-plt.rcParams['ytick.labelsize'] = 12
+export_graphviz(
+    tree_clf,
+    out_file="DataFiles/moons.dot",
+    rounded=True,
+    filled=True
+)
+cmd = 'dot -Tpng DataFiles/moons.dot -o DataFiles/moons.png'
+os.system(cmd)
+!ec
 
+!split
+===== Other ways of visualizing the trees =====
 
-from sklearn.svm import SVC
-from sklearn import datasets
+_Scikit-Learn_ has also another way to visualize the trees which is very useful, here with the Iris data.
 
+!bc pycod
+from sklearn.datasets import load_iris
+from sklearn import tree
+X, y = load_iris(return_X_y=True)
+tree_clf = tree.DecisionTreeClassifier()
+tree_clf = tree_clf.fit(X, y)
+# and then plot the tree
+tree.plot_tree(tree_clf) 
+!ec
 
+!split
+===== Printing out as text =====
 
-X1D = np.linspace(-4, 4, 9).reshape(-1, 1)
-X2D = np.c_[X1D, X1D**2]
-y = np.array([0, 0, 1, 1, 1, 1, 1, 0, 0])
-
-plt.figure(figsize=(11, 4))
-
-plt.subplot(121)
-plt.grid(True, which='both')
-plt.axhline(y=0, color='k')
-plt.plot(X1D[:, 0][y==0], np.zeros(4), "bs")
-plt.plot(X1D[:, 0][y==1], np.zeros(5), "g^")
-plt.gca().get_yaxis().set_ticks([])
-plt.xlabel(r"$x_1$", fontsize=20)
-plt.axis([-4.5, 4.5, -0.2, 0.2])
-
-plt.subplot(122)
-plt.grid(True, which='both')
-plt.axhline(y=0, color='k')
-plt.axvline(x=0, color='k')
-plt.plot(X2D[:, 0][y==0], X2D[:, 1][y==0], "bs")
-plt.plot(X2D[:, 0][y==1], X2D[:, 1][y==1], "g^")
-plt.xlabel(r"$x_1$", fontsize=20)
-plt.ylabel(r"$x_2$", fontsize=20, rotation=0)
-plt.gca().get_yaxis().set_ticks([0, 4, 8, 12, 16])
-plt.plot([-4.5, 4.5], [6.5, 6.5], "r--", linewidth=3)
-plt.axis([-4.5, 4.5, -1, 17])
-plt.subplots_adjust(right=1)
-plt.show()
+Alternatively, the tree can also be exported in textual format with the function exporttext.
+This method doesn’t require the installation of external libraries and is more compact:
 
+!bc pycod
+from sklearn.datasets import load_iris
+from sklearn.tree import DecisionTreeClassifier
+from sklearn.tree import export_text
+iris = load_iris()
+decision_tree = DecisionTreeClassifier(random_state=0, max_depth=2)
+decision_tree = decision_tree.fit(iris.data, iris.target)
+r = export_text(decision_tree, feature_names=iris['feature_names'])
+print(r)
 !ec
 
 
+!split
+===== Algorithms for Setting up Decision Trees =====
+
+Two algorithms stand out in the set up of decision trees:
+o The CART (Classification And Regression Tree) algorithm for both classification and regression
+o The ID3 algorithm based on the computation of the information gain for classification
+
+We discuss both algorithms with applications here. The popular library
+_Scikit-Learn_ uses the CART algorithm. For classification problems
+you can use either the _gini_ index or the _entropy_ to split a tree
+in two branches.
 
 !split
-===== The equations =====
+===== The CART algorithm for Classification =====
 
-Suppose we define a polynomial transformation of degree two only (we continue to live in a plane with $x_i$ and $y_i$ as variables)
+For classification, the CART algorithm splits the data set in two subsets using a single feature $k$ and a threshold $t_k$.
+This could be for example a threshold set by a number below a certain circumference of a malign tumor.
+
+How do we find these two quantities?
+We search for the pair $(k,t_k)$ that produces the purest subset using for example the _gini_ factor $G$.
+The cost function it tries to minimize is then
 !bt
 \[
-z = \phi(x_i) =\left(x_i^2, y_i^2, \sqrt{2}x_iy_i\right).
+C(k,t_k) = \frac{m_{\mathrm{left}}}{m}G_{\mathrm{left}}+ \frac{m_{\mathrm{right}}}{m}G_{\mathrm{right}},
 \]
 !et
+where $G_{\mathrm{left/right}}$ measures the impurity of the left/right subset  and $m_{\mathrm{left/right}}$
+ is the number of instances in the left/right subset
 
-With our new basis, the equations we solved earlier are basically the same, that is we have now (without the slack option for simplicity)
+Once it has successfully split the training set in two, it splits the subsets using the same logic, then the subsubsets
+and so on, recursively. It stops recursing once it reaches the maximum depth (defined by the
+$max\_depth$ hyperparameter), or if it cannot find a split that will reduce impurity. A few other
+hyperparameters control additional stopping conditions such as the $min\_samples\_split$,
+$min\_samples\_leaf$, $min\_weight\_fraction\_leaf$, and $max\_leaf\_nodes$.
+
+!split
+===== The CART algorithm for Regression =====
+
+The CART algorithm for regression works is similar to the one for classification except that instead of trying to split the
+training set in a way that minimizes say the _gini_ or _entropy_ impurity, it now tries to split the training set in a way that minimizes our well-known mean-squared error (MSE). The cost function is now
 !bt
 \[
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\bm{z}_i^T\bm{z}_j,
+C(k,t_k) = \frac{m_{\mathrm{left}}}{m}\mathrm{MSE}_{\mathrm{left}}+ \frac{m_{\mathrm{right}}}{m}\mathrm{MSE}_{\mathrm{right}}.
 \]
 !et
-subject to the constraints $\lambda_i\geq 0$, $\sum_i\lambda_iy_i=0$, and for the support vectors
+Here the MSE for a specific node is defined as
 !bt
 \[
-y_i(\bm{w}^T\bm{z}_i+b)= 1 \hspace{0.1cm}\forall i,
+\mathrm{MSE}_{\mathrm{node}}=\frac{1}{m_\mathrm{node}}\sum_{i\in \mathrm{node}}(\overline{y}_{\mathrm{node}}-y_i)^2,
 \]
 !et
-from which we also find $b$.
-To compute $\bm{z}_i^T\bm{z}_j$ we define the kernel $K(\bm{x}_i,\bm{x}_j)$ as
+with
 !bt
 \[
-K(\bm{x}_i,\bm{x}_j)=\bm{z}_i^T\bm{z}_j= \phi(\bm{x}_i)^T\phi(\bm{x}_j).
-\]
-!et
-For the above example, the kernel reads
-!bt
-\[
-K(\bm{x}_i,\bm{x}_j)=[x_i^2, y_i^2, \sqrt{2}x_iy_i]^T\begin{bmatrix} x_j^2 \\ y_j^2 \\ \sqrt{2}x_jy_j \end{bmatrix}=x_i^2x_j^2+2x_ix_jy_iy_j+y_i^2y_j^2.
+\overline{y}_{\mathrm{node}}=\frac{1}{m_\mathrm{node}}\sum_{i\in \mathrm{node}}y_i,
 \]
 !et
+the mean value of all observations in a specific node.
 
-We note that this is nothing but the dot product of the two original
-vectors $(\bm{x}_i^T\bm{x}_j)^2$. Instead of thus computing the
-product in the Lagrangian of $\bm{z}_i^T\bm{z}_j$ we simply compute
-the dot product $(\bm{x}_i^T\bm{x}_j)^2$.
+Without any regularization, the regression task for decision trees, 
+just like for classification tasks, is  prone to overfitting.
 
+!split
+===== Why binary splits? =====
 
-This leads to the so-called
-kernel trick and the result leads to the same as if we went through
-the trouble of performing the transformation
-$\phi(\bm{x}_i)^T\phi(\bm{x}_j)$ during the SVM calculations.
+It is custom to split to a tree uising binary splits. The reason is
+that multiway splits fragment the data too quickly, leaving
+insufficient data at the next level down.  Multiway splits can be
+achieved by a series of binary split and this is normally preferred.
 
 
 !split
-===== The problem to solve =====
-Using our definition of the kernel We can rewrite again the Lagrangian
-!bt
-\[
-{\cal L}=\sum_i\lambda_i-\frac{1}{2}\sum_{ij}^n\lambda_i\lambda_jy_iy_j\bm{x}_i^T\bm{z}_j,
-\]
-!et
-subject to the constraints $\lambda_i\geq 0$, $\sum_i\lambda_iy_i=0$ in terms of a convex optimization problem
-!bt
-\[
-\frac{1}{2} \bm{\lambda}^T\begin{bmatrix} y_1y_1K(\bm{x}_1,\bm{x}_1) & y_1y_2K(\bm{x}_1,\bm{x}_2) & \dots & \dots & y_1y_nK(\bm{x}_1,\bm{x}_n) \\
-y_2y_1K(\bm{x}_2,\bm{x}_1) & y_2y_2(\bm{x}_2,\bm{x}_2) & \dots & \dots & y_1y_nK(\bm{x}_2,\bm{x}_n) \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1K(\bm{x}_n,\bm{x}_1) & y_ny_2K(\bm{x}_n\bm{x}_2) & \dots & \dots & y_ny_nK(\bm{x}_n,\bm{x}_n) \\
-\end{bmatrix}\bm{\lambda}-\mathbb{1}\bm{\lambda}, 
-\]
-!et
-subject to $\bm{y}^T\bm{\lambda}=0$. Here we defined the vectors $\bm{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n]$ and 
-$\bm{y}=[y_1,y_2,\dots,y_n]$. 
-If we add the slack constants this leads to the additional constraint $0\leq \lambda_i \leq C$.
+===== Computing a Tree using the Gini Index =====
 
-We can rewrite this (see the solutions below) in terms of a convex optimization problem of the type
-!bt
-\begin{align*}
-    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\bm{\lambda}^T\bm{P}\bm{\lambda}+\bm{q}^T\bm{\lambda},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \bm{G}\bm{\lambda} \preceq \bm{h} \hspace{0.2cm} \wedge \bm{A}\bm{\lambda}=f.
-\end{align*}
-!et
-Below we discuss how to solve these equations. Here we note that the matrix $\bm{P}$ has matrix elements $p_{ij}=y_iy_jK(\bm{x}_i,\bm{x}_j)$.
-Given a kernel $K$ and the targets $y_i$ this matrix is easy to set up. The constraint $\bm{y}^T\bm{\lambda}=0$ leads to $f=0$ and $\bm{A}=\bm{y}$. How to set up the matrix $\bm{G}$ is discussed later. Here note that the inequalities $0\leq \lambda_i \leq C$ can be split up into
-$0\leq \lambda_i$ and $\lambda_i \leq C$. These two inequalities define then the matrix $\bm{G}$ and the vector $\bm{h}$.
+Consider the following example with attributes/features and two
+possible outcomes (classes)  for each attribute.  Assume we wish to find some
+correlations between the average grade of a student as function of the
+number of hours studied and hours slept. We want also to correlate the
+grade in a given course with the general trend, whether the students
+recently has gotten grades below average or above.
+
+We have three features/attributes
+o Trend of average grades before present course, classified as either below  or above  the average grade of the whole class  
+o The number of hours studies, classified again as either higher (more than 3 hours per day) or lower . Here we have used a standard for one $ECTS$ which is scaled to 25-30 hours of work for a semester which lasts 18 weeks, with 15 weeks of lectures and 3 weeks for exams, assuming a total of 30 ECTS per semester. 
+o The number of hours slept as high for more than $8$ hours and below  for less than 8 hours of sleep, classified again as either high or low
+o The final grade whether it is above or below average
 
 
 !split
-===== Different kernels and Mercer's theorem =====
+=====  The Table =====
 
-There are several popular kernels being used. These are
-o Linear: $K(\bm{x},\bm{y})=\bm{x}^T\bm{y}$,
-o Polynomial: $K(\bm{x},\bm{y})=(\bm{x}^T\bm{y}+\gamma)^d$,
-o Gaussian Radial Basis Function: $K(\bm{x},\bm{y})=\exp{\left(-\gamma\vert\vert\bm{x}-\bm{y}\vert\vert^2\right)}$,
-o Tanh: $K(\bm{x},\bm{y})=\tanh{(\bm{x}^T\bm{y}+\gamma)}$,
-and many other ones.
-
-An important theorem for us is "Mercer's
-theorem":"https://en.wikipedia.org/wiki/Mercer%27s_theorem".  The
-theorem states that if a kernel function $K$ is symmetric, continuous
-and leads to a positive semi-definite matrix $\bm{P}$ then there
-exists a function $\phi$ that maps $\bm{x}_i$ and $\bm{x}_j$ into
-another space (possibly with much higher dimensions) such that
-
-!bt
-\[
-K(\bm{x}_i,\bm{x}_j)=\phi(\bm{x}_i)^T\phi(\bm{x}_j).
-\]
-!et
-
-So you can use $K$ as a kernel since you know $\phi$ exists, even if
-you don’t know what $\phi$ is. 
-
-Note that some frequently used kernels (such as the Sigmoid kernel)
-don’t respect all of Mercer’s conditions, yet they generally work well
-in practice.
+|---------------------------------------------------|
+| Grade Trend | Hours slept | Hours Studied | Grade |
+|---------------------------------------------------|
+| Above | Low | High | Above |
+| Below | High | Low | Below |
+| Above | Low | High | Above |
+| Above | High | High | Above |
+| Below | Low | High | Below |
+| Above | Low | Low | Below |
+| Below | High | High | Below |
+| Below | Low | High  | Below |
+| Above | Low | Low | Below |
+| Above | High | High | Above |
+|---------------------------------------------------|
 
 
 !split
-===== The moons example =====
+===== Computing the various Gini Indices =====
+
+In computations we will translate all classes into numbers. Being
+these binary classes, they can easily be split into ones and zeros.
+
+!bblock Gini index for Average trend
+"See handwritten notes November 3":"https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf"
+!eblock
+
+
+!split
+===== Computing the various Gini Indices, Hours slept =====
+
+
+!bblock Gini index for hour slept
+"See handwritten notes November 3":"https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf"
+!eblock
+
+
+!split
+===== Computing the various Gini Indices, Hours studied =====
+
+
+!bblock Gini index for hour studied
+"See handwritten notes November 3":"https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/2022/NotesNov32022.pdf"
+!eblock
+
+For final tree, see the above handwritten notes
+
+
+!split
+===== A possible code using Scikit-Learn =====
+
 !bc pycod
-from __future__ import division, print_function, unicode_literals
-
+# Common imports
 import numpy as np
-np.random.seed(42)
+import pandas as pd
+import matplotlib.pyplot as plt
+from sklearn.tree import DecisionTreeClassifier
+from sklearn.model_selection import train_test_split
+from sklearn.tree import export_graphviz
+from sklearn.preprocessing import StandardScaler, OneHotEncoder
+from sklearn.compose import ColumnTransformer
+from IPython.display import Image 
+from pydot import graph_from_dot_data
+import os
 
+# Where to save the figures and data files
+PROJECT_ROOT_DIR = "Results"
+FIGURE_ID = "Results/FigureFiles"
+DATA_ID = "DataFiles/"
+
+if not os.path.exists(PROJECT_ROOT_DIR):
+    os.mkdir(PROJECT_ROOT_DIR)
+
+if not os.path.exists(FIGURE_ID):
+    os.makedirs(FIGURE_ID)
+
+if not os.path.exists(DATA_ID):
+    os.makedirs(DATA_ID)
+
+def image_path(fig_id):
+    return os.path.join(FIGURE_ID, fig_id)
+
+def data_path(dat_id):
+    return os.path.join(DATA_ID, dat_id)
+
+def save_fig(fig_id):
+    plt.savefig(image_path(fig_id) + ".png", format='png')
+
+infile = open(data_path("grades.csv"),'r')
+
+# Read the experimental data with Pandas
+from IPython.display import display
+grades = pd.read_csv(infile)
+grades = pd.DataFrame(grades)
+display(grades)
+# Features and targets
+X = grades.loc[:, grades.columns != 'Grade'].values
+y = grades.loc[:, grades.columns == 'Grade'].values
+print(X)
+# Then do a Classification tree
+tree_clf = DecisionTreeClassifier(max_depth=2)
+tree_clf.fit(X, y)
+print("Train set accuracy with Decision Tree: {:.2f}".format(tree_clf.score(X,y)))
+#transfer to a decision tree graph
+export_graphviz(
+    tree_clf,
+    out_file="DataFiles/grade.dot",
+    rounded=True,
+    filled=True
+)
+cmd = 'dot -Tpng DataFiles/grade.dot -o DataFiles/grades.png'
+os.system(cmd)
+
+
+!ec
+
+
+
+!split
+===== Further example: Computing the Gini index =====
+
+The next example we will look at is a classical one in many Machine
+Learning applications. Based on various meteorological features, we
+have several so-called attributes which decide whether we at the end
+will do some outdoor activity like skiing, going for a bike ride etc
+etc.  The table here contains the feautures _outlook_, _temperature_,
+_humidity_ and _wind_.  The target or output is whether we ride
+(True=1) or whether we do something else that day (False=0). The
+attributes for each feature are then sunny, overcast and rain for the
+outlook, hot, cold and mild for temperature, high and normal for
+humidity and weak and strong for wind.
+
+The table here summarizes the various attributes and
+|-------------------------------------------|
+|Day| Outlook |Temperature | Humidity | Wind | Ride|
+|-------------------------------------------|
+|1 | Sunny | Hot | High | Weak | 0 |
+|2 | Sunny | Hot | High | Strong | 1 |
+|3 | Overcast | Hot | High | Weak | 1 |
+|4 | Rain | Mild | High | Weak | 1 |
+|5 | Rain | Cool | Normal | Weak | 1 |
+|6 | Rain | Cool | Normal | Strong | 0 |
+|7 | Overcast | Cool | Normal | Strong | 1 |
+|8 | Sunny | Mild | High | Weak | 0 |
+|9 | Sunny | Cool | Normal | Weak | 1 |
+|10 | Rain | Mild | Normal | Weak | 1 |
+|11 | Sunny | Mild | Normal | Strong | 1 |
+|12 | Overcast | Mild | High | Strong | 1 |
+|13 | Overcast | Hot | Normal | Weak | 1 |
+|14 | Rain | Mild | High | Strong | 0 |
+|-------------------------------------------|
+
+
+!split
+===== Simple Python Code to read in Data and perform Classification =====
+
+!bc pycod
+# Common imports
+import numpy as np
+import pandas as pd
+import matplotlib.pyplot as plt
+from sklearn.tree import DecisionTreeClassifier
+from sklearn.model_selection import train_test_split
+from sklearn.tree import export_graphviz
+from sklearn.preprocessing import StandardScaler, OneHotEncoder
+from sklearn.compose import ColumnTransformer
+from IPython.display import Image 
+from pydot import graph_from_dot_data
+import os
+
+# Where to save the figures and data files
+PROJECT_ROOT_DIR = "Results"
+FIGURE_ID = "Results/FigureFiles"
+DATA_ID = "DataFiles/"
+
+if not os.path.exists(PROJECT_ROOT_DIR):
+    os.mkdir(PROJECT_ROOT_DIR)
+
+if not os.path.exists(FIGURE_ID):
+    os.makedirs(FIGURE_ID)
+
+if not os.path.exists(DATA_ID):
+    os.makedirs(DATA_ID)
+
+def image_path(fig_id):
+    return os.path.join(FIGURE_ID, fig_id)
+
+def data_path(dat_id):
+    return os.path.join(DATA_ID, dat_id)
+
+def save_fig(fig_id):
+    plt.savefig(image_path(fig_id) + ".png", format='png')
+
+infile = open(data_path("rideclass.csv"),'r')
+
+# Read the experimental data with Pandas
+from IPython.display import display
+ridedata = pd.read_csv(infile,names = ('Outlook','Temperature','Humidity','Wind','Ride'))
+ridedata = pd.DataFrame(ridedata)
+
+# Features and targets
+X = ridedata.loc[:, ridedata.columns != 'Ride'].values
+y = ridedata.loc[:, ridedata.columns == 'Ride'].values
+
+# Create the encoder.
+encoder = OneHotEncoder(handle_unknown="ignore")
+# Assume for simplicity all features are categorical.
+encoder.fit(X)    
+# Apply the encoder.
+X = encoder.transform(X)
+print(X)
+# Then do a Classification tree
+tree_clf = DecisionTreeClassifier(max_depth=2)
+tree_clf.fit(X, y)
+print("Train set accuracy with Decision Tree: {:.2f}".format(tree_clf.score(X,y)))
+#transfer to a decision tree graph
+export_graphviz(
+    tree_clf,
+    out_file="DataFiles/ride.dot",
+    rounded=True,
+    filled=True
+)
+cmd = 'dot -Tpng DataFiles/cancer.dot -o DataFiles/cancer.png'
+os.system(cmd)
+
+
+!ec
+
+!split
+===== Computing the Gini Factor  =====
+
+The above functions (gini, entropy and misclassification error) are
+important components of the so-called CART algorithm. We will discuss
+this algorithm below after we have discussed the information gain
+algorithm ID3.
+
+In the example here we have converted all our attributes into numerical values $0,1,2$ etc.
+
+!bc pycod
+# Split a dataset based on an attribute and an attribute value
+def test_split(index, value, dataset):
+	left, right = list(), list()
+	for row in dataset:
+		if row[index] < value:
+			left.append(row)
+		else:
+			right.append(row)
+	return left, right
+ 
+# Calculate the Gini index for a split dataset
+def gini_index(groups, classes):
+	# count all samples at split point
+	n_instances = float(sum([len(group) for group in groups]))
+	# sum weighted Gini index for each group
+	gini = 0.0
+	for group in groups:
+		size = float(len(group))
+		# avoid divide by zero
+		if size == 0:
+			continue
+		score = 0.0
+		# score the group based on the score for each class
+		for class_val in classes:
+			p = [row[-1] for row in group].count(class_val) / size
+			score += p * p
+		# weight the group score by its relative size
+		gini += (1.0 - score) * (size / n_instances)
+	return gini
+
+# Select the best split point for a dataset
+def get_split(dataset):
+	class_values = list(set(row[-1] for row in dataset))
+	b_index, b_value, b_score, b_groups = 999, 999, 999, None
+	for index in range(len(dataset[0])-1):
+		for row in dataset:
+			groups = test_split(index, row[index], dataset)
+			gini = gini_index(groups, class_values)
+			print('X%d < %.3f Gini=%.3f' % ((index+1), row[index], gini))
+			if gini < b_score:
+				b_index, b_value, b_score, b_groups = index, row[index], gini, groups
+	return {'index':b_index, 'value':b_value, 'groups':b_groups}
+ 
+dataset = [[0,0,0,0,0],
+            [0,0,0,1,1],
+            [1,0,0,0,1],
+            [2,1,0,0,1],
+            [2,2,1,0,1],
+            [2,2,1,1,0],
+            [1,2,1,1,1],
+            [0,1,0,0,0],
+            [0,2,1,0,1],
+            [2,1,1,0,1],
+            [0,1,1,1,1],
+            [1,1,0,1,1],
+            [1,0,1,0,1],
+            [2,1,0,1,0]]
+
+split = get_split(dataset)
+print('Split: [X%d < %.3f]' % ((split['index']+1), split['value']))
+!ec
+
+
+
+!split
+===== Regression trees =====
+!bc pycod 
+# Quadratic training set + noise
+np.random.seed(42)
+m = 200
+X = np.random.rand(m, 1)
+y = 4 * (X - 0.5) ** 2
+y = y + np.random.randn(m, 1) / 10
+!ec
+
+!bc pycod 
+from sklearn.tree import DecisionTreeRegressor
+
+tree_reg = DecisionTreeRegressor(max_depth=2, random_state=42)
+tree_reg.fit(X, y)
+!ec
+
+!split
+===== Final regressor code =====
+!bc pycod
+from sklearn.tree import DecisionTreeRegressor
+
+tree_reg1 = DecisionTreeRegressor(random_state=42, max_depth=2)
+tree_reg2 = DecisionTreeRegressor(random_state=42, max_depth=3)
+tree_reg1.fit(X, y)
+tree_reg2.fit(X, y)
+
+def plot_regression_predictions(tree_reg, X, y, axes=[0, 1, -0.2, 1], ylabel="$y$"):
+    x1 = np.linspace(axes[0], axes[1], 500).reshape(-1, 1)
+    y_pred = tree_reg.predict(x1)
+    plt.axis(axes)
+    plt.xlabel("$x_1$", fontsize=18)
+    if ylabel:
+        plt.ylabel(ylabel, fontsize=18, rotation=0)
+    plt.plot(X, y, "b.")
+    plt.plot(x1, y_pred, "r.-", linewidth=2, label=r"$\hat{y}$")
+
+plt.figure(figsize=(11, 4))
+plt.subplot(121)
+plot_regression_predictions(tree_reg1, X, y)
+for split, style in ((0.1973, "k-"), (0.0917, "k--"), (0.7718, "k--")):
+    plt.plot([split, split], [-0.2, 1], style, linewidth=2)
+plt.text(0.21, 0.65, "Depth=0", fontsize=15)
+plt.text(0.01, 0.2, "Depth=1", fontsize=13)
+plt.text(0.65, 0.8, "Depth=1", fontsize=13)
+plt.legend(loc="upper center", fontsize=18)
+plt.title("max_depth=2", fontsize=14)
+
+plt.subplot(122)
+plot_regression_predictions(tree_reg2, X, y, ylabel=None)
+for split, style in ((0.1973, "k-"), (0.0917, "k--"), (0.7718, "k--")):
+    plt.plot([split, split], [-0.2, 1], style, linewidth=2)
+for split in (0.0458, 0.1298, 0.2873, 0.9040):
+    plt.plot([split, split], [-0.2, 1], "k:", linewidth=1)
+plt.text(0.3, 0.5, "Depth=2", fontsize=13)
+plt.title("max_depth=3", fontsize=14)
+
+plt.show()
+!ec
+
+!bc pycod
+tree_reg1 = DecisionTreeRegressor(random_state=42)
+tree_reg2 = DecisionTreeRegressor(random_state=42, min_samples_leaf=10)
+tree_reg1.fit(X, y)
+tree_reg2.fit(X, y)
+
+x1 = np.linspace(0, 1, 500).reshape(-1, 1)
+y_pred1 = tree_reg1.predict(x1)
+y_pred2 = tree_reg2.predict(x1)
+
+plt.figure(figsize=(11, 4))
+
+plt.subplot(121)
+plt.plot(X, y, "b.")
+plt.plot(x1, y_pred1, "r.-", linewidth=2, label=r"$\hat{y}$")
+plt.axis([0, 1, -0.2, 1.1])
+plt.xlabel("$x_1$", fontsize=18)
+plt.ylabel("$y$", fontsize=18, rotation=0)
+plt.legend(loc="upper center", fontsize=18)
+plt.title("No restrictions", fontsize=14)
+
+plt.subplot(122)
+plt.plot(X, y, "b.")
+plt.plot(x1, y_pred2, "r.-", linewidth=2, label=r"$\hat{y}$")
+plt.axis([0, 1, -0.2, 1.1])
+plt.xlabel("$x_1$", fontsize=18)
+plt.title("min_samples_leaf={}".format(tree_reg2.min_samples_leaf), fontsize=14)
+
+plt.show()
+!ec
+
+
+
+
+!split
+===== Pros and cons of trees, pros =====
+
+* White box, easy to interpret model. Some people believe that decision trees more closely mirror human decision-making than do the regression and classification approaches discussed earlier (think of support vector machines)
+* Trees are very easy to explain to people. In fact, they are even easier to explain than linear regression!
+* No feature normalization needed
+* Tree models can handle both continuous and categorical data (Classification and Regression Trees)
+* Can model nonlinear relationships
+* Can model interactions between the different descriptive features
+* Trees can be displayed graphically, and are easily interpreted even by a non-expert (especially if they are small)
+
+
+!split
+===== Disadvantages =====
+
+* Unfortunately, trees generally do not have the same level of predictive accuracy as some of the other regression and classification approaches
+* If continuous features are used the tree may become quite large and hence less interpretable
+* Decision trees are prone to overfit the training data and hence do not well generalize the data if no stopping criteria or improvements like pruning, boosting or bagging are implemented
+* Small changes in the data may lead to a completely different tree. This issue can be addressed by using ensemble methods like bagging, boosting or random forests
+* Unbalanced datasets where some target feature values occur much more frequently than others may lead to biased trees since the frequently occurring feature values are preferred over the less frequently occurring ones. 
+* If the number of features is relatively large (high dimensional) and the number of instances is relatively low, the tree might overfit the data
+* Features with many levels may be preferred over features with less levels since for them it is *more easy* to split the dataset such that the sub datasets only contain pure target feature values. This issue can be addressed by preferring for instance the information gain ratio as splitting criteria over information gain
+
+However, by aggregating many decision trees, using methods like
+bagging, random forests, and boosting, the predictive performance of
+trees can be substantially improved.
+
+
+!split
+===== Ensemble Methods: From a Single Tree to Many Trees and Extreme Boosting, Meet the Jungle of Methods =====
+
+As stated above and seen in many of the examples discussed here about
+a single decision tree, we often end up overfitting our training
+data. This normally means that we have a high variance. Can we reduce
+the variance of a statistical learning method?
+
+This leads us to a set of different methods that can combine different
+machine learning algorithms or just use one of them to construct
+forests and jungles of trees, homogeneous ones or heterogenous
+ones. These methods are recognized by different names which we will
+try to explain here. These are
+
+o Voting classifiers
+o Bagging and Pasting
+o Random forests
+o Boosting methods, from adaptive to Extreme Gradient Boosting (XGBoost)
+
+We discuss these methods here.
+
+
+!split
+===== An Overview of Ensemble Methods  =====
+
+FIGURE: [DataFiles/ensembleoverview.png, width=600 frac=0.8]
+
+!split
+=====  Why Voting? =====
+
+The idea behind boosting, and voting as well can be phrased as follows:
+_Can a group of people somehow arrive at highly
+reasoned decisions, despite the weak judgement of the individual
+members?_
+
+The aim is to create a good classifier by combining several weak classifiers.
+_A weak classifier is a classifier which is able to produce results that are only slightly better than guessing at random._
+
+The basic approach is to apply repeatedly (in boosting this is done in an iterative way) a weak classifier to modifications of the data.
+In voting we simply apply the law of large numbers while in boosting we give more weight to misclassified data in
+each iteration.  
+
+Decision trees play an important role as our weak classifier. They serve as the basic method. 
+
+!split
+===== Tossing coins =====
+
+The simplest case is a so-called voting ensemble. To illustrate this,
+think of yourself tossing coins with a biased outcome of 51 per cent
+for heads and 49% for tails. With only few tosses,
+you may not clearly see this distribution for heads and tails. However, after some
+thousands of tosses, there will be a clear majority of heads.  With 2000 tosses
+you should see approximately 1020 heads and 980 tails.
+
+We can then state that the outcome is a clear majority of heads. If
+you do this ten thousand times, it is easy to see that there is a 97%
+likelihood of a majority of heads.
+
+Another example would be to collect all polls before an
+election. Different polls may show different likelihoods for a
+candidate winning with say a majority  of the popular vote. The majority vote
+would then consist in many polls indicating that this candidate will
+actually win.
+
+The example here shows how we can implement the coin tossing case,
+clealry demostrating that after some tosses we see the "law of large":"https://en.wikipedia.org/wiki/Law_of_large_numbers"
+numbers kicking in.
+
+!split
+===== Standard imports first =====
+
+!bc pycod
+# Common imports
+from IPython.display import Image 
+from pydot import graph_from_dot_data
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+from sklearn.tree import DecisionTreeClassifier
+from sklearn.model_selection import train_test_split
+from sklearn.tree import export_graphviz
+from sklearn.preprocessing import StandardScaler, OneHotEncoder
+from sklearn.compose import ColumnTransformer
+from IPython.display import Image 
+from pydot import graph_from_dot_data
+import os
+
+# Where to save the figures and data files
+PROJECT_ROOT_DIR = "Results"
+FIGURE_ID = "Results/FigureFiles"
+DATA_ID = "DataFiles/"
+
+if not os.path.exists(PROJECT_ROOT_DIR):
+    os.mkdir(PROJECT_ROOT_DIR)
+
+if not os.path.exists(FIGURE_ID):
+    os.makedirs(FIGURE_ID)
+
+if not os.path.exists(DATA_ID):
+    os.makedirs(DATA_ID)
+
+def image_path(fig_id):
+    return os.path.join(FIGURE_ID, fig_id)
+
+def data_path(dat_id):
+    return os.path.join(DATA_ID, dat_id)
+
+def save_fig(fig_id):
+    plt.savefig(image_path(fig_id) + ".png", format='png')
+
+!ec
+
+
+
+
+!split
+===== Simple Voting Example, head or tail =====
+!bc pycod
+
+# Common imports
+import numpy as np
 import matplotlib
 import matplotlib.pyplot as plt
+from matplotlib.colors import ListedColormap
 plt.rcParams['axes.labelsize'] = 14
 plt.rcParams['xtick.labelsize'] = 12
 plt.rcParams['ytick.labelsize'] = 12
 
+heads_proba = 0.51
+coin_tosses = (np.random.rand(10000, 10) < heads_proba).astype(np.int32)
+cumulative_heads_ratio = np.cumsum(coin_tosses, axis=0) / np.arange(1, 10001).reshape(-1, 1)
+plt.figure(figsize=(8,3.5))
+plt.plot(cumulative_heads_ratio)
+plt.plot([0, 10000], [0.51, 0.51], "k--", linewidth=2, label="51%")
+plt.plot([0, 10000], [0.5, 0.5], "k-", label="50%")
+plt.xlabel("Number of coin tosses")
+plt.ylabel("Heads ratio")
+plt.legend(loc="lower right")
+plt.axis([0, 10000, 0.42, 0.58])
+save_fig("votingsimple")
+plt.show()
 
+!ec
+
+
+!split
+===== Using the Voting Classifier =====
+
+We can use the voting classifier on other data sets, here the exciting binary case of two distinct objects using the make moons functionality of _Scikit-Learn_.
+!bc pycod
+from sklearn.model_selection import train_test_split
+from sklearn.datasets import make_moons
+
+X, y = make_moons(n_samples=500, noise=0.30, random_state=42)
+X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)
+
+from sklearn.ensemble import RandomForestClassifier
+from sklearn.ensemble import VotingClassifier
+from sklearn.linear_model import LogisticRegression
 from sklearn.svm import SVC
-from sklearn import datasets
+
+log_clf = LogisticRegression(solver="liblinear", random_state=42)
+rnd_clf = RandomForestClassifier(n_estimators=10, random_state=42)
+svm_clf = SVC(gamma="auto", random_state=42)
+
+voting_clf = VotingClassifier(
+    estimators=[('lr', log_clf), ('rf', rnd_clf), ('svc', svm_clf)],
+    voting='hard')
+
+voting_clf.fit(X_train, y_train)
+
+from sklearn.metrics import accuracy_score
+
+for clf in (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    print(clf.__class__.__name__, accuracy_score(y_test, y_pred))
+
+log_clf = LogisticRegression(solver="liblinear", random_state=42)
+rnd_clf = RandomForestClassifier(n_estimators=10, random_state=42)
+svm_clf = SVC(gamma="auto", probability=True, random_state=42)
+voting_clf = VotingClassifier(
+    estimators=[('lr', log_clf), ('rf', rnd_clf), ('svc', svm_clf)],
+    voting='soft')
+voting_clf.fit(X_train, y_train)
+
+from sklearn.metrics import accuracy_score
+
+for clf in (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    print(clf.__class__.__name__, accuracy_score(y_test, y_pred))
+
+!ec
 
 
 
-from sklearn.pipeline import Pipeline
+
+!split
+===== Voting and Bagging =====
+
+!bc pycod
+from sklearn.model_selection import train_test_split
+from sklearn.datasets import make_moons
+
+X, y = make_moons(n_samples=500, noise=0.30, random_state=42)
+X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)
+from sklearn.ensemble import RandomForestClassifier
+from sklearn.ensemble import VotingClassifier
+from sklearn.linear_model import LogisticRegression
+from sklearn.svm import SVC
+
+log_clf = LogisticRegression(random_state=42)
+rnd_clf = RandomForestClassifier(random_state=42)
+svm_clf = SVC(random_state=42)
+
+voting_clf = VotingClassifier(
+    estimators=[('lr', log_clf), ('rf', rnd_clf), ('svc', svm_clf)],
+    voting='hard')
+voting_clf.fit(X_train, y_train)
+!ec
+
+!bc pycod 
+from sklearn.metrics import accuracy_score
+
+for clf in (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    print(clf.__class__.__name__, accuracy_score(y_test, y_pred))
+!ec
+
+!bc pycod 
+log_clf = LogisticRegression(random_state=42)
+rnd_clf = RandomForestClassifier(random_state=42)
+svm_clf = SVC(probability=True, random_state=42)
+
+voting_clf = VotingClassifier(
+    estimators=[('lr', log_clf), ('rf', rnd_clf), ('svc', svm_clf)],
+    voting='soft')
+voting_clf.fit(X_train, y_train)
+!ec
+
+!bc pycod 
+from sklearn.metrics import accuracy_score
+
+for clf in (log_clf, rnd_clf, svm_clf, voting_clf):
+    clf.fit(X_train, y_train)
+    y_pred = clf.predict(X_test)
+    print(clf.__class__.__name__, accuracy_score(y_test, y_pred))
+!ec
+
+
+
+
+
+!split
+===== Bagging =====
+
+The _plain_ decision trees suffer from high
+variance. This means that if we split the training data into two parts
+at random, and fit a decision tree to both halves, the results that we
+get could be quite different. In contrast, a procedure with low
+variance will yield similar results if applied repeatedly to distinct
+data sets; linear regression tends to have low variance, if the ratio
+of $n$ to $p$ is moderately large. 
+
+_Bootstrap aggregation_, or just _bagging_, is a
+general-purpose procedure for reducing the variance of a statistical
+learning method. 
+
+
+!split
+===== More bagging =====
+
+Bagging typically results in improved accuracy
+over prediction using a single tree. Unfortunately, however, it can be
+difficult to interpret the resulting model. Recall that one of the
+advantages of decision trees is the attractive and easily interpreted
+diagram that results.
+
+However, when we bag a large number of trees, it is no longer
+possible to represent the resulting statistical learning procedure
+using a single tree, and it is no longer clear which variables are
+most important to the procedure. Thus, bagging improves prediction
+accuracy at the expense of interpretability.  Although the collection
+of bagged trees is much more difficult to interpret than a single
+tree, one can obtain an overall summary of the importance of each
+predictor using the MSE (for bagging regression trees) or the Gini
+index (for bagging classification trees). In the case of bagging
+regression trees, we can record the total amount that the MSE is
+decreased due to splits over a given predictor, averaged over all $B$ possible
+trees. A large value indicates an important predictor. Similarly, in
+the context of bagging classification trees, we can add up the total
+amount that the Gini index  is decreased by splits over a given
+predictor, averaged over all $B$ trees.
+
+
+
+!split
+=====  Making your own Bootstrap: Changing the Level of the Decision Tree =====
+
+Let us bring up our good old boostrap example from the linear regression lectures. We change the linerar regression algorithm with
+a decision tree wth different depths and perform a bootstrap aggregate (in this case we perform as many bootstraps as data points $n$).
+!bc pycod
+
+import matplotlib.pyplot as plt
+import numpy as np
+from sklearn.model_selection import train_test_split
+from sklearn.pipeline import make_pipeline
+from sklearn.utils import resample
+from sklearn.tree import DecisionTreeRegressor
+
+n = 100
+n_boostraps = 100
+maxdepth = 8
+
+# Make data set.
+x = np.linspace(-3, 3, n).reshape(-1, 1)
+y = np.exp(-x**2) + 1.5 * np.exp(-(x-2)**2)+ np.random.normal(0, 0.1, x.shape)
+error = np.zeros(maxdepth)
+bias = np.zeros(maxdepth)
+variance = np.zeros(maxdepth)
+polydegree = np.zeros(maxdepth)
+X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.2)
+
 from sklearn.preprocessing import StandardScaler
-from sklearn.svm import LinearSVC
+scaler = StandardScaler()
+scaler.fit(X_train)
+X_train_scaled = scaler.transform(X_train)
+X_test_scaled = scaler.transform(X_test)
 
+# we produce a simple tree first as benchmark
+simpletree = DecisionTreeRegressor(max_depth=3) 
+simpletree.fit(X_train_scaled, y_train)
+simpleprediction = simpletree.predict(X_test_scaled)
+for degree in range(1,maxdepth):
+    model = DecisionTreeRegressor(max_depth=degree) 
+    y_pred = np.empty((y_test.shape[0], n_boostraps))
+    for i in range(n_boostraps):
+        x_, y_ = resample(X_train_scaled, y_train)
+        model.fit(x_, y_)
+        y_pred[:, i] = model.predict(X_test_scaled)#.ravel()
 
-from sklearn.datasets import make_moons
-X, y = make_moons(n_samples=100, noise=0.15, random_state=42)
-
-def plot_dataset(X, y, axes):
-    plt.plot(X[:, 0][y==0], X[:, 1][y==0], "bs")
-    plt.plot(X[:, 0][y==1], X[:, 1][y==1], "g^")
-    plt.axis(axes)
-    plt.grid(True, which='both')
-    plt.xlabel(r"$x_1$", fontsize=20)
-    plt.ylabel(r"$x_2$", fontsize=20, rotation=0)
-
-plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])
-plt.show()
-
-from sklearn.datasets import make_moons
-from sklearn.pipeline import Pipeline
-from sklearn.preprocessing import PolynomialFeatures
-
-polynomial_svm_clf = Pipeline([
-        ("poly_features", PolynomialFeatures(degree=3)),
-        ("scaler", StandardScaler()),
-        ("svm_clf", LinearSVC(C=10, loss="hinge", random_state=42))
-    ])
-
-polynomial_svm_clf.fit(X, y)
-
-def plot_predictions(clf, axes):
-    x0s = np.linspace(axes[0], axes[1], 100)
-    x1s = np.linspace(axes[2], axes[3], 100)
-    x0, x1 = np.meshgrid(x0s, x1s)
-    X = np.c_[x0.ravel(), x1.ravel()]
-    y_pred = clf.predict(X).reshape(x0.shape)
-    y_decision = clf.decision_function(X).reshape(x0.shape)
-    plt.contourf(x0, x1, y_pred, cmap=plt.cm.brg, alpha=0.2)
-    plt.contourf(x0, x1, y_decision, cmap=plt.cm.brg, alpha=0.1)
-
-plot_predictions(polynomial_svm_clf, [-1.5, 2.5, -1, 1.5])
-plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])
-
-plt.show()
-
-
-from sklearn.svm import SVC
-
-poly_kernel_svm_clf = Pipeline([
-        ("scaler", StandardScaler()),
-        ("svm_clf", SVC(kernel="poly", degree=3, coef0=1, C=5))
-    ])
-poly_kernel_svm_clf.fit(X, y)
-
-poly100_kernel_svm_clf = Pipeline([
-        ("scaler", StandardScaler()),
-        ("svm_clf", SVC(kernel="poly", degree=10, coef0=100, C=5))
-    ])
-poly100_kernel_svm_clf.fit(X, y)
-
-plt.figure(figsize=(11, 4))
-
-plt.subplot(121)
-plot_predictions(poly_kernel_svm_clf, [-1.5, 2.5, -1, 1.5])
-plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])
-plt.title(r"$d=3, r=1, C=5$", fontsize=18)
-
-plt.subplot(122)
-plot_predictions(poly100_kernel_svm_clf, [-1.5, 2.5, -1, 1.5])
-plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])
-plt.title(r"$d=10, r=100, C=5$", fontsize=18)
-
-plt.show()
-
-def gaussian_rbf(x, landmark, gamma):
-    return np.exp(-gamma * np.linalg.norm(x - landmark, axis=1)**2)
-
-gamma = 0.3
-
-x1s = np.linspace(-4.5, 4.5, 200).reshape(-1, 1)
-x2s = gaussian_rbf(x1s, -2, gamma)
-x3s = gaussian_rbf(x1s, 1, gamma)
-
-XK = np.c_[gaussian_rbf(X1D, -2, gamma), gaussian_rbf(X1D, 1, gamma)]
-yk = np.array([0, 0, 1, 1, 1, 1, 1, 0, 0])
-
-plt.figure(figsize=(11, 4))
-
-plt.subplot(121)
-plt.grid(True, which='both')
-plt.axhline(y=0, color='k')
-plt.scatter(x=[-2, 1], y=[0, 0], s=150, alpha=0.5, c="red")
-plt.plot(X1D[:, 0][yk==0], np.zeros(4), "bs")
-plt.plot(X1D[:, 0][yk==1], np.zeros(5), "g^")
-plt.plot(x1s, x2s, "g--")
-plt.plot(x1s, x3s, "b:")
-plt.gca().get_yaxis().set_ticks([0, 0.25, 0.5, 0.75, 1])
-plt.xlabel(r"$x_1$", fontsize=20)
-plt.ylabel(r"Similarity", fontsize=14)
-plt.annotate(r'$\mathbf{x}$',
-             xy=(X1D[3, 0], 0),
-             xytext=(-0.5, 0.20),
-             ha="center",
-             arrowprops=dict(facecolor='black', shrink=0.1),
-             fontsize=18,
-            )
-plt.text(-2, 0.9, "$x_2$", ha="center", fontsize=20)
-plt.text(1, 0.9, "$x_3$", ha="center", fontsize=20)
-plt.axis([-4.5, 4.5, -0.1, 1.1])
-
-plt.subplot(122)
-plt.grid(True, which='both')
-plt.axhline(y=0, color='k')
-plt.axvline(x=0, color='k')
-plt.plot(XK[:, 0][yk==0], XK[:, 1][yk==0], "bs")
-plt.plot(XK[:, 0][yk==1], XK[:, 1][yk==1], "g^")
-plt.xlabel(r"$x_2$", fontsize=20)
-plt.ylabel(r"$x_3$  ", fontsize=20, rotation=0)
-plt.annotate(r'$\phi\left(\mathbf{x}\right)$',
-             xy=(XK[3, 0], XK[3, 1]),
-             xytext=(0.65, 0.50),
-             ha="center",
-             arrowprops=dict(facecolor='black', shrink=0.1),
-             fontsize=18,
-            )
-plt.plot([-0.1, 1.1], [0.57, -0.1], "r--", linewidth=3)
-plt.axis([-0.1, 1.1, -0.1, 1.1])
-    
-plt.subplots_adjust(right=1)
-
-plt.show()
-
-
-x1_example = X1D[3, 0]
-for landmark in (-2, 1):
-    k = gaussian_rbf(np.array([[x1_example]]), np.array([[landmark]]), gamma)
-    print("Phi({}, {}) = {}".format(x1_example, landmark, k))
-
-rbf_kernel_svm_clf = Pipeline([
-        ("scaler", StandardScaler()),
-        ("svm_clf", SVC(kernel="rbf", gamma=5, C=0.001))
-    ])
-rbf_kernel_svm_clf.fit(X, y)
-
-
-from sklearn.svm import SVC
-
-gamma1, gamma2 = 0.1, 5
-C1, C2 = 0.001, 1000
-hyperparams = (gamma1, C1), (gamma1, C2), (gamma2, C1), (gamma2, C2)
-
-svm_clfs = []
-for gamma, C in hyperparams:
-    rbf_kernel_svm_clf = Pipeline([
-            ("scaler", StandardScaler()),
-            ("svm_clf", SVC(kernel="rbf", gamma=gamma, C=C))
-        ])
-    rbf_kernel_svm_clf.fit(X, y)
-    svm_clfs.append(rbf_kernel_svm_clf)
-
-plt.figure(figsize=(11, 7))
-
-for i, svm_clf in enumerate(svm_clfs):
-    plt.subplot(221 + i)
-    plot_predictions(svm_clf, [-1.5, 2.5, -1, 1.5])
-    plot_dataset(X, y, [-1.5, 2.5, -1, 1.5])
-    gamma, C = hyperparams[i]
-    plt.title(r"$\gamma = {}, C = {}$".format(gamma, C), fontsize=16)
-
+    polydegree[degree] = degree
+    error[degree] = np.mean( np.mean((y_test - y_pred)**2, axis=1, keepdims=True) )
+    bias[degree] = np.mean( (y_test - np.mean(y_pred, axis=1, keepdims=True))**2 )
+    variance[degree] = np.mean( np.var(y_pred, axis=1, keepdims=True) )
+    print('Polynomial degree:', degree)
+    print('Error:', error[degree])
+    print('Bias^2:', bias[degree])
+    print('Var:', variance[degree])
+    print('{} >= {} + {} = {}'.format(error[degree], bias[degree], variance[degree], bias[degree]+variance[degree]))
+ 
+mse_simpletree= np.mean( np.mean((y_test - simpleprediction)**2))
+print("Simple tree:",mse_simpletree)
+plt.xlim(1,maxdepth)
+plt.plot(polydegree, error, label='MSE')
+plt.plot(polydegree, bias, label='bias')
+plt.plot(polydegree, variance, label='Variance')
+plt.legend()
+save_fig("baggingboot")
 plt.show()
 
 !ec
 
 
 
-!split
-=====  Mathematical optimization of convex functions =====
 
-A mathematical (quadratic) optimization problem, or just optimization problem, has the form
+
+
+
+
+!split
+===== Random forests =====
+
+Random forests provide an improvement over bagged trees by way of a
+small tweak that decorrelates the trees. 
+
+As in bagging, we build a
+number of decision trees on bootstrapped training samples. But when
+building these decision trees, each time a split in a tree is
+considered, a random sample of $m$ predictors is chosen as split
+candidates from the full set of $p$ predictors. The split is allowed to
+use only one of those $m$ predictors. 
+
+A fresh sample of $m$ predictors is
+taken at each split, and typically we choose 
+
 !bt
-\begin{align*}
-    &\mathrm{min}_{\lambda}\hspace{0.2cm} \frac{1}{2}\bm{\lambda}^T\bm{P}\bm{\lambda}+\bm{q}^T\bm{\lambda},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm}to} \hspace{0.2cm} \bm{G}\bm{\lambda} \preceq \bm{h} \wedge  \bm{A}\bm{\lambda}=f.
-\end{align*}
+\[
+m\approx \sqrt{p}.
+\]
 !et
-subject to some constraints for say a selected set $i=1,2,\dots, n$.
-In our case we are optimizing with respect to the Lagrangian multipliers $\lambda_i$, and the
-vector $\bm{\lambda}=[\lambda_1, \lambda_2,\dots, \lambda_n]$ is the optimization variable we are dealing with.
 
-In our case we are particularly interested in a class of optimization problems called convex optmization problems. 
-In our discussion on gradient descent methods we discussed at length the definition of a convex function. 
+In building a random forest, at
+each split in the tree, the algorithm is not even allowed to consider
+a majority of the available predictors. 
 
-Convex optimization problems play a central role in applied mathematics and we recommend strongly "Boyd and Vandenberghe's text on the topics":"http://web.stanford.edu/~boyd/cvxbook/".
+The reason for this is rather clever. Suppose that there is one very
+strong predictor in the data set, along with a number of other
+moderately strong predictors. Then in the collection of bagged
+variable importance random forest trees, most or all of the trees will
+use this strong predictor in the top split. Consequently, all of the
+bagged trees will look quite similar to each other. Hence the
+predictions from the bagged trees will be highly correlated.
+Unfortunately, averaging many highly correlated quantities does not
+lead to as large of a reduction in variance as averaging many
+uncorrelated quantities. In particular, this means that bagging will
+not lead to a substantial reduction in variance over a single tree in
+this setting.
+
+
+!split
+===== Random Forest Algorithm =====
+The algorithm described here can be applied to both classification and regression problems.
+
+We will grow of forest of say $B$ trees.
+o For $b=1:B$
+  * Draw a bootstrap sample from the training data organized in our $\bm{X}$ matrix.
+  * We grow then a random forest tree $T_b$ based on the bootstrapped data by repeating the steps outlined till we reach the maximum node size is reached
+    o we select $m \le p$ variables at random from the $p$ predictors/features
+    o pick the best split point among the $m$ features using for example the CART algorithm and create a new node
+    o split the node into daughter nodes
+o Output then the ensemble of trees $\{T_b\}_1^{B}$ and make predictions for either a regression type of problem or a classification type of problem. 
 
 
 
 !split
-===== How do we solve these problems? =====
+===== Random Forests Compared with other Methods on the Cancer Data =====
+!bc pycod 
+import matplotlib.pyplot as plt
+import numpy as np
+from sklearn.model_selection import  train_test_split 
+from sklearn.datasets import load_breast_cancer
+from sklearn.svm import SVC
+from sklearn.linear_model import LogisticRegression
+from sklearn.tree import DecisionTreeClassifier
+from sklearn.ensemble import BaggingClassifier
+
+# Load the data
+cancer = load_breast_cancer()
+
+X_train, X_test, y_train, y_test = train_test_split(cancer.data,cancer.target,random_state=0)
+print(X_train.shape)
+print(X_test.shape)
+#define methods
+# Logistic Regression
+logreg = LogisticRegression(solver='lbfgs')
+# Support vector machine
+svm = SVC(gamma='auto', C=100)
+# Decision Trees
+deep_tree_clf = DecisionTreeClassifier(max_depth=None)
+#Scale the data
+from sklearn.preprocessing import StandardScaler
+scaler = StandardScaler()
+scaler.fit(X_train)
+X_train_scaled = scaler.transform(X_train)
+X_test_scaled = scaler.transform(X_test)
+# Logistic Regression
+logreg.fit(X_train_scaled, y_train)
+print("Test set accuracy Logistic Regression with scaled data: {:.2f}".format(logreg.score(X_test_scaled,y_test)))
+# Support Vector Machine
+svm.fit(X_train_scaled, y_train)
+print("Test set accuracy SVM with scaled data: {:.2f}".format(logreg.score(X_test_scaled,y_test)))
+# Decision Trees
+deep_tree_clf.fit(X_train_scaled, y_train)
+print("Test set accuracy with Decision Trees and scaled data: {:.2f}".format(deep_tree_clf.score(X_test_scaled,y_test)))
+
+
+from sklearn.ensemble import RandomForestClassifier
+from sklearn.preprocessing import LabelEncoder
+from sklearn.model_selection import cross_validate
+# Data set not specificied
+#Instantiate the model with 500 trees and entropy as splitting criteria
+Random_Forest_model = RandomForestClassifier(n_estimators=500,criterion="entropy")
+Random_Forest_model.fit(X_train_scaled, y_train)
+#Cross validation
+accuracy = cross_validate(Random_Forest_model,X_test_scaled,y_test,cv=10)['test_score']
+print(accuracy)
+print("Test set accuracy with Random Forests and scaled data: {:.2f}".format(Random_Forest_model.score(X_test_scaled,y_test)))
+
+
+import scikitplot as skplt
+y_pred = Random_Forest_model.predict(X_test_scaled)
+skplt.metrics.plot_confusion_matrix(y_test, y_pred, normalize=True)
+plt.show()
+y_probas = Random_Forest_model.predict_proba(X_test_scaled)
+skplt.metrics.plot_roc(y_test, y_probas)
+plt.show()
+skplt.metrics.plot_cumulative_gain(y_test, y_probas)
+plt.show()
+
+!ec
+
+
+Recall that the cumulative gains curve shows the percentage of the
+overall number of cases in a given category *gained* by targeting a
+percentage of the total number of cases.
+
+Similarly, the receiver operating characteristic curve, or ROC curve,
+displays the diagnostic ability of a binary classifier system as its
+discrimination threshold is varied. It plots the true positive rate against the false positive rate.
+
+
+!split
+===== Compare  Bagging on Trees with Random Forests  =====
+!bc pycod
+bag_clf = BaggingClassifier(
+    DecisionTreeClassifier(splitter="random", max_leaf_nodes=16, random_state=42),
+    n_estimators=500, max_samples=1.0, bootstrap=True, n_jobs=-1, random_state=42)
+!ec
 
-If we use Python as programming language and wish to venture beyond
-_scikit-learn_, _tensorflow_ and similar software which makes our
-lives so much easier, we need to dive into the wonderful world of
-quadratic programming. We can, if we wish, solve the minimization
-problem using say standard gradient methods or conjugate gradient
-methods. However, these methods tend to exhibit a rather slow
-converge. So, welcome to the promised land of quadratic programming.
 
-The functions we need are contained in the "quadratic programming library CVXOPT":"https://cvxopt.org/userguide/coneprog.html" and we need to import it together with _numpy_ as
 
 !bc pycod
-import numpy
-import cvxopt
+bag_clf.fit(X_train, y_train)
+y_pred = bag_clf.predict(X_test)
+from sklearn.ensemble import RandomForestClassifier
+rnd_clf = RandomForestClassifier(n_estimators=500, max_leaf_nodes=16, n_jobs=-1, random_state=42)
+rnd_clf.fit(X_train, y_train)
+y_pred_rf = rnd_clf.predict(X_test)
+np.sum(y_pred == y_pred_rf) / len(y_pred) 
 !ec
 
-This will make our life much easier. You don't need t0 write your own optimizer.
+
+
+
+
 
 
 !split
-===== A simple example =====
+===== Boosting, a Bird's Eye View =====
 
-We remind ourselves about the general problem we want to solve
+The basic idea is to combine weak classifiers in order to create a good
+classifier. With a weak classifier we often intend a classifier which
+produces results which are only slightly better than we would get by
+random guesses.
+
+This is done by applying in an iterative way a weak (or a standard
+classifier like decision trees) to modify the data. In each iteration
+we emphasize those observations which are misclassified by weighting
+them with a factor.
+
+
+!split
+===== What is boosting? Additive Modelling/Iterative Fitting =====
+
+Boosting is a way of fitting an additive expansion in a set of
+elementary basis functions like for example some simple polynomials.
+Assume for example that we have a function
 !bt
-\begin{align*}
-    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}\bm{x}^T\bm{P}\bm{x}+\bm{q}^T\bm{x},\\ \nonumber
-    &\mathrm{subject\hspace{0.1cm} to} \hspace{0.2cm} \bm{G}\bm{x} \preceq \bm{h} \wedge  \bm{A}\bm{x}=f.
-\end{align*}
+\[
+f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m),
+\]
 !et
 
-Let us show how to perform the optmization using a simple case. Assume we want to optimize the following problem
-!bt
-\begin{align*}
-    &\mathrm{min}_{x}\hspace{0.2cm} \frac{1}{2}x^2+5x+3y \\ \nonumber
-    &\mathrm{subject}\hspace{0.5cm} \mathrm{to} \\ \nonumber
-    &x, y \geq 0 \\ \nonumber
-    &x+3y  \geq 15 \\ \nonumber
-    &2x+5y  \leq  100 \\ \nonumber
-    &3x+4y  \leq  80.  \\ \nonumber
-\end{align*}
-!et
-The minimization problem can be rewritten in terms of vectors and matrices as (with $x$ and $y$ being the unknowns)
+where $\beta_m$ are the expansion parameters to be determined in a
+minimization process and $b(x;\gamma_m)$ are some simple functions of
+the multivariable parameter $x$ which is characterized by the
+parameters $\gamma_m$.
+
+As an example, consider the Sigmoid function we used in logistic
+regression. In that case, we can translate the function
+$b(x;\gamma_m)$ into the Sigmoid function
+
+
 !bt
 \[
-\frac{1}{2}\begin{bmatrix} x\\ y \end{bmatrix}^T   \begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}  \begin{bmatrix} x \\ y \end{bmatrix}  + \begin{bmatrix}3\\ 4  \end{bmatrix}^T \begin{bmatrix}x \\ y  \end{bmatrix}.  
+\sigma(t) = \frac{1}{1+\exp{(-t)}},
 \]
 !et
-Similarly, we can now set up the inequalities (we need to change $\geq$ to $\leq$ by multiplying with $-1$ on bot sides) as the following matrix-vector equation
+
+where $t=\gamma_0+\gamma_1 x$ and the parameters $\gamma_0$ and
+$\gamma_1$ were determined by the Logistic Regression fitting
+algorithm.
+
+As another example, consider the cost function we defined for linear regression
 !bt
 \[
-\begin{bmatrix} -1 & 0 \\ 0 & -1 \\ -1 & -3 \\ 2 & 5 \\ 3 & 4\end{bmatrix}\begin{bmatrix} x \\ y\end{bmatrix} \preceq \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
+C(\bm{y},\bm{f}) = \frac{1}{n} \sum_{i=0}^{n-1}(y_i-f(x_i))^2.
 \]
 !et
-We have collapsed all the inequalities into a single matrix $\bm{G}$. We see also that our matrix 
+
+In this case the function $f(x)$ was replaced by the design matrix
+$\bm{X}$ and the unknown linear regression parameters $\bm{\beta}$,
+that is $\bm{f}=\bm{X}\bm{\beta}$. In linear regression we can 
+simply invert a matrix and obtain the parameters $\beta$ by
+
 !bt
 \[
-\bm{P} =\begin{bmatrix} 1 & 0\\ 0 & 0  \end{bmatrix}
+\bm{\beta}=\left(\bm{X}^T\bm{X}\right)^{-1}\bm{X}^T\bm{y}.
 \]
 !et
-is clearly positive semi-definite (all eigenvalues larger or equal zero). 
-Finally, the vector $\bm{h}$ is defined as 
+
+In iterative fitting or additive modeling, we minimize the cost function with respect to the parameters $\beta_m$ and $\gamma_m$.
+
+
+!split
+===== Iterative Fitting, Regression and Squared-error Cost Function =====
+
+The way we proceed is as follows (here we specialize to the squared-error cost function)
+
+o Establish a cost function, here $C(\bm{y},\bm{f}) = \frac{1}{n} \sum_{i=0}^{n-1}(y_i-f_M(x_i))^2$ with $f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m)$.
+o Initialize with a guess $f_0(x)$. It could be one or even zero or some random numbers.
+o For $m=1:M$
+  o minimize $\sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2$ wrt $\gamma$ and $\beta$
+  o This gives the optimal values $\beta_m$ and $\gamma_m$
+  o Determine then the new values $f_m(x)=f_{m-1}(x) +\beta_m b(x;\gamma_m)$
+
+We could use any of the algorithms we have discussed till now. If we
+use trees, $\gamma$ parameterizes the split variables and split points
+at the internal nodes, and the predictions at the terminal nodes.
+
+
+!split
+===== Squared-Error Example and Iterative Fitting =====
+
+To better understand what happens, let us develop the steps for the iterative fitting using the above squared error function.
+
+For simplicity we assume also that our functions $b(x;\gamma)=1+\gamma x$. 
+
+This means that for every iteration $m$, we need to optimize
+
 !bt
 \[
-\bm{h} = \begin{bmatrix}0 \\ 0\\ -15 \\ 100 \\ 80\end{bmatrix}.
+(\beta_m,\gamma_m) = \mathrm{argmin}_{\beta,\lambda}\hspace{0.1cm} \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2=\sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta(1+\gamma x_i))^2.
+\]
+!et
+
+We start our iteration by simply setting $f_0(x)=0$. 
+Taking the derivatives  with respect to $\beta$ and $\gamma$ we obtain
+!bt
+\[
+\frac{\partial {\cal C}}{\partial \beta} = -2\sum_{i}(1+\gamma x_i)(y_i-\beta(1+\gamma x_i))=0,
+\]
+!et
+and
+!bt
+\[
+\frac{\partial {\cal C}}{\partial \gamma} =-2\sum_{i}\beta x_i(y_i-\beta(1+\gamma x_i))=0.
+\]
+!et
+We can then rewrite these equations as (defining $\bm{w}=\bm{e}+\gamma \bm{x})$ with $\bm{e}$ being the unit vector)
+!bt
+\[
+\gamma \bm{w}^T(\bm{y}-\beta\gamma \bm{w})=0,
+\]
+!et
+which gives us $\beta = \bm{w}^T\bm{y}/(\bm{w}^T\bm{w})$. Similarly we have 
+!bt
+\[
+\beta\gamma \bm{x}^T(\bm{y}-\beta(1+\gamma \bm{x}))=0,
+\]
+!et
+
+which leads to $\gamma =(\bm{x}^T\bm{y}-\beta\bm{x}^T\bm{e})/(\beta\bm{x}^T\bm{x})$.  Inserting
+for $\beta$ gives us an equation for $\gamma$. This is a non-linear equation in the unknown $\gamma$ and has to be solved numerically. 
+
+The solution to these two equations gives us in turn $\beta_1$ and $\gamma_1$ leading to the new expression for $f_1(x)$ as
+$f_1(x) = \beta_1(1+\gamma_1x)$. Doing this $M$ times results in our final estimate for the function $f$. 
+
+
+
+!split
+===== Iterative Fitting, Classification and AdaBoost =====
+
+Let us consider a binary classification problem with two outcomes $y_i \in \{-1,1\}$ and $i=0,1,2,\dots,n-1$ as our set of
+observations. We define a classification function $G(x)$ which produces a prediction taking one or the other of the two values 
+$\{-1,1\}$.
+
+The error rate of the training sample is then
+
+!bt
+\[
+\mathrm{\overline{err}}=\frac{1}{n} \sum_{i=0}^{n-1} I(y_i\ne G(x_i)). 
+\]
+!et
+
+The iterative procedure starts with defining a weak classifier whose
+error rate is barely better than random guessing.  The iterative
+procedure in boosting is to sequentially apply a  weak
+classification algorithm to repeatedly modified versions of the data
+producing a sequence of weak classifiers $G_m(x)$.
+
+Here we will express our  function $f(x)$ in terms of $G(x)$. That is
+!bt
+\[
+f_M(x) = \sum_{i=1}^M \beta_m b(x;\gamma_m),
+\]
+!et
+will be a function of 
+!bt
+\[
+G_M(x) = \mathrm{sign} \sum_{i=1}^M \alpha_m G_m(x).
 \]
 !et
 
 
-Since we don't have any equalities the matrix $\bm{A}$ is set to zero
-The following code solves the equations for us
+
+!split
+===== Adaptive Boosting, AdaBoost =====
+
+In our iterative procedure we define thus
+!bt
+\[
+f_m(x) = f_{m-1}(x)+\beta_mG_m(x).
+\]
+!et
+
+The simplest possible cost function which leads (also simple from a computational point of view) to the AdaBoost algorithm is the
+exponential cost/loss function defined as
+!bt
+\[
+C(\bm{y},\bm{f}) = \sum_{i=0}^{n-1}\exp{(-y_i(f_{m-1}(x_i)+\beta G(x_i))}.
+\]
+!et
+
+We optimize $\beta$ and $G$ for each value of $m=1:M$ as we did in the regression case.
+This is normally done in two steps. Let us however first rewrite the cost function as 
+
+!bt
+\[
+C(\bm{y},\bm{f}) = \sum_{i=0}^{n-1}w_i^{m}\exp{(-y_i\beta G(x_i))},
+\]
+!et
+where we have defined $w_i^m= \exp{(-y_if_{m-1}(x_i))}$.
+
+!split
+===== Building up AdaBoost =====
+
+First, for any $\beta > 0$, we optimize $G$ by setting
+!bt
+\[
+G_m(x) = \mathrm{sign} \sum_{i=0}^{n-1} w_i^m I(y_i \ne G_(x_i)),
+\]
+!et
+which is the classifier that minimizes the weighted error rate in predicting $y$.
+
+We can do this by rewriting
+!bt
+\[
+\exp{-(\beta)}\sum_{y_i=G(x_i)}w_i^m+\exp{(\beta)}\sum_{y_i\ne G(x_i)}w_i^m,
+\]
+!et
+which can be rewritten as
+!bt
+\[
+(\exp{(\beta)}-\exp{-(\beta)})\sum_{i=0}^{n-1}w_i^mI(y_i\ne G(x_i))+\exp{(-\beta)}\sum_{i=0}^{n-1}w_i^m=0,
+\]
+!et
+which leads to
+!bt
+\[
+\beta_m = \frac{1}{2}\log{\frac{1-\mathrm{\overline{err}}}{\mathrm{\overline{err}}}},
+\]
+!et
+where we have redefined the error as 
+!bt
+\[
+\mathrm{\overline{err}}_m=\frac{1}{n}\frac{\sum_{i=0}^{n-1}w_i^mI(y_i\ne G(x_i)}{\sum_{i=0}^{n-1}w_i^m},
+\]
+!et
+which leads to an update of
+!bt
+\[
+f_m(x) = f_{m-1}(x) +\beta_m G_m(x).
+\]
+!et
+This leads to the new weights
+!bt
+\[
+w_i^{m+1} = w_i^m \exp{(-y_i\beta_m G_m(x_i))}
+\]
+!et
+
+
+!split
+===== Adaptive boosting: AdaBoost, Basic Algorithm =====
+
+The algorithm here is rather straightforward. Assume that our weak
+classifier is a decision tree and we consider a binary set of outputs
+with $y_i \in \{-1,1\}$ and $i=0,1,2,\dots,n-1$ as our set of
+observations. Our design matrix is given in terms of the
+feature/predictor vectors
+$\bm{X}=[\bm{x}_0\bm{x}_1\dots\bm{x}_{p-1}]$. Finally, we define also a
+classifier determined by our data via a function $G(x)$. This function tells us how well we are able to classify our outputs/targets $\bm{y}$. 
+
+We have already defined the misclassification error $\mathrm{err}$ as
+!bt
+\[
+\mathrm{err}=\frac{1}{n}\sum_{i=0}^{n-1}I(y_i\ne G(x_i)),
+\]
+!et
+where the function $I()$ is one if we misclassify and zero if we classify correctly. 
+
+!split
+===== Basic Steps of AdaBoost =====
+
+With the above definitions we are now ready to set up the algorithm for AdaBoost.
+The basic idea is to set up weights which will be used to scale the correctly classified and the misclassified cases.
+o We start by initializing all weights to $w_i = 1/n$, with $i=0,1,2,\dots n-1$. It is easy to see that we must have $\sum_{i=0}^{n-1}w_i = 1$.
+o We rewrite the misclassification error as 
+!bt
+\[
+\mathrm{\overline{err}}_m=\frac{\sum_{i=0}^{n-1}w_i^m I(y_i\ne G(x_i))}{\sum_{i=0}^{n-1}w_i},
+\]
+!et
+o Then we start looping over all attempts at classifying, namely we start an iterative process for $m=1:M$, where $M$ is the final number of classifications. Our given classifier could for example be a plain decision tree.
+  o Fit then a given classifier to the training set using the weights $w_i$.
+  o Compute then $\mathrm{err}$ and figure out which events are classified properly and which are classified wrongly.
+  o Define a quantity $\alpha_{m} = \log{(1-\mathrm{\overline{err}}_m)/\mathrm{\overline{err}}_m}$
+  o Set the new weights to $w_i = w_i\times \exp{(\alpha_m I(y_i\ne G(x_i)}$.
+o Compute the new classifier $G(x)= \sum_{i=0}^{n-1}\alpha_m I(y_i\ne G(x_i)$.
+
+For the iterations with $m \le 2$ the weights are modified
+individually at each steps. The observations which were misclassified
+at iteration $m-1$ have a weight which is larger than those which were
+classified properly. As this proceeds, the observations which were
+difficult to classifiy correctly are given a larger influence. Each
+new classification step $m$ is then forced to concentrate on those
+observations that are missed in the previous iterations.
+
+
+
+!split
+===== AdaBoost Examples =====
+
+Using _Scikit-Learn_ it is easy to apply the adaptive boosting algorithm, as done here.
+
 !bc pycod
-# Import the necessary packages
-import numpy
-from cvxopt import matrix
-from cvxopt import solvers
-P = matrix(numpy.diag([1,0]), tc='d')
-q = matrix(numpy.array([3,4]), tc='d')
-G = matrix(numpy.array([[-1,0],[0,-1],[-1,-3],[2,5],[3,4]]), tc='d')
-h = matrix(numpy.array([0,0,-15,100,80]), tc='d')
-# Construct the QP, invoke solver
-sol = solvers.qp(P,q,G,h)
-# Extract optimal value and solution
-sol['x'] 
-sol['primal objective']
-print(sol['x'] )
+from sklearn.ensemble import AdaBoostClassifier
+
+ada_clf = AdaBoostClassifier(
+    DecisionTreeClassifier(max_depth=1), n_estimators=200,
+    algorithm="SAMME.R", learning_rate=0.5, random_state=42)
+ada_clf.fit(X_train, y_train)
+
+from sklearn.ensemble import AdaBoostClassifier
+
+ada_clf = AdaBoostClassifier(
+    DecisionTreeClassifier(max_depth=1), n_estimators=200,
+    algorithm="SAMME.R", learning_rate=0.5, random_state=42)
+ada_clf.fit(X_train_scaled, y_train)
+y_pred = ada_clf.predict(X_test_scaled)
+skplt.metrics.plot_confusion_matrix(y_test, y_pred, normalize=True)
+plt.show()
+y_probas = ada_clf.predict_proba(X_test_scaled)
+skplt.metrics.plot_roc(y_test, y_probas)
+plt.show()
+skplt.metrics.plot_cumulative_gain(y_test, y_probas)
+plt.show()
 !ec
 
-!split
-===== Back to the more realistic cases =====
 
-We are now ready to return to our setup of the optmization problem for a more realistic case. Introducing the _slack_ parameter $C$ we have
-!bt
-\[
-\frac{1}{2} \bm{\lambda}^T\begin{bmatrix} y_1y_1K(\bm{x}_1,\bm{x}_1) & y_1y_2K(\bm{x}_1,\bm{x}_2) & \dots & \dots & y_1y_nK(\bm{x}_1,\bm{x}_n) \\
-y_2y_1K(\bm{x}_2,\bm{x}_1) & y_2y_2K(\bm{x}_2,\bm{x}_2) & \dots & \dots & y_1y_nK(\bm{x}_2,\bm{x}_n) \\
-\dots & \dots & \dots & \dots & \dots \\
-\dots & \dots & \dots & \dots & \dots \\
-y_ny_1K(\bm{x}_n,\bm{x}_1) & y_ny_2K(\bm{x}_n\bm{x}_2) & \dots & \dots & y_ny_nK(\bm{x}_n,\bm{x}_n) \\
-\end{bmatrix}\bm{\lambda}-\mathbb{I}\bm{\lambda}, 
-\]
-!et
-subject to $\bm{y}^T\bm{\lambda}=0$. Here we defined the vectors $\bm{\lambda} =[\lambda_1,\lambda_2,\dots,\lambda_n]$ and 
-$\bm{y}=[y_1,y_2,\dots,y_n]$. 
-With  the slack constants this leads to the additional constraint $0\leq \lambda_i \leq C$.
-
-Using the _CVXOPT_ library, the matrix $P$ would then be defined by the above matrix while the KKT conditions would all be collected by the matrix $G$.
-
-
-
-!split
-===== Support vector machines for regression =====
-
-Material will be added here.