From 2ae3cab1264546207c71a8aaa8f494734aea1e89 Mon Sep 17 00:00:00 2001 From: mhjensen Date: Mon, 11 Nov 2019 04:43:33 +0100 Subject: [PATCH] correcting typos --- .../html/._DecisionTrees-bs000.html | 10 ++++----- .../html/._DecisionTrees-bs001.html | 8 +++---- .../html/._DecisionTrees-bs002.html | 8 +++---- .../html/._DecisionTrees-bs003.html | 8 +++---- .../html/._DecisionTrees-bs004.html | 8 +++---- .../html/._DecisionTrees-bs005.html | 8 +++---- .../html/._DecisionTrees-bs006.html | 8 +++---- .../html/._DecisionTrees-bs007.html | 8 +++---- .../html/._DecisionTrees-bs008.html | 8 +++---- .../html/._DecisionTrees-bs009.html | 8 +++---- .../html/._DecisionTrees-bs010.html | 8 +++---- .../html/._DecisionTrees-bs011.html | 8 +++---- .../html/._DecisionTrees-bs012.html | 8 +++---- .../html/._DecisionTrees-bs013.html | 8 +++---- .../html/._DecisionTrees-bs014.html | 8 +++---- .../html/._DecisionTrees-bs015.html | 8 +++---- .../html/._DecisionTrees-bs016.html | 8 +++---- .../html/._DecisionTrees-bs017.html | 8 +++---- .../html/._DecisionTrees-bs018.html | 8 +++---- .../html/._DecisionTrees-bs019.html | 8 +++---- .../html/._DecisionTrees-bs020.html | 8 +++---- .../html/._DecisionTrees-bs021.html | 8 +++---- .../html/._DecisionTrees-bs022.html | 8 +++---- .../html/._DecisionTrees-bs023.html | 8 +++---- .../html/._DecisionTrees-bs024.html | 8 +++---- .../html/._DecisionTrees-bs025.html | 8 +++---- .../html/._DecisionTrees-bs026.html | 8 +++---- .../html/._DecisionTrees-bs027.html | 8 +++---- .../html/._DecisionTrees-bs028.html | 8 +++---- .../html/._DecisionTrees-bs029.html | 8 +++---- .../html/._DecisionTrees-bs030.html | 8 +++---- .../html/._DecisionTrees-bs031.html | 8 +++---- .../html/._DecisionTrees-bs032.html | 8 +++---- .../html/._DecisionTrees-bs033.html | 8 +++---- .../html/._DecisionTrees-bs034.html | 8 +++---- .../html/._DecisionTrees-bs035.html | 8 +++---- .../html/._DecisionTrees-bs036.html | 8 +++---- .../html/._DecisionTrees-bs037.html | 8 +++---- .../html/._DecisionTrees-bs038.html | 8 +++---- .../html/._DecisionTrees-bs039.html | 8 +++---- .../html/._DecisionTrees-bs040.html | 8 +++---- .../html/._DecisionTrees-bs041.html | 8 +++---- .../html/._DecisionTrees-bs042.html | 8 +++---- .../html/._DecisionTrees-bs043.html | 8 +++---- .../html/._DecisionTrees-bs044.html | 8 +++---- .../html/._DecisionTrees-bs045.html | 8 +++---- .../html/._DecisionTrees-bs046.html | 8 +++---- .../html/._DecisionTrees-bs047.html | 10 ++++----- .../html/._DecisionTrees-bs048.html | 16 +++++++------- .../html/._DecisionTrees-bs049.html | 10 ++++----- .../html/._DecisionTrees-bs050.html | 8 +++---- .../html/._DecisionTrees-bs051.html | 8 +++---- .../html/._DecisionTrees-bs052.html | 8 +++---- .../html/._DecisionTrees-bs053.html | 8 +++---- .../html/._DecisionTrees-bs054.html | 8 +++---- .../html/._DecisionTrees-bs055.html | 10 ++++----- .../html/._DecisionTrees-bs056.html | 8 +++---- .../html/._DecisionTrees-bs057.html | 8 +++---- .../html/._DecisionTrees-bs058.html | 8 +++---- .../html/._DecisionTrees-bs059.html | 8 +++---- .../html/._DecisionTrees-bs060.html | 8 +++---- .../DecisionTrees/html/DecisionTrees-bs.html | 10 ++++----- .../html/DecisionTrees-reveal.html | 16 +++++++------- .../html/DecisionTrees-solarized.html | 20 +++++++++--------- doc/pub/DecisionTrees/html/DecisionTrees.html | 20 +++++++++--------- .../DecisionTrees/ipynb/DecisionTrees.ipynb | 16 +++++++------- .../ipynb/ipynb-DecisionTrees-src.tar.gz | Bin 294061 -> 294061 bytes .../pdf/DecisionTrees-minted.pdf | Bin 542453 -> 542561 bytes doc/src/DecisionTrees/DecisionTrees.do.txt | 14 ++++++------ 69 files changed, 300 insertions(+), 300 deletions(-) diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs000.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs000.html index 39a7c6c43..b1cd969e8 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs000.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs000.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • @@ -285,7 +285,7 @@ MathJax.Hub.Config({
    [2] Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University

    -

    Nov 10, 2019

    +

    Nov 11, 2019


    diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs001.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs001.html index 020ba0890..c64a99e8a 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs001.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs001.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({

  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs002.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs002.html index f39b144d4..0ff176617 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs002.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs002.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs003.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs003.html index 615f518a3..711d3db2c 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs003.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs003.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs004.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs004.html index d10bc9026..51de6dd3a 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs004.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs004.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs005.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs005.html index 6110977b0..d9017f99d 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs005.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs005.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs006.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs006.html index 08579e547..54ecd74cb 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs006.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs006.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs007.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs007.html index 973d9cfd9..6008fa889 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs007.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs007.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs008.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs008.html index 3f4594277..b98b9369d 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs008.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs008.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs009.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs009.html index 2704094d8..382b83c55 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs009.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs009.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs010.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs010.html index 88f46e5ae..bbff5560a 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs010.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs010.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs011.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs011.html index c70fa55d6..bb9e2857c 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs011.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs011.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs012.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs012.html index 21388431c..4d264f91c 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs012.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs012.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs013.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs013.html index af177e3f3..a33ead00d 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs013.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs013.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs014.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs014.html index 3559b2d81..50a90ff1b 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs014.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs014.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs015.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs015.html index 85fa0c3e7..7aad9b445 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs015.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs015.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs016.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs016.html index 4bb186d4f..6b522116e 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs016.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs016.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs017.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs017.html index bd7b75001..ce6404038 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs017.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs017.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs018.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs018.html index 5c166ac71..e52a30779 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs018.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs018.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs019.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs019.html index a96f1adc7..7e8d938c9 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs019.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs019.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs020.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs020.html index 0e40e84e6..76fd563dc 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs020.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs020.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs021.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs021.html index f41b68e3f..645a3f990 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs021.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs021.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs022.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs022.html index 0ef377ff9..bf5f2f8c3 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs022.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs022.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs023.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs023.html index 9c22332ff..d11ef26e5 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs023.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs023.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs024.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs024.html index 66e3f6c03..6e205d2a7 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs024.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs024.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs025.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs025.html index 7d8cd178a..c425838c8 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs025.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs025.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs026.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs026.html index 33eb58512..6f0a90299 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs026.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs026.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs027.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs027.html index 24477fabf..f490442a4 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs027.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs027.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs028.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs028.html index db67c522c..38d4171c1 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs028.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs028.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs029.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs029.html index 8b4f0cb2d..e92deff5f 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs029.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs029.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs030.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs030.html index 7cdf1a908..ed9f87b49 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs030.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs030.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs031.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs031.html index 743fe7e25..f5e04b9ef 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs031.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs031.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs032.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs032.html index 42a5b09a5..0aef23b71 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs032.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs032.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs033.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs033.html index 38eb671aa..4bd6b7b6c 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs033.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs033.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs034.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs034.html index 223ffdebf..e3645d6c5 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs034.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs034.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs035.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs035.html index 0faebab83..23d3432a4 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs035.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs035.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs036.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs036.html index c63bad3ce..051eea3d5 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs036.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs036.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs037.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs037.html index dadb9d6b0..8013401f5 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs037.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs037.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs038.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs038.html index 2fbf3909c..5a69776bd 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs038.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs038.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs039.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs039.html index 43b55b06b..8829cc78c 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs039.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs039.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs040.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs040.html index 4cb7a033c..125faf41b 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs040.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs040.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs041.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs041.html index 403714f96..13eb76736 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs041.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs041.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs042.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs042.html index 134430144..6253f401d 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs042.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs042.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs043.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs043.html index 52938a941..2e5aec99c 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs043.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs043.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs044.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs044.html index 0a0c842a0..66cd83100 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs044.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs044.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs045.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs045.html index cf134e215..260c9ddd7 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs045.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs045.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs046.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs046.html index ac4baa3f5..47ca444c3 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs046.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs046.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs047.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs047.html index c0336091e..48d23ca4a 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs047.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs047.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • @@ -278,7 +278,7 @@ The way we proceed is as follows (here we specialize to the squared-error cost f
    1. minimize \( \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2 \) wrt \( \gamma \) and \( \beta \)
      2. -
    2. This gives the optimial values \( \beta_m \) and \( \gamma_m \)
      3. +
    3. This gives the optimal values \( \beta_m \) and \( \gamma_m \)
    4. Determine then the new values \( f_m(x)=f_{m-1}(x) +\beta_m b(x;\gamma_m) \)
    diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs048.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs048.html index 2f0910c8f..ada510754 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs048.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs048.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • @@ -266,7 +266,7 @@ MathJax.Hub.Config({ -

    Squared Error Exampe and Iterative Fitting

    +

    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. @@ -275,7 +275,7 @@ To better understand what happens, let us develop the steps for the iterative fi For simplicity we assume also that our functions \( b(x;\gamma)=1+\gamma x \).

    -This means that for every iteration, we need to optmize +This means that for every iteration, we need to optimize $$ (\beta_m,\gamma_m) \mathrm{argmin}_{\beta,\lambda}\hspace{0.1cm} \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2=\sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta(1+\gamma x_i))^2. @@ -293,7 +293,7 @@ $$ \frac{\partial {\cal C}}{\partial \gamma} =-2\sum_{i}\beta x_i(y_i-\beta(1+\gamma x_i))=0. $$ -We can then rewrite these equations as (defining \( w_i=1+\gamma x_i) \) +We can then rewrite these equations as (defining \( \boldsymbol{w}=\boldsymbol{e}+\gamma x_i) \) with \( \boldsymbol{e} \) being the unit vector) $$ \gamma \boldsymbol{w}^T(\boldsymbol{y}-\beta\gamma \boldsymbol{w})=0, $$ @@ -304,7 +304,7 @@ $$ $$

    -which leads \( \gamma =(\boldsymbol{x}^T\boldsymbol{y}-\beta\boldsymbol{x}^T\em{e})/(\beta\boldsymbol{x}^T\boldsymbol{x}) \). Inserting +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 \( \gamma \).

    diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs049.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs049.html index 3c9f43bb2..d067271a0 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs049.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs049.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({

  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • @@ -266,7 +266,7 @@ MathJax.Hub.Config({ -

    Iterative Fitting, Classification, AdaBoost

    +

    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 diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs050.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs050.html index c1ce98823..210eb2e34 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs050.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs050.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({

  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs051.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs051.html index 368345bff..c2411b8f1 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs051.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs051.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs052.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs052.html index 2e71bf81d..b7c7ee3c9 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs052.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs052.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs053.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs053.html index bd612b14e..4d9388966 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs053.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs053.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs054.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs054.html index 45e976c1e..22955d15d 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs054.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs054.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs055.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs055.html index 835ee26e1..b59026eb8 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs055.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs055.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • @@ -269,7 +269,7 @@ MathJax.Hub.Config({

    Gradient boosting: Basics

    -Gradient boosting is again a similar technique to Adapative boosting, +Gradient boosting is again a similar technique to Adaptive boosting, it combines so-called weak classifiers or regressors into a strong method via a series of iterations. diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs056.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs056.html index 14623a2c8..4a6f767cb 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs056.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs056.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({

  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs057.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs057.html index e35744ad7..f4d69394d 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs057.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs057.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs058.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs058.html index 2cda0d1a6..9f20d329f 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs058.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs058.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs059.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs059.html index 02d92efc6..9093acc0b 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs059.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs059.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/._DecisionTrees-bs060.html b/doc/pub/DecisionTrees/html/._DecisionTrees-bs060.html index 5af631131..b63aec67f 100644 --- a/doc/pub/DecisionTrees/html/._DecisionTrees-bs060.html +++ b/doc/pub/DecisionTrees/html/._DecisionTrees-bs060.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • diff --git a/doc/pub/DecisionTrees/html/DecisionTrees-bs.html b/doc/pub/DecisionTrees/html/DecisionTrees-bs.html index 39a7c6c43..b1cd969e8 100644 --- a/doc/pub/DecisionTrees/html/DecisionTrees-bs.html +++ b/doc/pub/DecisionTrees/html/DecisionTrees-bs.html @@ -124,11 +124,11 @@ Automatically generated HTML file from DocOnce source 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -237,8 +237,8 @@ MathJax.Hub.Config({
  • Boosting, a Bird'e Eye
  • What is boosting? Additive Modelling/Iterative Fitting
  • Iterative Fitting, Regression and Squared-error Cost Function
    • -
  • Squared Error Exampe and Iterative Fitting
    • -
  • Iterative Fitting, Classification, AdaBoost
    • +
  • Squared-Error Example and Iterative Fitting
    • +
  • Iterative Fitting, Classification and AdaBoost
  • Adaptive Boosting, AdaBoost
  • Building up AdaBoost
  • Adaptive boosting: AdaBoost, Basic Algorithm
    • @@ -285,7 +285,7 @@ MathJax.Hub.Config({
    [2] Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University

    -

    Nov 10, 2019

    +

    Nov 11, 2019


    diff --git a/doc/pub/DecisionTrees/html/DecisionTrees-reveal.html b/doc/pub/DecisionTrees/html/DecisionTrees-reveal.html index 5c102282c..59ee06a8b 100644 --- a/doc/pub/DecisionTrees/html/DecisionTrees-reveal.html +++ b/doc/pub/DecisionTrees/html/DecisionTrees-reveal.html @@ -148,7 +148,7 @@ MathJax.Hub.Config({

    [2] Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University

     
    -

    Nov 10, 2019

    +

    Nov 11, 2019


    @@ -2051,7 +2051,7 @@ The way we proceed is as follows (here we specialize to the squared-error cost f

    1. minimize \( \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2 \) wrt \( \gamma \) and \( \beta \)
      2. -

    2. This gives the optimial values \( \beta_m \) and \( \gamma_m \)
      3. +

    3. This gives the optimal values \( \beta_m \) and \( \gamma_m \)

    4. Determine then the new values \( f_m(x)=f_{m-1}(x) +\beta_m b(x;\gamma_m) \)

    @@ -2065,7 +2065,7 @@ at the internal nodes, and the predictions at the terminal nodes.

    -

    Squared Error Exampe and Iterative Fitting

    +

    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. @@ -2074,7 +2074,7 @@ To better understand what happens, let us develop the steps for the iterative fi For simplicity we assume also that our functions \( b(x;\gamma)=1+\gamma x \).

    -This means that for every iteration, we need to optmize +This means that for every iteration, we need to optimize

     
    $$ @@ -2098,7 +2098,7 @@ $$ $$

     
    -We can then rewrite these equations as (defining \( w_i=1+\gamma x_i) \) +We can then rewrite these equations as (defining \( \boldsymbol{w}=\boldsymbol{e}+\gamma x_i) \) with \( \boldsymbol{e} \) being the unit vector)

     
    $$ \gamma \boldsymbol{w}^T(\boldsymbol{y}-\beta\gamma \boldsymbol{w})=0, @@ -2113,7 +2113,7 @@ $$

     

    -which leads \( \gamma =(\boldsymbol{x}^T\boldsymbol{y}-\beta\boldsymbol{x}^T\em{e})/(\beta\boldsymbol{x}^T\boldsymbol{x}) \). Inserting +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 \( \gamma \).

    @@ -2123,7 +2123,7 @@ The solution to these two equations gives us in turn \( \beta_1 \) and \( \gamma

    -

    Iterative Fitting, Classification, AdaBoost

    +

    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 @@ -2358,7 +2358,7 @@ plt.show()

    Gradient boosting: Basics

    -Gradient boosting is again a similar technique to Adapative boosting, +Gradient boosting is again a similar technique to Adaptive boosting, it combines so-called weak classifiers or regressors into a strong method via a series of iterations. diff --git a/doc/pub/DecisionTrees/html/DecisionTrees-solarized.html b/doc/pub/DecisionTrees/html/DecisionTrees-solarized.html index acfe40fa5..27309b5be 100644 --- a/doc/pub/DecisionTrees/html/DecisionTrees-solarized.html +++ b/doc/pub/DecisionTrees/html/DecisionTrees-solarized.html @@ -144,11 +144,11 @@ div { text-align: justify; text-justify: inter-word; } 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -214,7 +214,7 @@ MathJax.Hub.Config({

    [2] Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University

    -

    Nov 10, 2019

    +

    Nov 11, 2019












    @@ -2054,7 +2054,7 @@ The way we proceed is as follows (here we specialize to the squared-error cost f

    1. minimize \( \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2 \) wrt \( \gamma \) and \( \beta \)
      2. -
    2. This gives the optimial values \( \beta_m \) and \( \gamma_m \)
      3. +
    3. This gives the optimal values \( \beta_m \) and \( \gamma_m \)
    4. Determine then the new values \( f_m(x)=f_{m-1}(x) +\beta_m b(x;\gamma_m) \)
    @@ -2067,7 +2067,7 @@ at the internal nodes, and the predictions at the terminal nodes.











    -

    Squared Error Exampe and Iterative Fitting

    +

    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. @@ -2076,7 +2076,7 @@ To better understand what happens, let us develop the steps for the iterative fi For simplicity we assume also that our functions \( b(x;\gamma)=1+\gamma x \).

    -This means that for every iteration, we need to optmize +This means that for every iteration, we need to optimize $$ (\beta_m,\gamma_m) \mathrm{argmin}_{\beta,\lambda}\hspace{0.1cm} \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2=\sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta(1+\gamma x_i))^2. @@ -2094,7 +2094,7 @@ $$ \frac{\partial {\cal C}}{\partial \gamma} =-2\sum_{i}\beta x_i(y_i-\beta(1+\gamma x_i))=0. $$ -We can then rewrite these equations as (defining \( w_i=1+\gamma x_i) \) +We can then rewrite these equations as (defining \( \boldsymbol{w}=\boldsymbol{e}+\gamma x_i) \) with \( \boldsymbol{e} \) being the unit vector) $$ \gamma \boldsymbol{w}^T(\boldsymbol{y}-\beta\gamma \boldsymbol{w})=0, $$ @@ -2105,7 +2105,7 @@ $$ $$

    -which leads \( \gamma =(\boldsymbol{x}^T\boldsymbol{y}-\beta\boldsymbol{x}^T\em{e})/(\beta\boldsymbol{x}^T\boldsymbol{x}) \). Inserting +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 \( \gamma \).

    @@ -2115,7 +2115,7 @@ The solution to these two equations gives us in turn \( \beta_1 \) and \( \gamma











    -

    Iterative Fitting, Classification, AdaBoost

    +

    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 @@ -2319,7 +2319,7 @@ plt.show()

    Gradient boosting: Basics

    -Gradient boosting is again a similar technique to Adapative boosting, +Gradient boosting is again a similar technique to Adaptive boosting, it combines so-called weak classifiers or regressors into a strong method via a series of iterations. diff --git a/doc/pub/DecisionTrees/html/DecisionTrees.html b/doc/pub/DecisionTrees/html/DecisionTrees.html index 1c1b06da5..f596aab80 100644 --- a/doc/pub/DecisionTrees/html/DecisionTrees.html +++ b/doc/pub/DecisionTrees/html/DecisionTrees.html @@ -149,11 +149,11 @@ div { text-align: justify; text-justify: inter-word; } 2, None, '___sec46'), - ('Squared Error Exampe and Iterative Fitting', + ('Squared-Error Example and Iterative Fitting', 2, None, '___sec47'), - ('Iterative Fitting, Classification, AdaBoost', + ('Iterative Fitting, Classification and AdaBoost', 2, None, '___sec48'), @@ -219,7 +219,7 @@ MathJax.Hub.Config({

    [2] Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University

    -

    Nov 10, 2019

    +

    Nov 11, 2019












    @@ -2059,7 +2059,7 @@ The way we proceed is as follows (here we specialize to the squared-error cost f

    1. minimize \( \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2 \) wrt \( \gamma \) and \( \beta \)
      2. -
    2. This gives the optimial values \( \beta_m \) and \( \gamma_m \)
      3. +
    3. This gives the optimal values \( \beta_m \) and \( \gamma_m \)
    4. Determine then the new values \( f_m(x)=f_{m-1}(x) +\beta_m b(x;\gamma_m) \)
    @@ -2072,7 +2072,7 @@ at the internal nodes, and the predictions at the terminal nodes.











    -

    Squared Error Exampe and Iterative Fitting

    +

    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. @@ -2081,7 +2081,7 @@ To better understand what happens, let us develop the steps for the iterative fi For simplicity we assume also that our functions \( b(x;\gamma)=1+\gamma x \).

    -This means that for every iteration, we need to optmize +This means that for every iteration, we need to optimize $$ (\beta_m,\gamma_m) \mathrm{argmin}_{\beta,\lambda}\hspace{0.1cm} \sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta b(x;\gamma))^2=\sum_{i=0}^{n-1}(y_i-f_{m-1}(x_i)-\beta(1+\gamma x_i))^2. @@ -2099,7 +2099,7 @@ $$ \frac{\partial {\cal C}}{\partial \gamma} =-2\sum_{i}\beta x_i(y_i-\beta(1+\gamma x_i))=0. $$ -We can then rewrite these equations as (defining \( w_i=1+\gamma x_i) \) +We can then rewrite these equations as (defining \( \boldsymbol{w}=\boldsymbol{e}+\gamma x_i) \) with \( \boldsymbol{e} \) being the unit vector) $$ \gamma \boldsymbol{w}^T(\boldsymbol{y}-\beta\gamma \boldsymbol{w})=0, $$ @@ -2110,7 +2110,7 @@ $$ $$

    -which leads \( \gamma =(\boldsymbol{x}^T\boldsymbol{y}-\beta\boldsymbol{x}^T\em{e})/(\beta\boldsymbol{x}^T\boldsymbol{x}) \). Inserting +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 \( \gamma \).

    @@ -2120,7 +2120,7 @@ The solution to these two equations gives us in turn \( \beta_1 \) and \( \gamma











    -

    Iterative Fitting, Classification, AdaBoost

    +

    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 @@ -2324,7 +2324,7 @@ plt.show()

    Gradient boosting: Basics

    -Gradient boosting is again a similar technique to Adapative boosting, +Gradient boosting is again a similar technique to Adaptive boosting, it combines so-called weak classifiers or regressors into a strong method via a series of iterations. diff --git a/doc/pub/DecisionTrees/ipynb/DecisionTrees.ipynb b/doc/pub/DecisionTrees/ipynb/DecisionTrees.ipynb index a1baa3fae..4bfc45466 100644 --- a/doc/pub/DecisionTrees/ipynb/DecisionTrees.ipynb +++ b/doc/pub/DecisionTrees/ipynb/DecisionTrees.ipynb @@ -10,7 +10,7 @@ " \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 10, 2019**\n", + "Date: **Nov 11, 2019**\n", "\n", "Copyright 1999-2019, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license\n", "\n", @@ -2107,7 +2107,7 @@ "\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 optimial values $\\beta_m$ and $\\gamma_m$\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", @@ -2117,13 +2117,13 @@ "at the internal nodes, and the predictions at the terminal nodes.\n", "\n", "\n", - "## Squared Error Exampe and Iterative Fitting\n", + "## 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, we need to optmize" + "This means that for every iteration, we need to optimize" ] }, { @@ -2172,7 +2172,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "We can then rewrite these equations as (defining $w_i=1+\\gamma x_i)$" + "We can then rewrite these equations as (defining $\\boldsymbol{w}=\\boldsymbol{e}+\\gamma x_i)$ with $\\boldsymbol{e}$ being the unit vector)" ] }, { @@ -2204,7 +2204,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "which leads $\\gamma =(\\boldsymbol{x}^T\\boldsymbol{y}-\\beta\\boldsymbol{x}^T\\em{e})/(\\beta\\boldsymbol{x}^T\\boldsymbol{x})$. Inserting\n", + "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 $\\gamma$.\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", @@ -2212,7 +2212,7 @@ "\n", "\n", "\n", - "## Iterative Fitting, Classification, AdaBoost\n", + "## 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 any of the two values \n", @@ -2557,7 +2557,7 @@ "source": [ "## Gradient boosting: Basics\n", "\n", - "Gradient boosting is again a similar technique to Adapative boosting,\n", + "Gradient boosting is again a similar technique to Adaptive boosting,\n", "it combines so-called weak classifiers or regressors into a strong\n", "method via a series of iterations.\n", "\n", diff --git a/doc/pub/DecisionTrees/ipynb/ipynb-DecisionTrees-src.tar.gz b/doc/pub/DecisionTrees/ipynb/ipynb-DecisionTrees-src.tar.gz index 2a8f44d75cb205f2c21babd75ba5699cea908703..c31a23f489cca7dc963b66c73a1d55e84cf0036a 100644 GIT binary patch delta 29 lcmZ4cQ*iB1K{okr4u*3#8rfRe7+cwxTG^Plvau|$0RX5m3Ecny delta 29 lcmZ4cQ*iB1K{okr4u<_~jcl!KjIC@;t!&I&*;tm>005aW2>bv5 diff --git a/doc/pub/DecisionTrees/pdf/DecisionTrees-minted.pdf b/doc/pub/DecisionTrees/pdf/DecisionTrees-minted.pdf index bfe73b849079c96ba2037aa1e7520dbfc4fb2397..93861c875986aece54409e4b07aa87ab6d099137 100644 GIT binary patch delta 38818 zcmV(`K-0hVjUnNVA+R|De{=J%#;BC^H!P)(uk}^OT^@_Lxarc zcCrmEb|HHN-tIVztvj8=-g)7B8b%@>{TYKhp)Glq);ped)ikxAe>16SfB%#pZj~)H z%OvSac#)WRj*w%5+Aw$3^ykEA|H@ufGR+~_b^+>&tX!n$s_ltt8(R_OQJl-P z0jGd|K>#eEUvPnDKDh$B&w4PT30^f^I}KBi(AcRS+lxcAf%|Cj$yRP2Z8+h}34yB9 zQ`DWmx~xJnSoFF>e>fn#9BGch=Z%4-^Dw!9uywwS&P1#Vz})(bs<9yeph{5@SB*#Ohz_Qbst|`fZ5FkAVwo;o>#j$j6EeVNaqEN}a#ay- zuz^xHc^Ogz@&>O1kfZ|;&%3;U53_eN%>Atr!xUInw*a$gf2M2^*($d=ab?wY>mo{D zMU@`zcTp@2)Nj&1oQf!(<}M^Rn9hA;$L?jB2RkQ9Ce;Z3vw@z7;}Z0cVMV1BsRM+@ zV`tgo&)a~2+AFY`xCVsW!nI7p;@VLX;R;$@orN~Q!etzM!iHQOwozV&HXy`jltGUC zW|-Y&4&R>ee*}m>p1K*RMHc3tTm7RqdCr6}q<@bn?m8?f(xg+>HbnX0gfXZR7Egij z0pqD94yY#3;?P^jCwm3Keu!Jh$}+7X^8~IZ8Y%w6#hHlYH)0XaWVtk3BVznQHQ|P7 zXg>B-C#WQ>bOa~bH{w-AHyH)p5l>_$nN;DLSHW0Pe*%NC8Tx=Kn9Mz9BG+~_{Sr|C zLsi$q8A5eBg?WfxH(V}D4gl~tJGOM2{)Cf_0>4lOycZh6a3)K$#70Qz-ly?iWojw) zs}ze7F0_Gf8fdJ67mCguze{XzOl;Pma6M74*yBzQ`#b)G2D8+0&r(@VD% z@%!M^e``g>wB{q^#_tZ@@p5VFU?A>lpV4)VFPz9qw=qPAzYQqx3gg+0E+ZK?xEovF zfKIeamV%Y0*q-&UNziY>IxCpU)OIHrfk6+b@NK}?oW9mX>4IgsywxOL#BE9tg=R{d zb)T>^IUekwI-bgz|@B9YWo;HXyQ!k$SDo&~5Y<+cXz3`RHMLKZ@Zonv2k9wUJO}cp z@UHSsmqn;!P}Ju!ER3Lc{CbY2b-r8)FyAK`_!LKaOkqy|cL{Al=y_n$_O*=;h~0%= z9ZCC7RXE7^{EJ$rH_Nkt5?18lI@Vo{e=xb1>qdCg`C^1CSVeL4m*5F75V-nfPvigG zK*~op5`fa0RK{6d^_ZkGl^#R)(&GmONIVA?cxue&i@({2HxU7=r0M4@&uV-N0l{g>cNVa1#`-*2e;<-C z1zcf^^g^=$Ep!7tepXM_k|!G8`WfHkI0dB^9zrGDPZqlAn&PfhEB?J;@Hk2K8H;~h ze9m|t7&iyQv%)NYlyGCr9^bnDiKniA%mBUlV+NtBMf#fsMA1tTI1ni4w2`JCKG67n zC&9UHY#(Rz{{GX@e;CeZJJzlL9=`xs6wAV2OOgrbBsQ&6*VZU=;SWQkZsu(wl5y3J@(}5 z_wXV$sixYO@nDe`w`7s5VzF4~RFUQ4dQmR^eo}_tzn-0Z`;O{GMJmPA;_PZslB!bp zRxap*3PUty3nevSTrSSmix11c^rr>qVp(teZ}U*EZ&%G5|8}w6^zCKS`M2Fo;~#!& zf3Mfu_R_yx-EA&T>2lX@!w~nUOfN&rJACPz(6S8;R@>{g-?bP1t!u8krti_6oc@0H zGZYQ#a!#}|3ne6B#%ueN8-YWUdKWaT`vx-JjG~Pf&G99O+jV$?8moHOBopCIcc(Z6 zfBOzn_Lif@5Ls1FF7tPbtgHFq{}IhBgnGK_pSt*x>uj2!aT_ zk5Z@%ZK{PLmFC6|LZw+0ln9LF8m=>j&09y5wgjB+@LB&CLF3>S$j zCXPW#S@KCVk(Gk;3BNz@Z=A{e4WshQOgyiUGDHcHb9HCGe@mFGTeq7g z*z?>1;OD(jOt++F zgbA?+c&17ia0^{J|FTDUrnN%0hXFf`zn|5Sts+Wde`K|g0?Si|8T^oiGQ;^HR$$UISn)X{6>heVo8M%g5lxFLVqx@npu4u+{P71a2}^0a|f|5D89| zw&E;mg>b=S26p0&1oN>GRl*+PvH1#G5^WT)AF2l%yw9kFR)V1we9tmZJ+otKf>C|mk`O616m@P(Re}y zIMu1jl17vuHn`^%h*RJ&t}UCzq9SVbT_&8`JcLcd6x;(GC}#E&p{MQrlVkRHQUgB8 zqjY`PP)Ke{GZ{Q*fAUx(ra;vAq|?kgjDnPM@g?c=@R%S(W(jjm1<^-L#&kur1>Crw z3&M4I;uR(jt)Y2qXvVDFo`iy$)tx*SoHL$b`IMdV+iTiNyDhs3`fh>v6aF?`CUfaBsC`)d(f_xr@;efA5~N*aWm0I7@ayyabEH zZcQYCB8WWrj~a5!#UOD>`n~b8LmG;w_NZt$d^H*2D$g?aVx>XOkoI|dOoplm1%ppI zygw}fM^BlgK`@`#hMX3HSAJuLHGq{f%P^%k0*-+3IF0>uNXoZ&t56N-TTqbC{@*^y z{Nx&TyTEgNf3x7J{$HDraBk0c^}x5Hq;%k~S@Oy>q8^LjI%^eVtA_29ez%PdtzJ+| zI1wcaFyV^Q>5$2?Rz&GNZe?&19ww;*GM9P^BNI9z70`6wE(fM|>%sCi9*0csEHjjH zrHdkJ^**nd9V=0cX4!&Mj1ncqvnj(wFN)a|3QnY6e}9fMByfb@ry`e)a~Aa?PouYM$J}jX8pyj|Zad$&zPxPXwraWFf;<`oX>7FXH~xS3(OAeY+I9O!3W+K( zbo;bt=(hQNYX#qAa}U9RLKdy9`q;&{PI7Y*Hr}Mx!;fz(N+l@;JKR=Na>8Y-*-L9;lUkQli9`~HqLxNh5)9t$cQ6E{Y#B>C6;T&N&CZ{@~Zt4tH7!z~M#VQ$6T` za3g~FJrg>J>pBg`+$Mx3)KJgdvRZVtX=*(%e+9(!(cLVpcGe$}8#XS=;)CO51LMH& z)poVo+9B@4KP+JiN%1gxcy=C!w~Mx3c^4i^iuU&=i0nV!)hvE}{)Z6841{3x*9LSL{3UYrK$al@8%v@p*08lxvK)~s-lzGTvjCRS_k+9PXi+G<{S@*b^)4ci~U{TtAjah*6ZZvrZ)4x1Ki471xL8e+cGF zwMes>nH7-pD`4O~lM>1+oTE)@aZ_5W9W~HIVCv7=&A$#X{H#rvtKd^xp;e2hF!nfc zlIh8w$+8%E{J8R2=W8!MuGCMbX7c!PrJ0thGVYw{Jc0FsGUZCe1vsM(yS|P!X=Q8Z z2mTF=wzwELKG^!eU62&JJ>s5Rf0k;N*LUD~N6S(Yl+~o8US(Xz?i&5*c~q=0QvzF9 zq(eViyh5m8lSw{2CP)icpq&buDPBeWj14$RH(Okg91Pl9aV!fMcy0mK7wqf~CWhzE z|295r+W5CWnr`bm4&VFjI`rtkx3}w*Ifm6`fGt)9lHrOEcWe9i+^0faFn{PrL{Gek zPuhGaWLbi;8*X*U>@7s~a^z9#eEnMPPkZ0`2o?g_ce#=VTz_c&YnLkw1Ra0#!{?uY z450~pHHph#`v8GH@dSZAk?*7-HLi>q8=^(b|R6g;zh7(0)a$Ga1cOc&>qDJ@KE_?e4;W1NTnbbrhPVCpWydx4U*-ul&oWQz@7A%3exq zvS@AVdof3vrDQ6%)#hqh@-=H&vB+6F+RJO2As|vXm{5C5Ud&nYZ^IpivG%u_Q~@{5 zy|*o_sH_$9Cd3Lxw~?NfnjL3<;%g=bsq#KCc!};`Gt7!t3_`&G?goEa#sKyT9(`#o zcP1qff3lS1nNB%Jf1_F@vllLG0on7-f1euZOQUpTzIeQI=iOMD8icg+MOnJsUip89 zGAZ5P|LOhvzG_RCH+48i^9&l@F6}u$e{t9QTe!?#)LpIoH+#s6T*h*7VU|B#jpL2I zgG%VsHBJBKSS>_TSCW5?=Q#r#wmLCA|D^`qhYLUBc?0-opB!S#Wq-SB!-0p7LtR^+ zuZCvhS##9PpKff!J=US&ZgMytJ~!RQgN3x#@dq;fo5|pFce<#mTWEQqQX$LM0sJ9s|uclD?jit4O-+&#I0?xAus z+OB$P-GAeqy=wSp-QNy(#|y#959ha7qcYcqxC{JJFz)USs~G?v9{r9yE8WF*xT8Cg zo6tIhyO+vi9UkxY+f`TZwz-4ObHn*!>B#3u_2Ejhte@W2K%;t9cewQ*ZtU5e-aeOY zyWjNBkL|8Mox>&1cKa)mHSW0w-8xEA9Z-_ONq?kNfC_ttTR(*bQ+Q65>?izEh$68K z%eCVDoR`EVB9~^EaGltJ=N#S#C`2STQG^C3AxHw9y~iJw;TvVy7pdd%oDmuAz7(=Q zXw2Fq=Y{B}I%+Mch7A;r>2>o^w+`)_s@%3?3f%Ri^c>rGDtcDjZNGH$o&)CDRz2r* zJAY5dr>+J{Zt6d)1VWZQA2E|_DrQL}0u*pR$1G%S6w{tcpg5L*0^>K^Bz97{0D%c6 zU0*GUR3X*EO-(LzdQ*04Q>1U%+;8e_Vox9_C$K5aITOK^mV#v0zZ6tBDT4))ps;DJ z-7LQuqrql?LmFVF=&>h*IWnhoU?Ac*;eQb+id<=XB%_U+_us_HVHO8Qq9mupPpcE~ zXs+RJ5AS`EKp3zhF%o~25a$c$ABi&|eqsvwBQX+fgEACRs3?-`q4;q0yEHhZQw$MN zMdgJwXHZ3rs4~)B95#u$YF0R2(vo3#3PlmP1d0SnqB4@jc{FY z2{}JY$kkg3xjIM4)#sFP{x&kM-cZJM!X}6CWKJXrg`1>2958jV5K|sbEHKW)dDT5H z6=!`!r%;@~o(V0uw(3NHi(UYqSbx|2KNwqBcZq__UpAe4-0d6pFt}Il^>*-|sq?A( z=PHP8VoZAIjG|9vV(52?%LnF+yL%-i`EV}?Y;poFb+q#Y(+;QMEGO_!wB}_74&YBy zdISd@WlgNx3_rzbA}|FPlOw?S)slm!W#hc^Hy<3)J~*Nfr8N+135b$x%zp*`ZF>}S zLC`2bK^7Ri;-1fS_uzcrrfM+`gjK6<@P?03BXw>JIhLO2CMM^4Fi044CQxjwZ+~Ri zM*9?Om~(L1!9l+T_q;0yFE!j5?qD_ik^m!X;uilcp(_N$&cp8C;)HkK-MstwuE!Sz znPml8aHD}TWU<=3`}x;mv44g?e*|O0j9xt3sT;I|MS*JWaq-i;fBQT^9P#*oR!WvypvK83ea_gSK>99`)_zlkuxPQZNTedV+S+38fH$&!NEKx zht37aJw{JntKKA)1Lo?N!K{N~i@>E6TconIRo}&h>I&8J#060^D>lAi?5_zqk=>tA zDKww6@`M4?6Mvw90>wo=?1y@Vr1Rm%y)O&3eX%DMYhZ3*E>eehP`JvHZHD@ zt0@m9zWCx?Dmnw{RG7>wtPww-i|MH>l7xH6Fc&f-r+@78$muz*{_DMp8GZJW1c-Pl z-rTcKB#=P2gC6#|3n%>wrJpmAJyzwq#YdUWG8{s1jNwhZ*Oaq;I0A+6f{=LCd(58j zDT)sOxpUd0rPB2(f`ZN5Fe8-Q_mv+u1ZP@j1voQt&Ms z$s-Gq8#a?IjFg}2Aqaz;UM+!BqJ=;8jm@1>k$<(j$GtxTyxBeg%7<_UhtQyboinz@ z1_r|T243%nl&gKkMiBp4_nYpU?I~_m=rpnTe1X}JCp!jaynXQWb3MBIL|%4?4^VjM z(Q!obI7H}6j>AE%3rgb6(I&W{V4#2)f&*A=>!cAA!4Xa2nP^)*yQZV+<*BLK*7MVx z2!BvGK>=<^wAW0%K}=W0!edep6OLjMZat&2C@VlAuz1bO93Tdhl{x}6L4?CPtZs_%uo`ZAxF2rR6r-s#R{0r z`7r{lOUO?r86YHMVB5YEkoryF-^^>V^UjNk%pQk=ZkG=iInObxDM>{*8^jd_iow^Ylyi_U&T{T7ML; zvg)mK*tz<5T(~a3K|M{fUE8IwZEJ9>@7efg<;%vdt(0afu%7k0@-^hr*N*FL+4y?# z>i7&W`qb=hAFGXf=?j>2u+Z8y!D4l~TkSSgvH=rBnI7sUudn++U<9**Nz_GAcN1y+ zD0z~WydXhvoo>812;OCuzwOE{pQ329+deB9pBFt@~Bwt1qSd z0S|dyuX>GYvaW9@!71CJd4I=v$`;;K;X@P#!2pn3W#!|(SF>?tvp2QCPLhv_H^I8$ zMLAe5rbY`-2{V_}7DJ*1*K<&$%q6gS3QGDu3cMC&;sy6JL<6BcB3d4M6+T48W)WN$ zgsAan4b9+8SOSHx$gF6I(`OPcww$co_dB8UFKE=SI?AkT zAPq5~+|}o7UDtn6S8H4s12gAa*FDmINH!CXD&3!0py%f}Pkxe23F^3PvmRA`PL-Y~dw(}4+m7bS@?I@KtPZA3hiah+KeBQq7#Qkx|EjXuzp=y0HEyyARg8jeEV{+rSFH5zAew(SPUi#-E+pA%_u#7;)&J z(&y!RnC^dwV|)rqW-j8iA>-HyAq=Fk@RaEMH+OCIabH>E+12HvD~*AR^YjGg9NwZY zQ9x9W>P>he_^+f| zpU4FB;4BJ1SWlF7*_H0eVG*|9uLcX_|8kdKRq-C-^;pC~$)HSP#||4yvx!=sHB5NT zS52S({oHxwQM6+H=b{T{Y(ENiyA9t$xvb1E2Q1$I0FaVk6qiw`0u#528w6Pu0XLWN zpaLj=-CA3Z+qe~epI_le4-jj`n@D95V3RZnve*_aG6f1G`=GHk9wD~nNOmUaukUwA zN)~0>bIERj_GLz*c{%sUtL}N^p@#Kp5z(NCRaBTmhr?GHp!Gu zxKNX;#pL7kQ>xMef7s>kSHJd)Cm&}jO{Z^vtKH>H2r=E=mK%DK)4~lbtFM>bOx>uM z&R0d-R-aNK%f_#$-F&R{15HtFs?E~9?c4sbpK*7|^|Y*(x3u%FC4zo>%gulTt}}1| z3N7h(envjYl3WWzGjzfPpUhYybLnPPZdM=(kqEpGCM6;FGj22vqK#Cd@V^(;^|7db z(j?FElup=$nKV(km}HVAnG%z$HI}`u*W3Nh@~E949ob^28ovRTDlwXe8zU}K!1E-_O+b$b50>2_feCegGn;y|#NI|8>4!nj${ed_e#(;nQqs;Ff0j4m zFZ*IcHn)SBgQD*di_5cwWxd6HN|}flU6u$foY8r&TE}#pCIVhJnDqkk(`LG=;R#IV zmILNi3N|!X5Ru9Q5ou~!d~8pn7Dck<;LR;*i@@4hu{h{)_j(G*TWRik?JwT=@-=hh^GIHI^nI z38kLwwC0@YzK;;65F^;(xF}6z+dWC&BlJil9C@$2j+B!{=t%<~I;**NQFk>$Jb<0v z*1kp9Q3(C=URygP;-kngO4tMG9VKfm5}svvkcmulH;E@PPaEXon77$~9E42QA(@Eu zgh&*>SA6m*;tNfR#Fq|)VpD}zDE1$axXt^LOrlSB$;b$3Q# zA{RgrDkL$gury&@MBLz2i3 zz^$|KlnsP0*#zK!2ZLVyL1iSSRqK+{s%VydRef7*x~IA=Jx08xl;xG&RnlG!Hml+1 zQE??vzF~A&!5J|ny-PKH*}=@)js!Q^&YhR#qJ4Hqpr?!q5FO6%7^|<_vVrO7-gHE2 z0(Mg~q&fA*oz3^FoqMRokD*oz4X$g^+3T@PPGsGqk~GL=)+W zh;VHhI{2n;=GNnAwDu&{o4IAxnk1aBbmJYd-lXi%SYttb@5KVpmZ0G9p{wvUOwpMg zyq~aCXK>YjW`w%7P5HH|_ukkJp*-;d!$B;`6>x-2f}ktW!Y7uOmx4_f#VfK3P}|bk zT20@4E7sdp*`ANRJ^d0kibc*_-XmC+~Y=#%Vo|}apbM{wYp0k_J)r0LrA6Scinf}! z=S{q$O60DRwp4XzAgt0E%uHQYBA)XD&rG|(iQkK5VU3R2piCoewXP@x*s)#d9C=fH z*_VEQsCM&S(kY>g@$0oa$(?G?BcXEc6p0D*dQH=>t4-OSLN!7gFzTChZ9SVKVUvr< zEcI!>_)KqX6tc;xlOrj%C@cDcuMZY9Ru`MxXJB~T4-))UBbVueJtFMU#lXV zTsBqdOwk^66SWkP!BV*CEI4r5fM016&GR_+dagNn9e1+l~q5*zkcV)oQhGeN40O7BykQwX-)ncugYu?w=dz&%1n7 zPZH3Bc|NZlg!)XjoqDT*HqkgT>&~{(M4gHzuZDr7 zi)3A7dsr+Xz?~YOtv>9xTkpn{@Qm_*=!2f+0n_y(g`4CQLr=Y8&g6E@6P5Gu;MeD2 zDK}{%QGW+(PSkAuIYy!$ADwe^&p8OJO!K32sLKEx80|82klf-9ZAn$i#3&kt8;rx{_WAq@4%hD&bFH!PUS;R&NP_oyC81*FMv&C^vsqy~{vYk5Qsy9dTo zu#_WN^$>ZGI`QUeO;;Zi;`VTVdtLeqK=;~V?zIOSasD=l#v@P4ps6tNdgSK1_D52M_+F`V*2}kcnr=tlHSj+IPkO6`={Uq`)keaF(CX ziRLIL2CfP_%rp9@m3(*~`BA=b;!;o^iOP&F1^GD*A`q*`bn<_!^_`7d--B!JEVvof zGM~`Jv!BZ&biq2|D1%9V&m%zt$_4#I9(`(|{ay3b;DZ|KeEFiM?+ZqIdcF(OedV(W zx-xwgOtnMN-uBf%*C2E^(!)DBgc;D!au}aC<>-an`%Oi~Q}yq0%qfBO*Hq5M!U!N& znvWHqg?6S*ptP7_6V}T|O8bLu>mGT9$%M2w6C^N|>2p_@Z>dv%TgUDuMGCR>uDp(W zvSNcTENndUVeNVBA{|6`v;P`PT6EFg-<3TZe#SLo5P(Aq9rAbL)3Z^Er!R>ZidvqU zE?xOMTq%5yK*9ny<7=iE!xRidV)j01R0)xemyYDCRV*am{4{O;R|8E?FnW5ON_NNFTBmu z=`X|+g?c=gcuJy>f#V!vBWDJFz-_)a9r$A=nDplA;y+=*QPG#-xCRrqw>$*?1}HT! z3NK7$ZfA68G9WiLH#iC}Ol59obZ9dmFbXeBWo~D5XdpN?H8hv;paLj=jdul5n_br} zTBJpbYj7!&;7}ZjOOXOa8$tp>5(w^ATnZGI0xhm>(H1B!#i3B#U5i60^3zwo_x=8R zXYQTMB+uGs?S1w-XYI^9Y)nt}c;&2+mQW=m!i5*a2b2IPXlkm;gMa`aP>>G@6u@U= z(}%m*L;tgY&t?EcIm3~E2#J4D6i`sG%Uw(n>~hD{L?QrcuJ!F9}qTidwYZS?mefD^(60Evr>^8AqwkaK{d z;1DnZpb2)dfjZpn2m#vz^pFrZ)W!3^N^nZrxVSh<@bkO7yYqp69h~`)C~Ijh9)LUC z#Ri}Yb%vtcpjLq2q5)c92k770`0&{P`ZjRqKUegSFc)_)3JSOj*ux=Eg!5gFE5Ztj z0^Ds6&{Nd_Xgfj?f08x+B=7+KIvoIr5A>gMe`WtJ1V{Xl42D3E4vt`iCmdl7fWhsd z0Bt1=J{J!c9sn4BVfC8`ws%I}<%8Y8aC@-j9pR7O!2l&W9RT=F;$H%tAt<<`i!+}y z-2S&n{@-Qp&iM(#N&)HM07bYsNy*jT{tp7@UyT2Q0Cs@ed;Uea zv#+bm-6&`x?*;<#zo~}MKL=PIX>awv*Hm4=cLO1Rhp@ggH7`hr4=D6!9PX?H_kdbG zg}Xp({M{Rs7I@!8-!% z9|#hE1@OE63yK5y-Tze#2;lel7rZOw`3L+j1rS#h>Q3Sx9=enFH~!;)0SfhiLhz@) zBO#Ap+J1c5c=2A2)}424P(X;F>nQRB*KqX-wcCbHkb>mTW5IV(lX5L%>rp=!Np}Q_ zZtLufD1dh63qEO45)@O_jo2{D6&{~qQmM;-9cK+!fpzufX5$xv2rFxfAN9ID?kBR} zxF~s?4%fyrx0EH>>nqO7ZGPZz)b8>6NyArgioS2rqX{m++v&m(pmrw+S58PcpZqQ#ad0Xw+W}1HQ4XwmqTx6~`HW zbkg`Tkkit+|7lXW`eyp%uWnC@qscTya;gow=qHqlK!4j#Tu-p8eC-vU!L%7_+gH#~ zzB^a_r1r>LG~(L7MD_W_q~E$vU;kdIm_)Fn`@M*o+ZAz6r2il?$g!46@)VVEMLaSI z*0LP?VnyC<-3`$@{8)FEaFV~oVr%>dVQe|U=8g5f}$1+^W(Fsef@y_$vh!XUi z8hy#D)Dp$IdQjS|2x3(B;g|va!U$B=15v$ITzkgoF?ljF;l01cE_|l zt`4=pi1s$P@h#>6MVegvX5P&PnK`W0_{2NAQSP9gX!iBNcM%43Rq`U!!9l;aq7pmg zV(eFE2034obMcbg(K$-l8PQ1%G%gz20^j96HV3GvQ(1*X_GHF&yZBc7J7GMJkZ<}m zG2ukl#7dHpD}i2lOpKwK<{_$oZQk@%>3AH_qH==1td1EhVh&&QE z7rV9~$-~%>ffQ*6$dxCZg2RFagAAZ_Ied(0rp3fz>%a|*egj-{XNZit%l|{j;Pm?nc`4% zV*V(>{A;o^Fg+$WaBa>|-ic|BUb<7kjYp>D(<0ONSfvs2%P)s{8{+QCG;xJ$98ViO zQ|eSkdv{xf?O|AdGh@k}n0VWMYo8o(!^tbf#}TBIn<(jb(Y<>&%OH;Q!+n*Of(CkL^Lmh9ol+L2i98xMH%W%YqY*VTs&JS;P<$e;nNFvP>f00wP;D%8rTIX`AIoJ@SX6NY&&s3mXgOp z^liQ^$BRz2W*UP+Z>S+`AHl=Xugu&}RB)=n@JI$G%3cpsDLL%SyUU zl`&jQYod85A5FZPaoJEJ4cDsHekiyP^X`d%Z|4>Acy!g>Kz2Oh3dH%mV}TpCy&gbQ zih0clvJ1zAiHG=9L=OjPCQ^WMy<$O|(! zMFLoE_eeW09OhOWz@VTsrJdIgITy-*VT4UFZZykUj1#hLfP0t%952P&hz{WTl8RslFvDbH+e8uQm%mkiN!$Ao zT<0_Uy8LZFgT<-sIJ|io>eOMgTSYOeqjC@(drZ>CudtI#Z>wB zF-xJPEL@)~HzQ9pNLP1?Wx<;g_QuE<{X{)?M5xV(j9l!ne(QB$;h=53wEKt02AtfC zVd5mg?NfC@&la`pDB74~tty0nU@7=E*Q!|^K<{eL%ZRQ7zi7T}hAyrAh4bcJlZQF? zKXAzQ$^9CSZ@OP}+&R>*RPdo<9{t*Kik&~bhdDj##GKeli1!VHnO9$h7}it{5!(DAWoO&W?cD*nXdz|e(d<#?lk)cJ zU>e-K8dzA*mWRhv-zG1a(+|#OF&X3a+SIY{NWY_k{Tf#~-$%&J^{3d(37QBk*Y2c4 z=r(04*`*E@{m5!K9{4hUhT(Bs6tRlKP2=*I&SPDL_0-gL z4(|XgQq-(uHL3X;cQ2Ff+|x*ZTcz7K5q-JP<;i>+QRDH~eb2amM|C`(48$D>i%aiq z4h!~vQ}5yJrIozo+9}i&NS0^^&BeA;A^-BrTke6QL!AxRpn!A{ut9|z^t2%_pa2{B(nAoOX z>h!9M-zu(l1{9Aadp`KiT*T-AILW`akRtSY{RXfAAI`9U{YbCy&2~GxrcRT5tnpdn z_Y}pHDk7UHLmt0JVGS%VrO11lUefGP=U`;wK=scCU6%}h;v%%1E0@V=cbFA_ve`f# z#Tw(i4qy71R2E4Nr9Cr_3|_`bSCJ7HClPq20W8`vM-43X@z|W{J+Ixgj+Bh`ntY&N zz#=OBJ<)`k8N#R)-+uLJK?P8H=FW@1Lr4`#&=n#5QEFX(;b6V3&05erm;3p*5Dt&e zOmLC>${FW>oSe?G=@sMU%j{9aOIMK@uEd65j-nMO)BE)g_y(Lw);{vko!|x=>vn363&gP{sxMyOg%0y-*P-OmE%pT-Q)GBhRv*ggn=zWO=Vln=zirdlE@ zlC}p=Ke`z?y~|01>GAScFa*8~S0E3p`3lDx4PBUj1SXyLW}XF1!wG{}QKErGW7$$k zO8mV6T9Q)yuB}VrL{!fvF{CWorKxK8-&votFU{R6f^o9wR#z@M3EXCT3%5#4GugIY z$;jxg=cBXGrVW7G;+w^YS(*i#@=ORBZ?cww{(2>B$;_|b_enYY?ONUXuTd-d&G=;*^i9m_IX%W*no?BVUsNr7gmtYjgj()grd%Ax%v4QL0{$?^-*IIc`p9fX{|OFV>(Fgv%c4s1c! z8=>cM2j8dhHmd8qhDT;bJ`FMdl#hFet%`GEW=3@7?L)pxD3;lP=cnUFqpQ(kw-Mfd zr+ZM`%RH6-NIg~I!i}vnSPZl z*COIUHDf5gm!oRshTP4!wX1r6$%aL!+!<{AR7dWe-$TA;gP($*o{eW5J`@9oza87e z2}IM!$J2pX19F+;E3Hdxj?%0XMFz)zReO>xvYD`qZ&p54QLIF5fUO&1dLbTM4b;YF z;K#7o{Ae*_J>SC(!x(>6blR4CY7_B4jP*dAYNC-eKhwSvhJUP2u$LQFa{3VqDtK8( zKk=)XC-4-)pFu2P3ufgamA7&o34A)8}mIeocE+)a$T)t4WCZ5qP@ zFULUYfSICi_=A~H+KfI4)47qrMb-k(^K)Te;weD(6)k9xw_LMb4>fIdQ2t(LIWjP_uchKxdd5v$AaG6EL7iPoncPcX4fnx(FLZ-QHCRNUV(CCno)MIx* z&Voe~H-q%;?Q#WH+eW>Lx9w;_Q!9pht;kGpW(8be*p-oV_cNjH$6&oY7ds#SqZi^y zKXJW>mEqNxkL>I8(d)Hm-R?PwfwXOlQQPN2B0~o6UUoJ*kM)GDtD8j_=!E_n4JGw` zD$o8TsR&d@zf9;|(!o!Eub$SosgT;GZ)?Ff%+nhdbALu5lXlk1es&uZEjaeu3ymb5 z5=1$n!1TSu67h2zd{GQ+-0fAXeD|eWN&UV0P(ECRGdpxK7?VZ?n+g+OIG9%*N~m#v zS>o9T3H(PJQ4-L2zPfXGM# zmld>$p=yR}GvJ=tfe5Z&8+#&SxA*45P1>bd{h#~6&6+qXyD&DqCl@r=ip96@CSxgcJOHd|j}2t{vpVX;=Kli06vggoBT>n54*>g-l0J$egw6sTr$ZeV^8}xGhvJ#vM1m!+Cu#?Zb{}BuAp$n7yMQCRl)_qA3 zmuo@*MnJj0OclLRIQrT%%F_{%D|>O8>iIZcj`{qdMyGttbr;A12e)$p4bcXt(2ovH zwO?g!jGqUtGKsi5e}U{zuMt(@*mGn%MZ2{pQq0zAA7?@0{p&Vbu_lZJGvByn)SX`W z^Dv(AWGxIbW&McHTdrAb3?9PQb`iN{Uu=3AgyxYfSo>T?E#0=M#^S+3hRIb!e^RJY zZ|muODT|t$Upje2_Il3=T_RqSwyko{n-WD9DD+lkbCmDzf5*6McMj#35Dt>ajYLXd z6`%bYrv5Z!Y|&oOA8Bhi=`~}2_bw@gO0>Agls@_{?kL7v&mMR{`*LD10qcu@-;bGb z(Y@>oP9pwadL`c-9i`GL!j5nv%8&)D1BGpsmL&6ZiyN5L0yDvWDmo>Y&mZE>+{pEN z;PT^wVqFfPf3^{a5Zkzf%7+{|hl7T2rM*mYlkG$&L!!lH-lycQd*TWQ-19}P%qzVb z83kU{jfKMIAb;tCn|`$SbTunguqVt|Q&O^*h6ADr#nOBs(ECA#Rym?iQa_E10S|1j zjos9DCzc+%+147=MuC&kkuojUtAn;u;}Ma3*a3t+eB)_%e`PxIsiU*0!NkTK;+R0wqdeKl5SIMihG8^)(Z@Y0aY{PMADlIn2Gkaolf*v~ z(g;i=WrTp*G=!n+fUO5bHJ&Z?Qry%WS%AmMTQM33`Q-LOWv9K67{6be^WnZo|5My* zfsZ9r70Gc8CVDK0uSevl8lhQQF8+?8?gJ%_fAN&1VjF<#lLV|5^=csdXiCflKu}6O zP5q}}OXFl7r>Fkq zg9mQ26`Zs`MkpL#U_@Bq;}fx-a0o~iTD2XWzEFva)&A;1U-#lWad&-miQKKxjl*%0 zeeDC5<}5AZ9OoYlFNAy@f{Eey$s@7`5o@w1kZDlj>++PJwrde=)l6 z2nBVzVe_4q;oja<1I^L|X} z-sKDCX+rPtg^>nRYA+feY6aTfe+l_2Xj@ep9hpaqXt~JzqoK)wG}o>TQg8!J*|w5$ z+CD7rN7V&Cdk`|1TcEd_%YrF0YcR)8>PtBbIr$4#|agzP7XD?I2*#=@j$8Fg#3 zXGC5WO-Qm;AF8KST~eQ|l@3Q^~eEranjzySJ|_?zc*Rvk?2%@D;?)LJlL7 zVFHnuxYUr-Xiyr6{Ol8Dq1qv!yw^oy`|JZcTLMwW78~1xRbc94q`6jmJ!goJV8~&@ zs^@KsLjd1%0R@K|zqaQ-Vw3PC6^vG39 z8~=JASYW+UWj5o^`z{vrp{Xi`&F2W2pa>6S2r^X3@1tm;5MiRu*I#JOiPM`Z&G6r9 zuE_Jf)o?`TBG4FW8Sp%nli0y{wLR-CD?P}X@#1^p6Nl<>;8hT1skVr?mk%|mXSlv(4>vV))bVlZiIUC;q_c6X;_ zip#>!_9jp5W3gtq_PLbZw(np0ldP`wU?;mxVs7H_tr*|7_Um}F$+viojY{*#rDfpe zVHvPo7A3j1%jQ@QB`33YBU2@3gtHXIynH`=we6i6Oxd$QGZiWNe}=jwp&@=K7ptUFsyC5P zNkk%eY|!Si{6%xXT?{RoX{XR{iWqy%=q;ektr066(ptzZnC+xrzn2|EyS4Kg&E!p4 zLz+R~AgY&0WGS5co`5=@!cR-z$_9o#x`rq#Fj{)j$PXN0nKo`-k@`X#T>Y;7l8gDG zG5uB^^H_P;fADjg@=%iGry1s$)q|D|3zR&F^$XCh$f+{=Go$_5obZfeL;?S|4TjIk zZ!R^7bf{IN51`xOT9;l7eUc$ZJLA$R#j45P&YcpJU-1Ks$OLL`;>nsQ4xKR~HMWI+ z3Nqj>ksTXNk;am}QqtmbPg11&&U$FPRavr5jlsD)e^{*Ys3L*dZ>LmOX&DVyfI0TT z`?=RCuTaNK+(lxQ@^Kx0##V1mTX|KhpRH|XS)1aGI%qIw)sbENSn-Z- z^{?bs)*3$p$-bk0dWcO%naPkF;AGoOXoQ1}T}VvwrqvXu>oacuHu-d3Vi3i&DQ-Mj zDhbYge>W#o3c@6maC3<=96U&ObS z6-?uzPc^zAp#*XL(=x^u+Keis+U%Q8X+o z;yroMD{QmR7?<~&2S$!$XO! z@fUrQ#{LruEFePBFN>_J6V^FcG`2Uz7X*?i(!}ecU+FPIF>+i_GHS9r_^CB7z5t)G ze{pCtcgjyH?V{yMHc%VsvNmJF4*9b(T=8G1XL7o%PIqL9{N&|c<2)eRb0%ho*Rrz( zBr2w8M|N+lm0Q|Rhh~h_a!O2ieTuvW@B)>H*`w+>a@w72G_u9>;u0PE4bv1&NR@i| zUXLz|uaYY^TU-hlIK*$t3tqytP}`d3f4OSl?O@HhQU8yPx6xbP*o~*QK3U7A!n2xl zJnEkhO)x#f$H$iaw9_A!7s`e$(egZMjx9hAkpnG^Y>%5+51$wSg?3rrzDarCk!1nJ zObO7B5y{wjrd?=_X=W<@n)~!o)q&oeT~=R{ft6E~AWQE1$Sj1Uxxw6ULcG8s$(nvQ>iEH5i~2 zW;IVpCVv-$SOLSXWNEdm0#SAXdXzoxBOJ;Ai~H=F{RErJrF+k9P9S3V4Su9w(z-W= z2U&-$NAzjc%@I19gllV^rKUDae|Xn%g$A&d=Q+^#iLi<$D|Jj!A=5Q4I~%`4IYG)6 zUM?Q-%|k}co+&^0hyIhA<<@{fHW|mk@T92wm%vjAwVf^uKdHHiIMiwFV@Qi|;|29D z-n(msA94a=6eFRIQxClKNLMBZG!0r^+>)`n8xx}u|Xq;*p-ZJ5LKs7*~laPs{qWA^>sM-uIyTSanu z=a(!INfj+ay0g63uy=H%fvB9)hmy36G=4Mr=v1Y;;3E`=Y)PlU`B_CT)4b3`{&LiFHB`_ zXLM*XAUQTMmrh9q76dRkIWm`Fc>*Vwk4Xgxf8CuT;Q-P(ba!`4!vF&eF+&W}DP1Dn zNQZ<9N`s_;f^j-GS9@1Nf_*TlW@UTd$t_ivc!b@aJp?O--wB^VUW z1?1)x188V!g5Zv-aF7ec7Rbw`2X^ps0eJv`+Ie6AT9#__~3C09Md%pab)SbJ>7A5o$2h z0RjcHAzb8PZoVE62S@mQ41O-I`$+fRa@+tlkgYS!+tV2W072~lYTTOK04fup`LD9ssik7=VodhWd(n`T%7;e{Dk@eKu}HV0|w)H<-u&aFN$HFjVFQD9CCV zC<4GnoB(A*eS`Z?127c9-+>dLWq{D%=Rp|WyK5>M$QqdFC<1xzYXATNPr)9Zko&a% z#f=%k4Dg2N z#J3071^l})UQjzkJK>Jte*)aE4nPB93x;}v@12xj|0rD%twcB>dk+aZ3}~XLOkJ~|3m}>>>)1Td-k69 zYX*V-mTAgrsVXVz8*phLW)I4x2}5)T$_@8{|F*u5C##?#1`rYz000s1`#DvF+R4LQ zT@kE3aqiov073K#4)gHk`9B8S84B};`u|^AdkECd{(cqgyxe#Up%8a3f3T{;|M(z8 zIDcdgU^svm0Cop}eQX_h?oDS5fLXn|b8tp9ss|Jxem3UTrMU+n+O?g{vwH>(!R z!xiN6Upk1V62u2=rvri8I{wq-f8_s!l7%|BfB{@U0d8J?A-;b!hWES81u@o$qX}|< zxZUf8{!53LSX*Z(*wYgr!24STMs()C*dvzXo;!d?PuW0T&Vb|pe;B6U#)?o|m>mS_ z0N@i80)RX`K)yJQ236d|$s%Rm$X5%Uh=>FK3nDE3CWD~v{RiH2_4x}ThR63e{LkUmxj*}VA7#9M zru}~|#ot&T?g4WKKY`dG9tr+3LF@nzh>saB;v51Z`7eBbH2}b1AFwUX+%(KqJlrYcePm^j zV)1AdI+G~(el>-ZvDyo?x$Mz4>IVuhzcGNNOG2AMa+phC8Y*HIL204TGGnH2a+lks z=G2-mKkRg6)<8j;LDU3_vY!f!xDCT{G-pGqIvCh?f7A+!Ov(kub6Q@s&;blT?aGU` zbzUa$jocBRq%)XTwR|l^_b#0THj;U`koe3JJ#43*9c@R3-(8YDj|X+j&aqEe7e!Kh zQp0fT>Lgu9jxUwIiDO;Q;oU}wcQk)~*DXQ7QRmK;t-g&56&XA$QC;L3R+YdZMM6HJ zea-!?e?uq!wCg)orae{2lQ{YGYoPBb@>60YMp|aVA=@pjuOkCk>+Lx24Rq5^d7^2u zZbiJ8#Rrw1%ztYKJoVd|4>V?9W`e2d^#$iK#7WWgeNFkbl;}(QX7lR7O7mMe5l3Hk zjzL_Xxl23r#~+PHZ?E0OvhnQ&@n+(m4=R>Pf5u}D+*#PK&dk|18U?cuHa#7`qkDRM z^KceLX`OUJxcmX6U~3->P32}--E>xwFKX8LBTJV|+1sx3$D*Q62Ll;XO8pm)i5bEk zMc?KW3UHvQ?G#I_r@xw-`RPp4HUWGY?>Z@5o=C1Jw)Wn%p24b+r|?>HGJ0}Wc=5F+(jGsHTB^UQNtcKr04V0XuqGK9>Qu$%o=$m7 zU9W9#qfk@LLxuUevVyp5d^(a*S*7O1C*3={)iDh*Rtu|7t!OiWy=j)=e-zKN`**pS z2Ku~Lc6W`Rq2m+2i;c^W56I10M8Vp@-kPfp#E;0F@kdQ*6D#D)tXK9bmFBkZE^qw0 zuBo1#rzed0Q4MXlp5ku9$#69hH$l81BYx#rD))E>onk18)g!tCI#DE&W)57DoQIcY zbH@`jh_V?>p`GcNTxE{5OafV^T{~R&FnH6v^rM6Zc%!4CaOdB94IS zY&Sn%Nic7Hv$9f2ID27m^Mhpfm7+`6H*aaf$g-SnpjxF&Q!JaWf2O-EZ;O03ayF*q z1r3ODy(?#E*e)aZ-1mOSd2M#-`;I9`#~MgHC(oTiJB`z3X%R?|ON%RZYf@44gynrl z8MPNM@Mcrs9`3_S%8QV(8IzF`m-PuEQ{|+ms-eI8nnz7DaFtoLUK>UU@N%WNm}*}$ z(v%qw^(THX_-J|!e^ocmCg6c06^RDVg2~>%%9n~s;rzsrJ&$!4%2@iyCA(^7=5!o6 zyT;25uQ0yG!wpZ(2g`GP_QvUYK-ovcgrDiAC1^hUpdN6MzBX4kYA{#XM13}db5j?J z<=$jU>Ge+fts_$F4Xa#~yeAFC`a}g$PN#S^6<%CE%?+K#fAKS$o8#0n*F2rCfPi%O zTX~j(tV7D=2lNgP6y=g4>IfT(BWWB;3^_q)8Iy9`CR+lc@TWeyRSFr#hOZogD_pS& z9xE9bPaq#nkd>^$vpyFO(ub`_rFzzM1l4Ds3+4`MY*Bu3YX8Dgl&EQUN?|I2(kQHI ztt6T>MCZ$-e>F^c^2E23jds_gEGF0}y^Cd{i~5MUw-|l27nT-qcksjMG+59SoxJzuN1%hwJ zE@?3B%#)(XZJ;K5G4c>rH!@1uoimvN|Lava8fL`e2_>EX+iK$>anrvC#Lq7e&)r$q|dGkD!gC37@)aUpK8Z zX0Q2Qg>sPA@V$-e9_M?w2Ke$&Xqpp(<{Gk8F4Fx$5*-aCU^mL>s=Qk5M2PN-_4zc3YwM>?o7)KPx3aTQWNbc&BoOxG7($YxsgoyM5mZQ)Ci( zy$l6M2}?hQRt7i%mNmLa`^~1ia&BvtU2~yaQf>B$p~xldoFe7w7_~}W93GU|PII*9 z`{HX6U#2_xSsty*totjhBuz0nlp@Mje*qd~?_>N70nLrrfQFaP-Cv+gNcU=KTOY6R z^+R5 zc~I!FO4|bINl(f4Zr#B1n&-u4fBGCN zc2kZGRL^8;0Quah7G&S7zOhG#AI1YOnde84Bdv;sWeU2r2>YV!XIQBiZwfQx2=+c3 zzkLaKaDZAdOaZljnvlH3R-X4(r&(lBtj2`CN;$_&<--4K`_>ROTS@o}PZPG|aFhkU zmGLhW6LYby9-CcRDOFr!MTxJUe`rwDx(H;Dzx&z2=b@$(JeMvw>x|*ONaEM=20WR1 zuw91OO{S4A;Hj* zkw&eq3@LjgGOoEc`%9hjR;m3K^8GIN;<{VbF=kU~2n-prn*EVeUZOy>e}1#8))bC0 z{kp0XoUXzd#z;-w8eWIR5-}8O2|FlcYuZ(@5-|&T{sW)5_tcmW>&Wni&}5e(T7F!x zm`2bPxfZ-58yBmT7_xonXDM_c23fvZXF{%Mk?TaCQBOpBWR}%gKtOxZ`@vQ_Wwcd3 z2Dd=tx>~Eq_p=DL>&C4te>UUt%2w3S4N>DgOZ#PbKJ<8;+*0* zPAUSmhubk{eY^Zc4x}|irEJAK>s44My2R$eN%>=3KQ(~Yoc+0uKeYY&^s4Wzbue@cKt=7TmVY@Zi$8$821J@*s3f|L#jdr(v8zFrtVv5NLHeQ(;c znEt7L5o_d=Dt^l4Ufz=2*qMpvYXBgR@%D=ng@ zm9jQYh3^&8e1smgP8&?g-6iZhZ{2O6@YuA8h&>ms11>!#f4Njr3Bxjn!F>5U?feI1 zhYRe{i3a%D?>53JntVn?cQH(geIaXyvN%10ouV6_;%)NJD}8Lo>O~E;spEu%rrIl9 zJD%%?5N}z7II~7UUHt=>Akz-+5wdvwNOjx~(iv5$-Cui|u-nltY?G7dY+_H>oN@}q zE_B=asmZ#5fBS8N4#eL{GuiWpOwdn+v0m5Fibj>1S6pIePPM@a2epR^O{tBtXu3wI zUA_A66alL70D)Vn(PR$>0pi%6M+&MZM>lo+;le&q1_xK4lu^09Fscjlhcd~?&LV`+@f$1-dL3CRx&Zn;l4OOT20ZLpR|{}j%s1T zZRJLNX*recAtgt2peTdgWN@l3Zz~U6mWbR*35Ibz59wR3QJ&gX+hTt7men9Cmw>oJ zMJ=O`e_avIf>EOF1ECQwcl^$f zhZ#RPRw{-<15j6w(zb1sg-AU4%=N9fGnq${e^&r6NgEIF7N3GJXa*!}^`5C_NY4wQ z(O$;~7z4@AzvBCRG&AmeFXLk*<1&aBaLXZ$45hij#32~8Gz;YcFqA95u477`v%k8`pR37A z?}Xzl)#7#n_wmf4QLD1l2=Z*4Xdu7S$IscI9s%~e?U4j z+2(62mJW(~HNVmzY43na$DvZ%llDaNCM_(QPqBY0+e;KI;~o!^8;MrbD|;>5iml(dN;651Jv;Bi~C31OlVe#HXQ zXB?~>0RZnel;PG6zc(Kh4xB(ze=Mr`WA-cU;#;O084KYa*IT!01KA_{x0NdUWQp7O zme0p}_#-cyWM9Bb=DsvD`p$e$QEyZYoFI3R>0E=gmuO_SLPM)yHx(N0y}Z+{B8ouW z7P<4J4&RwD0gnzPqZnYo6Dl)VLO5%%ay6(2nT4WxLz$%&Hd&7GTJY3ne?Jj>K4g>H zyYH>+TEr4ApR(Sfr%Gw<$51s6g`Ulg1&XH?{&FsrltLj$EGN>XS+{abXQ^djiPe4_ zGhvMLS@tZP@EXe`ELQ@J;Qdo8+Lazh?M6G2qS{pgWi%&OmVMiJJk55-8{7D<%r4|I z4sru0jG*OpS})(?5c z#Vvw_f?M@jib+w{yXwA~F1K2A2tDIjA0tx*%+9&9zrugN;JPG=e?+=NW`Q%=??z3s ze$pBgKrXWIOFM$C!AN0~sQ3fd=#Pfn5mJAC21>naO{clku+~NJDqfV(xbvWMzAd5H z2UPB>Lu!a^CM`78`{W0!-E?E=Q)AB6*3<%?9xNxfmVmiV7`;!i&RGkrtp|IExxqTt?kH*cQKTw$v>Lh zNP>Qaf36p87hBJKewq}1OH7$|5kHl{7%XI7Y7&I8)NXrGC(xJQmyfYUHwNI8R{CwfYqLeBIb>Ijuu;0c*{+yv*k~8NV?eNNABgu9?@nTjyP4&qr0oBXE zx!Uowpzy5=DZ3w0Cq-0fiBa((tEo+XL*L71J7q3~$tj&}jH!pGAmN_e&(Dyua= z_y9WJfBN7E%&O%Z%4 zbtQ4jzf`IekU2+!Ho2z5{Dop?(F4kkUycbC&yl z5cL|%$kp$Mp-yt|u?WB#OXsGD7~{itv`61ALe82m`^6TY^JR)5LC>kaQ}MSZN{B=k ze=bxO716AvCBT-`b@hZWn&h}E+7@kE8Z9$qik`|Zzj=n7U@?-E%0hB8Km?v@c|QU~ zzB7p9(7evAj)KSHZ}RN#(>QQ<(^ybjD7MI3ylhS zs8s)*&2}}CHh#``n}NmJWvm<|mLm%TNKIi{B298i{{^!b36Ag-g9`JWX0D9{e`?|O zPD)9|0aGSM-~*bE%}62FWf|iy(!IPaOc#&wWwREP(~$5gW+N6u!j$> z$JZI>vL9yJV*4{N7xUH7iPW!Kps9lkwH2rjCDQSl32d(Y*eXPaF@mZF>N0?e-Zzqcqc{JkE?k4pM~!&nCVWZTwAxjEbX=O!1-_W2u$|PyU3W2wEv)?84Rbg1Hw!R ze2=hfNm5L=9@r^-?o!=UVrBvkj(;>*?KEf5UJE!qWN>h;$k^MDe>7yAOcp*OfUB{A z46Empvwgk<_SLnDw$$_Zo((-s)^F@Q8@~?1VMifq4o`$N({TEE%q7R(uCE4wGeYyj zhscOL8ugHV-t1~hvwyjkW3ZjF8qKeS|~1 zthuQ$*1^(M+OR`Ze-sDCFs)f8u*mn8~cZtwPn3SwQrmFmn^h!c= z^C9dul_Z9HKNsmqmR;Cp!&s$Y0SfG{@%T+AHlL#vW$cGTz;+NrJq3~~;}a*VEZ!R~ z?1N_V&04Rr^6gOa>2T_rc$3$Y?LiKoHrl;(b;t|ocJOY6e;DUq&VJz~efiN=7Ra4s zw9;Q(kKXXWka23DM#18(n1s{CMn+_3W_cDFdtKV=q;W2K5Urs^X3&v=F@6x2Wb%bF zGg6BRX7w}Z{)JSL{N3YfE0muj5qebFyb;-vl&Iy_yfbXQUr%;)xr!?H3vs2xKU&<< zPy|g8yNU#Ne@f3kY>BZtRjF%ePd;(iRihT=u#^>4vbt~xyH+CC%xSZXGremL$SBR( zF4GWusE;8^`Znd(EM?*(%J5~i@@Z!6YAwIMTl}2GTQk&7``gkEbLl7nCOagZud>hg z+}r7vOPJuXtw|BDFZ0qxG!!i}-q(&u7eI@B&Pz`Ye>_rCqCb^p>PuO3-$<+>cmMi) zjEeMYl*i&zfTFLjpgx_3~ z^NQf@d&C24-Iq5dQJ+Ys^CQ1`CH9Nnic%wm6ta%(Qnl?jfI->*x&8RgQqlUogfOyU zzo@~Rf65qHuB7*GPgKOa8nTy{fz@0eXtGEhIQ;RBn7sEoHLxpKgE;p(-2jj;1Ai2UGO$;k3@%1}UJP$Es295CHMhW#3VJA=B zta@Nz{!?@t=@sGK&=}B|%G)Y!f%Wd zf6~B$4m$tas!_$-gMtWpPrq;14VXvKe(`i49B?<7Na5(hNM6rgQhpBaTz8N4=fcc5 zIv=`#pr>2srm>1H<#9^9Sj!MtVL-w{fm6^LU-8e>^87>)k$8>SEwwT@%`Y1wPT%V|FvEtT0@p|ZONptoVW#b*%Z3B6XV!je@>T< z!zv*PX&Yv~ztWh^%Bv7Fze;h{Pn?y3_HER;k;Ao>ihWq^RWPnqv*jbqgQUq;;(&*J>xrf20h0=+W!* znk8}dKi$6G8G{H{82dK2!p;my(t_v%&zmyT4XzkTv`fz|MB7Q4;=kf%c`Sr6CYl9~ zRm^Ri%~VhAcS<%_fk4a9GJ_g((>cG2OBU*cemkKJ`VIzMh`7@T0d1B1wP|B{QQ$Ihw9p` z2uF?Enc%9RKUAg0gxPHOx+s&1Pu&)tNA+D6H2dDpXTvk3`z*RftEWo~j-JfGYLemY z$Xj!uw=YCzzRJ3OH_2$Skv@&_ZA|xzipkolZ9D#P`OJJz}==K(s^FCMuz(QsnTGK)dC%`Q~> zXhh8BV%#wMh)~2&NTaQHHcF1fzG&BwkPvJkr^}X4yUb;$x3|{lKQ*_FN_sC>E|gg@ ze$1nfHu5EdC`zc3yq8!0DLKM04 z{_OneW=~m^a(2qroYKaOH@M0p(0(?~KT~%H<-fLtUTUdD4j~SToHr{sHU(BWcYG1o zq%Fw9lrSm0^}|@Uhs=pA%EK%3My}tJO--0ze4Ua8D;R(}#pM@G3=W>Mhe;-BVi6xo z6{BhR>Cz28e|glJ>{@@`Qpcg-wVF#S5J5X56Eh!-e_Ek%otaGSmAIE;==YAWMg>P^ zi{f}u_-&j)=o3!{@)t{()iD(sn7n~J;T_5s?tH^o2E^?b-2{R(H0-FNJ#~O8g?;<3 zF!a!$-3JrI*i~O?Gw`31P~rUkyZw_Q!TEakb`@pDfAE!4^EP+l?!eEFO%v64{-oiL zso5wGcrgba?X;iK_jZwRy|$z^{mge}Ar1-Y%s2h^!VX^f$o`T=iVv&OMsLpUxl-I^ zg!%l`HVR(WFJ9CS^SY0P00Yc^6|$!Df;1dSo}jlQ)6BxR<>8GqPA{UqL_RJp9Et0W ztzch$e`(r{#wBKH=os8-(<2Vr=8t)vT~oqgd0g@S<3jID-0IBWEUR2TC&`!via!0t z=HoP6hw7g#UvjBJ?>y?8w{KS89ah~wv}H4Tpjkv;8bwS?%FnpQC%*}56A(W?yWGN> zi;jWje$NYlV|Pi?;x4XsaZ|mvpi4`%KC6RSfA-5Uan+a5Gizv}mr7%lLvA}Y?!u^1)2;bF+0sm3yXZR2(5O3HnsHA3t;MG8+TrFrJ|VYxq`}`A zc>GWkjo$>@98PR--d@r|Rh;tz|Hntv(#DJxRvr~IxU&m{IH-3Nm5U*>xd`PuHR%uJ zf8kZaR~miv71T6~6qZvbN2rM#wh3iA8K(X;M>L0uif-vCBbZm?CK2Di-}S~NitCqM z_`{YJ*GCrpk#fHpUdwj7BmN+e<*QTFz&)KUk5LYe$V2}UYJJu91+kR(ktCw z>-JUh@;r7*3NtF8xpGfGTk1_-rxkHJ8I9Bws*0V|JN2*fwlh2nbm%I3<*!nse&!K( z{AiW1DW5f+b&d@_K62O!U>;{Idf0InerR|w#~SudeQJN8;o9V|zw85`ck9B=+rLek7NE0s&zlBNXr{(&aqc~$Rpx+_pbe@9PWu%C6{T4v zg=EgGiF!;vvzB2b@)-bQoajvxf0v0;A~&>Ea<06(=9!0jZtk|glkv#;^=lkW$}5+d zlZ?z&-@(5;F zyrGfQU;XMWsXlgEZ3AFB+|V9LTTTilj0-J3FJ59-*yg-tYkU4tfrB}bf9Y}DoNY8U z&X1SrPKEqeBdoKDWL`)w7yNU0IWF@j>A==}&&W2$;1q{HC-uYoeZ9GsL=UWU&Cd3s zJ^+WRCeU5pa%cp-ccAPzaB9qXuY<=NMo&EpmXhw8O-by2ot-K0p-rzGF#98<;Z}V{ z<@&&02t$gRNup?`nyfS}f55~9-Ik9}jm`JF)8Y84Ss$K=JC(o#OFTy?DJHQ>&O;ql z5p|~QM>Eze=cun4{BXWsQ`VwiJ+cgLvvoG=&EJ#LT9Ca?FZng<@bLZflL+YGc^1Eo z0&;jsn>JW$#}f^^qT;Zr;vp$#D_vOOb+BhA$MB*kt>}}^B-Zs}f7VKm#UL|GT`SK} zoQ`9ewMI?yX(kn2NuLjLd&D=hvvPq>320c|kXm@o`~ni z^xb5nn>~_M$bNKCL^dFTk?)d}mP16?A2H2)vbQmL6AYO9I>0@W)38lAwE;6eN|#u% zkchb>sLAQ1B6E@#JOo^Ric0%z(CggJc@8-09}bjyMi<@YXa0`u)M> zG_h%znM^V{?JDpZKDY7z0Ow1Rx0m6#1{1d~bp@ibe@Bvi=U4ROOiZ}ze%nw?gf1qn z8B^AX8jp7sdH|^wi8V;j0O+CT>-U_j2GHP5ElRtdjUW=$Rav)_S(#bcpz*rVjq#@O zp)t-jg-^b*C>kFdXA{0tWK4ug+ek>f5 z6R^M;xY}4i5_%@`Bn+4xCQKMGr?6)d9*Sj6YOHanPNlJyJXxb4Jq(cG!Z4_HQO0O& z;K(^NR(6MvP0|bFB#xYOV>3t27%+1OK-Ct*e;SKIV2y2DU4jS5!V?Sov@DBpa8Hf#)8~g6Ho}OLHDFtgYL<*20D@oYcNH286%@6 ze*-ou3&dghh5g`*SV9ki0Riw$NDd3Y9zqI~Ky0EwXmb-S@T6>Fh*&L_o-}814k~3R zCEuj%!3;rWPXS`AELgeJ*mTaj{I;6B`J#%-)tZW4a-8qn^)a^TpBSCi&AZ=tx$TTzV3v^>}2 zn2m1j4ao0h#_?$_48=HRom%Thf3Ddl+rrq4WcH=Sc^%2DeamArirFRI(&~<5#xfEP z!YF3zJp`D)s%j5Zct^DdDo-nhQ3#}bX-bfBkNm+9Zu9bg~R7oCe}Zk>xz9(yjL z9)B)&d3#MTxrIp{$LxY>^IbZ2C$ovI9mi^OYMss_nPX^OF(a9MY#GLke`9v0)uwz+ zbsVFDLroj1<525%R1;yF&$rZZsQZsq&17$>X3~#U&1A-_<}U9iO!i&OHp56||9hH} z9XSp}bhq`m*YrGEcKNpK;-h7kK0|it4nG%C>s=nljDWaJ#{I76Zg`9a^4f`r1~T0} zj{`xX@P%^B{+1lGeXJa_e;Y5yd{^*UOk2Dg$LtK;xl7}iZSXMTb|f>rR8D0aGj5Ui zPX8zBGz=2%i7IxE{5@u2sAgZm6cl$$g^GEs3KctEg?4#+GaggR%k4PkVB6x%NM>H6 zd>YBDd5yB;m_yF3AG-{nP0OqNIA-1!(n#nWVD@(8GGI7bCQ!yPe@7fuKbF}$oa^%c z^*~u>xxU)rvPR!LT+DuqIb2L{2;b9-i#djRC#e~(g_#+C-gkgwGJ|@0 zvYO5>)RFr8o0t6jf8lJkx?KK;Qm50^+0}7-GCx-z&Sxj;`t*1{U#|X$~x!*AIBUD5BJDr;6B$fA3BaXxYp~D%W$};g%n0Ie+zW(cQqR$!tn98rY&h2 zb$R;na&daxyxXi_z36_Oo^)>xp1tQ#?_xk$3(095L)&sCu9b`^Xw;~UxJ~0ASW&Ww z`wkQW%9+{%5ym#Jd9!UjtO>be4s5mLqyohz8QTU>9Q-B)@+Xtn7e^mFh~HvUAhOz@ z?gxE*YMG4(e@-`uGX-;1Nta%Hn_lHe=t~0 z&WPZef_(!otQ+;Xnmg>nVBIMM?7g+26T&zPed-lgskxo#&HrYe^L%?4Uszu-h-#gl!>-1MvoB) z&IMEZE%VThE99*wz;7oi1PAuEtyUXD%9gzeq5`O8Sc|IA z31aDce{k~saW1%%fn5`_GB;*iG`gq zBd<8%=mD&?b~{EENGs2lyK*?3J;=MrcvT>S*17>HLu|o`{(#HMDK@YcR9Iszn{%Ps zIajFdQVu!XSsp`;ae342b32b3y8pVrWjOb4f6GqS-MU`0gg(&8Sz4HNADElmQ=<*7 zXaA*vs~3aWQ_;Mh4m5j6Lp9+-hxKl#235H#Ei@vPR_*O|*j5uQbTHV4G6si!e@V$Gun5`Hv^JwUXesWRRr4p<_o_%Rn; z@2%I$5w&;{1ml%EB+ZfVZQ5D~PS3L)N~lJUFb1|pYh~mpt*psbJPP`|Nb?zkr&rQi z-z2vz+anEp+KjL*1xM#p6){V+K_i|mexBVA++{J1Finp{axFC7M9LJ@ngvev4 zV`%yPwyAt-Q!|EUTbYBfGM!|-Yx_ZDU(k5(-QG5$VnOY|7#B4+i8DY~A{kH?HC1}ir!$67*(}R$6Q1**0w^Eu__ufKw=I3#X~rh>{z2Nl(bY;g8?=uE0wgKL^q~IdB2m66i`Pn$b}5gzs5;=e@-@s65Ge zDlueKn91#W@0<8|e0@k$v;cNgi!@`#o`Sj)XYb&ONL)s-$Tf1G=7kZJGFC9EePw{U!lViGK z2-jRy_hw-5gl+k)SH1}2CjEeBnQeJ@?!zVR$1G1Ca1-}%$IYliYk0dI{K-Q!E5G_1 zTW$~sq-uzAc<2KnLx-JNJMIYEP!Ef-8=m!(?~@aDN$;xD=kyd(bY-Xe z!iJbEuwVXs35In%`@McRnrh!t6*H+Kgsv4|tUl3DiyjQW)H?z{9+e@Oon@(pn zi-}byY2FffbYF3`^)?1D%3IXKk8BNBRkzs&>#`6jbLOy88$keUY0A>9FUK0%8QR&e zL>ytwt6%ci4ICy&0ZPAy6ky^^MT2??@tO9zr?!7jU*|gbo(jm+C6f9>b2<)6mSTJX z;#Sx{1mMN=(I;3v=}KCufIF4)eTOk}<2G!;_&>o5s{L*vqG|O@iM^0$R}B{ILCH^G zfV|*q(X=Og!&K4G77iOVW%Byu`cKhcm6b-0xL{s9A-`@~wHCABp(D>BOK-^~%kp>@ z0*+?|qfbBSOiymSXz#+Sy zuoKd?K^SkFmBXoO;2*AIBjmP46a`7zdym2Bvh4L@`}%i@mfvb|6*IR2KBw(vd1t=T zhNt}=7?{9gWD~is2{Wq{I>nr%iR7j+Wc)DS7{u>RWl&AkD4_>oK(f?Qu-4O(ySjsl z-QNFD>mcxSzbeXq$41C@SZ^5R3jS4Q*1`Cja67jGuzYII-)24vxEP)OIVf(HM3_WK zc=>dGe|)?x_-#TU7s5kpWtu-Ix#_9t++LS}dy^ib6oR~(n%u@+rR({2^kXM|7}>JB zJO5eRM+d~v_|b0e+FLkBODEqFGA1!qxZymV_oUX^UKYC7deVV6bp{)Z0P{9K7KZBg zFS`~?T`!Pz{2l#_7p~^OIr>oCvUJDbTt~#lseE#Xo>%fha;vJBN;Ej zb8_~K5$q(z=GRCeG(8%*FP?}9M(k9cD}{6Nm(s?(Q%$Z`^aJ~MFXo8${D>A-r<3;> z;n^=%20w^&ph(LW;G`^H3INXW9!66OY&m;Y(H=$fH&W9W00%X-WRMx$E+gSn9g|H(GV?ZjWz}cF7^E9ECk`MVYU zea9@PDXBVR!$TQ`32ip6Jf0TZdR*7pf>n=SC{my|-0lPL; z{12Dc8izHoHhbjqN_hEjQTdRZe6T|{gb|z_D`CDudFqsd+^Ml*4803pTW5&slpHd7 zoJf;310v;@0xIt=VF=HIDNiz?98)b z3u!F&H|JJPtix<+M}zk}fZn{5?%dn$8q3J@qX4gy@o>z&Qh-d=@O%Fb+obLD+i0Z8 zgOAKPzS!)_iPp>N#IxNPdMIIN~i_3DmQBPUVl6lQv>-fU)@&D71;P3`0L3I{OLoycsC zN2iX{ggZb(wM|EvS{r6i%y#Uzku}ARy^nK^G~WPcs2$<250Si~udIljLYq=!CVPb& z1LBlRBirTr-J}F^3DzfLm9wOliv_QVquSRd3E+^WNaChbO^kqwxh53QndR#@=&<5u zu&S{haVIyu8ZsmKS!`;XQ&`q#NUpPRrham(mtewB%!`lvIytM>uo%)XcEfMtCIVln zZDfNFaXG`qvNvf!3u$Ru9ui`mjuZ&KCKsdl1qlEYD6 zzyIT>Bc@*IFUN7;gvCXojc!K3fMQFZRfresr@zLj0dW*tCdukNP2=C-64Mb#;=y!C z3(rVkYh1D;A7VL9mKg$S;p#b*H^-y)2KG@G2t#mA zqAc@Yh6i&U{j?sMtKEAJ&Ostyi9zQZWWAM**pN-W8r4DE@Ct9o?}3)pahcHD`ydC4 z(*!?a&kzJdz8rC+-)(CZ?9{JsaF!~-`+7GS_K*Q`Yz4h21{1x?-0yAG{@4WS^P>@k zP)_2Gw)0GO(@JRU%$E!t4efa(`G66XCOj!sf{)!HSXM-$xyz4dfaL4F?=Z!5b>hTQ=bd~N~Uh-_^H3--`= z3edb%!OCB8%r#gTYl=n}A_^fC;M>vtn}}N} z@>RGhlegZ?FQx}o%L#S^J}-LvN^}>pGBRoNZ!)&y=ih-*C_6N*`zk^s4{v`}{bIwKG~580c&*=5N(^*nM!6 zCpfax!%VB~lzawt*@-T>@fU!s{7WY(A4ML$jDFn%lz$|TURj=Kg{)9`r7L=B!eVSv zlZx=ZTV+i`3467a328PYf~HN|wGoC(HXK&MX!XAfocy$!ho1%!%;3WK1IFx4?w=J3 zh9-b{v8w<$z}OH>d217v#tr`g$5rH5z8XB|mM-qsjnGEa=$7SsbNbfmrc@(@OXiZ# z$sRNzKYZM?9iEE8*g2PODQ-`tMCYlGSKG??P&N;bQ$0B%DUxA?EiOLOxj+beT!RjJ z(~YfBNiy%d9&xSsdFkxbj5qY9wO9~C_Q1V=lt(OiTwUN3xdR+a=hjo@ZAMLshWC5q zXsUQoX8&A3c&=;a;U%O!yZm&tDf+=?{Co3JFqAfxpF8d_5CzO zme&r4h}BT(YPV8lI<-NO3Pds9m`$14Bn7Q!L@kOC#!1&@M2(hYKYEVR&t6MYHUA~F zF1&x~U=&<&h8V`ztF-G^Tcz%PQ&%Sox5=A8HpiR?7;U(E!Of(NG;D#JL7yE1BZ^;1 z2~N;GE|ej#(8>Emx9n6H+v^2i+UIU-?J|Oh|Bi&V3>*^_8TT9a_p8b>;A64fK`x3D z)T1h(9MVHO@uV;^=iP~`-CIjMU0Er5{4wo_BL}eAl1G@0EmgOF zmu_*IMbJ1{F;KV`w!eAy?M#?&0&_$fT@#~oizkfU+#G+!5+gyHq6#P1!%om*u4Epw z@d#b(Ia$UpRI1>-gaJ>pklUA$R)XU%5_01Uy9A(CwMHfv#Zi3~%!B)r0vU-`tgP;h zz)k9&j6*}7tFWE6z^rn2l7XP? z3j@5uo9L<25fYBREMQNW;hI$|Uf?)A*v~ewYHGw`pa35++1hy}jj1JU&S}zVvDmPh z^3)3?ivxEsfBI@NAxUjYVdx&OwHM>k@3Z2eV&sM+INUg<_8@#xHe+*DT(XI3jUNbJ6$BD!7YiblB~L63)R_wYG)&cD^j!)CZ{l7RjeqG* zC2m3USF(GS(5r5#SpU~MBWvfnZ=7}41uJV5s;w1zPUIYW4&@x6Ic!r6X)Z-M@_&1_ zdfjGzI@#y7di8qB(}I(%v(@R7uDP?LUQs#5#p+f0R)3PvwNay?u8z@jq-o|G0-}FP z7NAHH+mfg!vg-t@l-{}JjFxw*{fTJkRH9$N$75k~3hQBnUe0z-B>{|5#59X`drlfFP?ZtR zJqbl&hevA(unlwE_)?7h1L`K|{+hYZF3Wnwce&iJgK05w0U*q9Hv4x>=oJETnXk4Q@0R?9;5b#&uxVthRLy;PTk)H_}W6%@2A6rL8t@_Kn~SxR8L`Cf`Imf4tVzTxji_ z9nE{<88d(+-fM~QeH97RUG7e9$}vj8c1T}mokiTo6uvJBiR!fqCu0A|G#h_Qwv%Wm zw$O?Y7up$FHzv^&`8zNEs9j$zd^;7>-4Y$YP*Ys3Rm`JWEHe~WAvzyyjVHSAdT~o^ zr5aHlA72h8Dk?sz|FpSGfarso*C1W%t5x35zR}BitrxIc{Bh}^{M&lF#cnzO>|s;o zrm?_dx7IYX#Z14p=@k-sw7|Of_Vha@5e#!+up-1Y6iZBPxIQ~6c={@=Ob6Q#qHY9! z+1C`g>#icp1wC^x-6FdW^pY8V?fF}x*hu4vF(DmIF)LFM2c$2cs3b2a#3!nx%qOp? zD6Swc!Y?K$uc9I^tRyeMFQO>(QtJPE#o~VyQ!hmYMPB~*de1ubOUFQuJnSUo=(p&Z zQl7j&HRgm3W2Vh2096?S1E4gYW#VdUTbzhH@p*&gFBNS=?Xiw-h=wE{CnuTRfu@kZ z*lOKPP&KmMVdXBuXN*30=h=kTQz}*T+_oB$q~6OaJNG3_7osu$S8+r#4xsR&q6n>` zkV?Cj*P0_j2n<<&(kBLFUAxZ4ui=o_ZnI44iDC6K4Th7RoQ)sgFY2?0?tbGwR8iK& zXVo2|cF=UtY^;p(7W)(KT^`0!PTtYrJPTu!9<{#%x`w%)1o$-m2~{P4eDX;NH!F_} zb`DoLN&bFV|Jpg4>sT4PKWGI;@N;F|NgONT)Yjdg9UGr>Tcm?LQ!kn z8Eez~ecr|!%B5%u6IO+IHw#eh!NyveLZ(utQqro-mk*;!e^XAW zr#;juzox!Cd&-O|`A^Qv66F*Zh)i+4{Tg#d>M5MnT>c+I!XLj_K2uYJPKs|lr#z_o zf3vGo7U*`l=$L(=^*4SDT5l}XZ;uShA6${%1@EG1+%bIl_J6SdJ}f8nR;Y!!tG7sc zw*n*JJJ2kBEj^=rpET)Uv7TvNGgm8@R3^}4GBZkN*nJ`Mcv>;npJnB0@uwMlw$@DGga&A-Ez%g$iV V8d1Svd?73%{(_ZNSx4o?e*lB3$QJ+r delta 38718 zcmV(!K;^&Tjv@7pA+R|De`lzTl&Q$L2_S+<=FH@|2p~Zqs$^djjK^Tu2Aabldmwti z2!6&NJ62_YYkyxnmaTX#B#z4OBNG>k+%`ZET1LR<1It#>@_s%dIJe`iwF{{AUH+$vjY zmPyi;@FFqs93jU9wPEh6>CcJL{*}F|WST>;>zrje*n2pkS-D8hRofHQHnt+lqd1po z15N?`f&f@Rzu*GRd~yYLpY>ow6TE7;b{eK2p|MjvwikzJ1NYJ5ldar5+Hk^`69QGI zr>HxBby|v=V5XIVe5Puorzc#fVuTyZ4Y;IUL5}#Hj64k z3uYjdt#`u>RAWN`K$W5*t{RWl5gkk=RUr;}+AM1K#4=sF)?JT4CuD%l;?@Z{xP))L4|YzJOsWz5X9GPE$0g_?!-`5NQU?f) z$Ii0FpSJ-4wO3#>aSaH$g=?9H#kHd(!WFc*Ity)ph08ejgblemY@@skZ9s_6D1#jN z%`m&m9KJo_e+dwOJascri!97PxB5qK@|+1{NdF#D+;vz|q)Df$ZHV&031d(tES>`4 z1IANJ98gW5#i6&5PxcCe{Sdd1m1SB(<_TO+G*bMBi!%|)Z^R;;$#QA7M#T7qYQhcE z(0uHvPEbi$=?G4=Z^WyLZZZnGBc8}iGO5BfuY$3pe*^|&GxPyfFqwPIM6T^<`X!po#!mI9joLVs=`Sm0H)IVgffeO>RUkbw71e-6;uoa3zrHU60Yx=d?qyI<>oPNc~H znsk{V*l(m`B7@p=2;~pMI)u7wZ9rrdBlTLFq1)&!wrMV6^3lWgeiXxFG#8;yn}+?H z;ixjEmY37xSoJj;cDf&g@D39M8hfI^x@1A5+pZ((T>ZHe3mtk~lw-dT@xR%QFGRNe?_MzzlEWJycn;)G z;a%mOE{jmdps3GdSQtU?`1Kr3>wLKqV7^Z>@F|Y;n8Ka_?h@L9(DT5g?Q0ty5W5S# zI+FIEs&J6+`4_cLZM==WDm{korN<8nka!L(@YI;k7k{%4Zz`T?rsHG7-!a{NmfZjY z7UI*>w=pm$2xS~w98m0t4Gm^~N2}NJ3%pC$NYl?*p4Ipk0)o?$?<`>3jP-f4e?BB( z3b?`+>4jziTIdFP{H&g;B~LWG^)tT7aSBQ;JcLTPpDc9KHN{=2R{VRx;Bk`dGZz22 z_?+=PFm4WpXN6h*DB;GKJ-&7Q6Hi_Lm;rk8#|%PMi}W`Oh@zJwa3E07X(LTPe4z3D zPJ(mY*gnqa{r#t*|1g}-cC1_f9)AI_D3*o4mLwC*?%U!z-Lp+`{y$mfM@TPA<%QiGvZExFt*IxOjuDR`+zDIX*{@cY*P&BB^ zInl~2l#qlOukDX+1P)E=UC^-Z8_0MwiZ)&}$D1H-*WnIotm<8pOoTh#o#GJu^$(D; zw;VNw$f|-W7uS${8Jh5@i6BgWtD(VvN?B%x;lyY$w0X)5B8jTT2H$2z5Jcd4ltN`_ zQ!Ny!G&g<_D$SyxL|`n}aGepQiBx(KHGZyX|23Ll-z|#-ggA zKVK9~;y*daqL6Fo zAhTeMEx+=Y=IQ>tph`#cZ`+e||zt=;-0opKr{k_^ZHBJK3{n3PHo30F!wL_0_2cTRB5 z@`O69RD=taj{emVbr?kl&R?kx&4&b4EuuzGCl-rw0TC06)El>b**+ZJ*Im7Cb~d1P z{gD}Q#tCQY(+oVTEawV=RU9gNL?LOzsQfe&&nu)1Q9|Tg-P!Md8Yb)3?WPI#ytDxL zac|oks!1adWh+0{&k-Q<<-bNi^v{TvVz+BIxBjtS-EO;fcenO%34}M(EvXq{LhJ#a zsS*a`|U+ZBgFhqX&AXsCNk4?rPib{M{Jh-P3@x_WruD3cAU=hT9-x1%uRK zzz*Z@XLV$&h>{q8SuLc%@|0l)-({i9aDIptnDh)*e9TCNo9*M~x0w+V=)K7zRzd(0 zu%QN1VYUgc=HUh`gG-GGLtDV|F}U#y9fM6gnQ zMd@_NWLYbs^d7e|I0+AvQ~{YwJ%y189gzxXx^I^QQ@izGc^i*ICU=n;O1aWS5w&`k zSImx;C`Pkv!6`OmyL54^&;d_%NBlD@7fQe zw`<4TZDbnAyIyWP-?+ZMZsWFUx!!_28U$%eJ&;(I5#xe6O^QtRQzw-u$5l!6^@t0_6*GFPBVLlQ!*F>!@| zY9{Tg2-objHL*#pOR7X72}4m!BP$68@Af+wf>O4OrJahd>5s#TbLv(}>g5luTWq_2 zNP&_}4sU(@l4|+u@8<X7&FomRe7(LuO55v<{Td%we4<$u=y$d4y-ydq1DDI7ce{F5q+9T)4u(7MGctj2U zKhUYMD!!nMP_40c%rO8Te8{8?DaTz7=v82L*hYj0qkZ4+>dRHz-vJb*vDjpPFMjjc zyFoY>UQ~3sFM}oNLd-=hor@)WmnQ^l>{1a4<>M|hdG9C%&SJ}8ahETX2FH3@-e_Ge z)suR=yv<<3zJ>!i5=KGp4+HtGnUtAJ>;V8O=M@MzJ(e;LS!|p$$01@rl-*V`O@~EV zVKvl-r)ogLf8=4w;yBV>q4?*2&#e<#h0S#loC*zmGM|(*M~i0=-6}ofWckWlMTrqr z*2$*%v`C_HE}sVO+rS&2P-GKo3l#Doqn^mqY4rc)_}m2|yM~Y#%w2FWp0vk_#9YiL z=CLZGM3=)oG?S)}>Ii#ch5au4sh{h|@(3}Cl6KaKq{r6NhPmSU5FNpPe5n>`HZ!vV za()F2yk}BE`3&c1lUm%A)@nx$G!dBkOLp@w0}Q`t)8#7o)K+NKA}WlXCQdRv*)v%d zBaa_fzUX}I)yI|k`P57vKdv;>QdP#C6P+ipUQni7ski`Vv|-oRu_mo-4gJ8sfzcKh z1IGtjAGiyWVz)=!oy$^x&GPmEJnv{(N`kVQbkv)S>)2hRPo77`3Nt0Jg+)5_$>J44 z1)EIr>69QXT!D5fXr_1-^)oi$B;9OrL2@u?Z^f}JVBon0SYNQSJD3>mJAZ9_*0k|Y ze>C0JcO2gL+jZ#Cfp2fuDRT^~%K%%f3M9i7A0F2B_tK|AU4Q5aM?_D&h)>$QFP8`m z1RM!BxsnE4e`x(nmn{qg9e>Ni=bwQLp$U97iOXO50D(U91c5!1@1!9$v9>0qs$OGl ze(d!5C0O#APAszAu{7qymlDZAj6>W4Q78wNa9_dI{|znEL2S*beCBZtCyL7CxMouC zK)zE%jUKyQ_Uk8v$4rG~QY$U@hpQ{G|>wf)2tyW@J&E^ zLE&LS?M-<#XUe|~A2E!@zs;l!xM?1}abZOTE6$lv#-AfSFE=|*|HPNf3{vHNX7C!_ zzhs;hpD_vr1GrDnGJgiJ-{8^L7ISBE67eU?NnYrbbNn|dRxfa!ujuWBYka@ zj?@>o5AMJlD^r7!R=z4rm)jfvXQ-3X{Rcq3f8LjE>GG})2Wg%{quaGT2e{-xDqN%IN#((pi!3|rUm>vLAgYLr@ zKjVP|2xy<2V#{TJziPvYhmS*DTVJk*X5(RV)XX1mY{Q4FL&M$VbUeInx{U`5X|3Z= zWcoLg!RJ2eqN;MS&rhOju3J|W4iHuECdZX7;9_-xMBycBD{+!q*aD>;xG><}X83wF z#nSID&i)6XCjy~Gmr`olul?mq zv3-JnOPrPNL9z3LerU_OXsbW2kN>@^+P2y~T>p0aA5E$@V~m%IxlXNn}>GK;AqM*!x)R!~OGP(Ym3J#nY1;y=%TQ zk94*vTmM$$#;nUfu8GRZ&Hl+fXdg>AqG`*2r^a30+wF#bR^9Flb+aK@`EmW0eMBIo zvD?6}Ic09|@HU0X`%6!eWrf?=-cWQ)as!&4^Keso&co&XVYh0l{SLX=8Pbd`mQIdN zX78>Q&AQ=DwSC%DtEx4mKe%VlZuR!MXqv;eyWTYWZgq~AEZZNhaaOzQZtBKKlBhs` zk_1*FgalUDHQf0jG#J7Vk)j*$YtHio8H$u--I$jI5srlJ2iznmFvQ@#p9S+I5s68# z5^O}EvbW|(X6BC2>~E>%F+{P5s4uzb78pY&_(vsus=GUBXXJqh#z{b{;Qf`zdMoB;nct`@yl5{uOe-kT*Q4I8mU_`KX!>Yu51S$C2&-XS@Fbp&iYhivQXXo4Ihs2r;e_{#w zB~cuaF(wpAD9@AXAz8ol`}A~Dr4%BPibT24XGld3q%v?_42Z;9HESF%Y0J!quwSY z1H0$oz%jT2PN$8WeS|vw9~e15*U0&tk+ZXmT)vi(%X5reen}f=uVdr#6>VH4Vsbd2 zKv>|=xN!>LK&TTr9|JhC!8m~PQTx2qoYje(LUVpIlUXoj-3bSTj=^|;Ei(Vvq!tz} z5kUM!-MY*Dp>`KNx^lO7J$xn(Qg_#7klMtY^s`fq#qT~h zB3*7oAW5qs)#8vO*^~=^;B9k~b3x9?!9ils2}|?2Y9Af+t;@!wf$-L}zHlUN!e#{ZE13te4aNZU@N)0K)9;}C7Vo+q_mHzJu zU%?@D?x}xk23&o2d-eXR;}-#&r8!*q>FPheUJxgIo|b~e!M^>XjJFu^K$yd&(RfAUI2`T4#$l0gcyugmkVJ$9HwK&l zsR=e7UIAq&kq)onWC~m&bFq^L?$cz6B$*`w!JN{IohU!YiSo#a&W6;P2+qeNI7b}r zqbZZI$6V2iW6unK^T_$k6K63`B$yXw{VGfCjR*uA8i3izo|i;0P!B*T#JPwG2cf|D z!NQgjY0RgMm`)in0lWux#5jZ0X3~Z+5;pf4wP7G-;4EVdNx;al4fC`ObATgX%!ctN z*szLML%Xv&&CSonOf&$!pu>7jbBp+T@#mq)l>)jPPJ6XWk@#Hbqfi z9a|N zB239Ff`$!)ZUlooccug;k&x`O!}t_d!3JS6W0E35> zF?HoIAvjupVKiOJ3M2bc*4N}0#yjIfMmA{Nh$O=u#^8dBkxKN4{u4A8k}#-0cupq) z@713ltCfM99veUSBaEgym`DL*lgQ4cl>nZP<@UEJ7DTaNT*wrOE-Kz8$aohqUIkJWtRy!yItOHrQ zONs;Q$0FO5#k!fO5|G@~CuvM$1Q7;PlUxJJ8cc1=19VK(M^Foy{NT-TMf+G^F;Jo$ z<5?wt;)%+@V6p&5jV+lcmOOM^gC&E?uwVAAt?&_^HT%uMS6SiC9##_zz#Bin_ANkw)3y%;9)xflZbJsZt?5?qmsLmf zQ17OsAq`rh@%n;zA=2mCJUU1il^Ix%8ti?fdg$K(EN-PcTnhq&Kic#l7C_%4gRZQ@%XaWz+bIIiIls0xK|CO+@RwsyI$t zcfU*m;gUgG6_K9vSTsk(1ibK_m$`{3hh@r>iAIEQe9vS^e6aUoG9Ggdrhgu@-Ct^;qwAtM3})<<0Tv6gBo;g#?GQwV zN_?rUCeF5kL33=S6!3g5F4jR4L^vPF&;~P|{d4b6Vg$H>YDzDuHWiEkm4DAn&9JG} z4VK(i+tN4PffqS!T7OfeLl@B$>D)Kj`7=9eyX|(x#+PkOBmlBSSJmnJsQmGN(y^5; z8Tia1@!({75cYK3-ZgBfTw*RQ!m&x08#vw_`IWo-c2j^8H9J}L&0X;LALfO1@y*`$ zT<)7T#bkRxM3pat{;Tw7ny%%MWGe`o)w=WtoP|Hwtae514>MPTx!u#JdVjYmx9+Ap zT_b&9R<1$Q9G$kS{kBXt2Z9BElVef;;}0cFDU5~oL0r~!KT>0kiX~}-3km`^>8^ld zeIs=~A5l1?$JMGY4L$!P9>5%cE#TYem1_FH77K`G;4!0jAP=K?I=}4ak^&s<;Jurf zxfG;|lEm*!B_UXuf%LLTv^QRKRGGA!B+$7&C>SSS^gzGUdG+Tr}36&69icE3^F_Bpu!HIM}vv-&!9#8(^p} zQ2V>W{WvoR_ZV1u2ii{@EisPPBgH9y*{cv(9FAZ6ek2_`zM~z-!k1<+2v)E@qpIOM zFLM(q4aiWPOf-g?kLo0Uta!rLqI#+}P7*wk&ep)z!^(y-mOF;$@MpQMR-H%HS=;qG zm?68qvqXEz<~~B1izxen2Ow_9?lyyLvLbk7|wJmdQHnAq|a>~(j<&Kg8{a8lbl!H>bqQZ2}kT2U< ztv^%GWn?2TgvX74)juR#T?gl|pRe=!h@Si)k+Rg``H31m%=tcfs&6_2^1rNua-;+r zH5a&+`h`6F0?3`Yz@yIjuS)CwYx`Kbw!+2@@whoRA0GM0>)i%CD*PBr zviRNI9l^Lo?B@T+ed9@Y&Iz7c=Vaa5DbBQKv4VQgXzOBsBD*W<(w-cd>*%8@ViG=E zI)sNcEEAyVe(P=@4mP1OA21o50jZg<;SIluvX2}`q~VBT2NAv=*UjnvheXDgopuu<%J+xTM$fh?HtsA8FwWBxth2w1KHpLyi{Wu9-!akkVPAu| zw*K9{d$lfqeKR5Od+q4UY`eG5)=p0AlU$YOkj)$S#=Q{yCp^uupTLC_b_^-&qAlE& z<06nhtojh+|8kpOl`)NQdo-C9|X z+sF}qpI^a839x$E7rW_&fwdjm@M04r@B|28?*m8TXb4NBOiDAJ{Q6Y&L9wZ(=J0NS zH3 zrUm?=^WQIj?H4DX7cxy}@2mZPe8D-N?QhB&pTw-N1MBK`&h*TV^4W4zbY1l&<)Uoe zipI`|m2RLds=BJz_U+L1kNu3hIa9N;THoN#`wj{E={?f}4w%Y-1EA25e&=T7Nk($T zHO^3ka+WM85xKClDmyEX1ScHc2a`z1{eo!)2jNDNNci82>grTfDamtxcuEtRP@NK) z^CT0LWRfSBTUhqC+3pT|!=ti-v}E(4YTO1)NTL-BHz?aS!o%QLmH{iHQ;T|KPei4x z53M`DIqdDj-A%PbIp5gV(yP;o-i#G~5o&i%XHLOx-(B16<_lcZZtju3o{3%xEGJp6 z1A2IPu<(ZjCgf+fO}&PHy|p;e4+T9-eX5@2lqLVA#GPw*mNVjShoVNC+k=?{MfVYl z%QHf=-r~Nblt+v%BV2K7bk3``F&(FbgVzmay#o0uJ=-+k2~_2V18P@tIy6@xA{9p> z($ui{?(*W>g%>(!45&~^MsuQ*<@Vz9?`g7vpMC{P;h=>(GjW@LPzKJ=;c=6Ey7-3; z69JvaodlsYvxoLJ7%=YkdWdsHB*<_u$!lpOY!0^Hv-v8*qof41_($;A@APaQdkY)+ z<$|T)nmmUUf}~m6s`5ht;&6JU9EKPtUqT{<1F+!Y# z7(tK6MQK9Y?n%0j&?6w>hzI3$qMR&3Pa62pS7n&4` zFC7TQqzb1{*ndEh&4r2xuxB&>@3Amu)xKavWNw-RBH*P%E6kZ3XByZl3p|H0<>%1I zSqn;mByfHO7)6KLA8rJ2l;*vy#^eMcRAdY$=2}12o5f?iS^W3BSvqi^q7g zc%nDMqLg`m_y{hhKuuDDF=B2;`1pRV*{s3ZBA3o_b1|#-Wn1j4rta+fUA4c738_cd z2ov&f=WRqUkn^+bC^IHvaPCMk%LD8&|^7&3ve&6ctb+BFLxc8F9gqLmq-rF zp=h{rK)<`4%WUQ)jz3v6pCg(#uLiEPZr!%(LR61~NEZI@b$3Q#Jm-KSWJqFEVJV@E zN8%O_D<;po01b1^!;NXY%IenutN_pvuDL0Y<69JgW4ge98YPh*ft$DSln#V1Xae8| z2A%qUgUX1{s?H{*P0_CVs`{p={ZrGG4kO-R%5qBfm9$fX$!hTPsJIePzQO3Qg42A4 zdKYr`+QZBnPlBCn<}S-})x9_-&?%z=hz`zg8EdY(vIWyoz3K2&2kfSLNOS6pJ6j$$ zd;4zM3Gn}DPxo%1Y1?wS-?-PpZ!m07;5hz&%TboVi4`ZjRyFJ-4=_~L%aTkolPZ$t zPD}3!Pc>7ig$_*70bxsP)#hh(r*i>GLdev4c)<3&8Jb`-q6zhcMA$YB9(>)jOXG1k z+Bg!MdTCfSCJD}0+Hp^;GbuAP)>u&AJFx(0BT#U-&{en^rtoG5-j7(y8(dY7P}i=1 zEpMyl;Ee4U%83&g97IVj0Y{i52y_v}U1E7X=XACz-k?H&IQ&J6;8Q*_m`J8$BiDxTX)+EC@s zfUt@gM6sF8N;WKco^jlJL$NNZ+CAETUA3(+1X#Byts~dfw?i3v%4_@qWLy}3A=}U% zJ@5rSOQT8aVOv~zRn=uT1#7s{z^UufbWU0&h#aDczqqyC6b`(Jn&TY$e zTXtPz$5Zq#1PhIQE9~TMk3BRyv@OBGje)G}kGEA3PA=Q3w6D~qANqH(8=F3>l~ z-$m1^*w`ThpzJgg@nC2J=72_rV@_)zF=Rwj$bvTKVbDFeNlO>d=?ErgeCic*Cb=s{WX{5a-=4>%Os7OZ0UoTGDBAjSw19$q zbk6ky=Rjnoik+N;Z3X~>(Kds}oZ&ZG=(zKB~Kaw(ZdejDQ1@M2!9<$|3q|G4iJ8cMOKq;N6~o%uh{nqKPnEs+H-{ z2TN^Wo#sSf$rovX{)_Qe7c#14qbiq=(ffF$MyQqvh$)8(RUThp3@ z-|$cvn@Ww85t=H<2{0yHwtxHhrlk;5P0QHbEo6;8vHTo=*#IIv>#ugNqc>qn=95kM zFxlBAJn$#vpOEZ=OuT^1s_*z*WHx^9=tfDIVWQ zev&Ut+zN^lQR&gGAUmf)2*mO!t^6PBeP`p|_u!s83$8|$$P>JIc5_*TF0f8;l)-SpLf?-1H6 z>EWdugc+cp=3sozl%qFtAL|N>r|RG1m{S7Q-(ooz3nKuz(tND=EVMJF1Es|b8?jzK zQraJMUH8O0Ov;71nm~ccOr5*Ke2=Z#CU!Z=QxHpk@5I)zLQ%_7?bDUJ#1-822n1MQU9PR6 znpB8Ct+9-B4NsHeanMd!?!5p-w8p_(nRImQA$skoM&WFG9_nzs_mMQ*`Bm&jqsO)x z;o}8=6XDRt6{n0}05D|H z4&ifdKJE!+{Y#9slxQX%ONs=epORRd5HP0weDTB*35bKe)c|lpY!Z@*2}`6slzP;L0Ui+kq8%VAP=7yKu%3f zSr*6#;NuhE;p5}SWns~UyVybh<-ldpgE~3GkqEK>pnu3YLBTHfHhHkiJyQ*d06cfK z0{{g8Kw&YUh!`IqfS->K^bZl~BnFTNyTKsCP`ad8mi<#l&= z=K44@+y_ z-+x=^dOAS=IDx;-_g(mSIUpSXu=_@!K5!WH{s-5~8SDlHxH!2&eZ2nD@!tp+2n0aj zmM#Dbs1+Q6`|s@cW+?10d@lhfxCg+P?_PmG0N?M=zgDLAB7-0icAo#H|G8yeElpi5 z8OK0QP^2-Gonw&+`5Y`2Q{WKOz4&rT@*vs?L7ac+{@S1<$e^@koNJ#p_Z@g4~Qyny!;2Z4J0ktu+e2Z40C4*}el?gM}!op68On5Za#*WkD54+8Q7 zcrBnV;C~zV?k%u?L!bzN*Y!Uj2!Fuq{vXkN0A7#(fItAR=O6fAC0n{WIoO~S z{Dc2I2%u08s3q>qH>Bm$V4KR|rXL?=sNK2Ohxi5Yx=$j{9_X(rP`Yhu1#XX^zki|9(qi;d z00FTgKdjg8bw86H=XlKJc)UKDv8^P|T3>!yW_^gwR=daZ2yG}P3Y$VyRNxU>aI;KC zbGhs%nemA)%FNCu(nI@+w2Iq5Q@i^4g;u&}i@8wY%c;~mJ9rK%XBj(^Z@1cosZ?H$ z0LEF?+7&2%#jyvTH3bK-TYoqYXeO7bY^6>A>hUB$nNF1_qu8X0R(LGW=V#OPz!U5$ zTYHV8H*4y&<0GIi+moYmR(s+l5^>{KtZevW+IPdde_+2vR4mBB9V?>dZWY9i^czA3 zI@D5#pF5>r6OK-U)h#BzK*)NmdiYfFBBT(rMreYUD>Y_s$n#1lvVYM+^>-+HV(G4B zXoMuzxxe#Ri{SO18+^&FP#3|tepJ#T52RP}W}5^4LJLsV0aCn|Uw=XGF`{<=$?Rn!(zP&b+dlWRB_y<`a*;3Dcn} zlNFi_4f(bg7TY41V!t}m$@mywf{Js-79LB_iA<}aa8Oa_`>gb{*h57eOUuQwr_*mb zK-(Q3gm64U#&v6A!U=8&6~!Z01777a(1&K2g($as(N?A5uzx`d%kcWQOIjWjMW4WW zsS{@HNAgDAN(P8Y9zVx@tlHcsWlyDwxyb)Kh!vwi&qMTU(c$h(_TkKZ^v9^P6e{u? z_GW2@*(^K(8U;DulrYq06l!;7wQ5l*l7tENZ;s6^=ShEjk9PK+C4dIj!AYe$b?3i~ zeH9m(6JYmE<$HYm)(D^F!&{i{oND8x*gy{j_LolVmaR@NnnFPh?`h^iSQNtw#+3T~g_SJpQhJO)hQrV*rvDpKeBSJ@nRflkU z0w&I6rL*GBSOE;-$cr=gnaF00^VEZ*Yu#*9rH<<%yrzVzDO+TkowP1HO|#iduu(H|KCp6-GdEGUk}W8n@v)vJkE^v|uK$Ft z2r|5|c|y6K!Vu>~BkJcQkat6R0j5RgBbK{b#X;@P zynpB|kFi9)Jf%=pWcwVt+l4S1cRA4ZpGf)irIlj>@6=3Yjy9>Oc+)VYl9&91M4qWN zi7fwHfm|$)cDSU49<%*AC8j`DqBisT4D*isC3B0skF`Z zO&AAkcj)>c5=NNn%pP1QINYzGhn*zqv}C>C*qip)B#$O{xkT97N5Hemek9(qdvb~# zz8fCQvKx1}oH7xjYdyXaFEZ1XVFV1lrL<(J1dl|&Hg!LHhF#6_DdhD_4FDOp*?&^x zau$#DDe6Gpn9!Ta&iv=ES=+3i$dZ0i<|F$_PF4jWIL@HAKHWs>MMR#Jrf9-7w2RdZ z9w^rnb%k%7=*gD0HZEgtGQ)4nF)<~Va2rym`T&mW0m|UihIf`v%ZD31 znA=CXG%S>wWZv!j(p}9Xj(~A>XMYSR`aPavHg1=0QTv38@rs8qU+i%>qP<7n+v7%d zAni+r_Ytu=e^iCYmqRt}w%}>M+o#J2(<^LuLu@QXO5F3gS1%}hYR+S8b7~Ljew=)Z zO+>DusA#oFi@VK~(*2lKL-9~LnR-3vvZ+WMu3oJ{B5(ln>W%N<7W8;>-G9?adOGUL zhy8ii{6W~xhCfva`VBqMHXI!W3h^$F9tl)SA_wNYigh|ImOb54vdbUk=`Vxvrh9H`|c56*3-8TC6@yd+lg zezFZfBA+FLb#CSD?DQztF(poiX34sQlpATA zbRy2pG|W0*5$=KZkN~E;ed4Yk_6w`_U|?XX;%*`d`(hakzd6Q@YJWwYeoDF>fQ8P_ z77S`9)bD?tb;j;DjFyRJi6lbL+ZsWR{3% zbBvOau_=&XsCfUBslY-Su1lJeo+}cltv$oE==B(uWnhGQrjj!%*ls{dCVE`IofuFs zWK%Ec{^_Y6J12b@h$yIIrY`WslDaKhe=7fudSfxe%A3~9n3j3#*>zDn$At~U zW%h)*zwQS$Wr3L^tP|TMVimAFeI?_Ys+^eETJUNrNUh|^0NKZ%=Te2iSt^U|x zfsSw19{_AVOMk;6I2ur-n1eknA2lX63}K4+*f^AN{Z)`|{n;@>gEyq~Vu!J#$3F)p zq;xWxHOg#S)=mXX^&qzf7S_A%;i2i%{7UNlle1||`eeNZW$XvyZ%#qJO{-lWBV-l^ z-dHaP7z?h{?xtDNY)MtHN*v4klGd<2@}UpIVf#_Y41a=~#N}N;V|~UB_7PC4t9HFb zO+WJJs{MhkOaddjYL*dvWveRX73+?U6D#$d6p1&Q0)n2j2B0Qtv+v5)#yg(?8w2hK zJ$pRP6Ws;%l$3S$9{@}el+2_xZ}U{|-yYpLXOVt3ig#HN{W;K;={zbCqshen7o1~S zo(h9;M}I;f$^ESnfxdB-UhY0>@hgtaqU*&>vF)9&rN`Y^#Y2F%LR;CG@5!e-)YoV` z<5uphIOfaj)jsw5b;R988D`*523+poAL88+#qnro+4=t)X z1xc|VL70qf-d<_-seta}*Sh?RCQ>{fePb-7w+Ec%VJ*H9Ox(ByEW$_9?Gm5p6lCpm zu&QY_%f=dAG<|y`|E7w-dPbki_eoeIQ?LYCZ*wr!F6BG43~Z?G#gOZ=UR;E_bHxfN z^?xp-{7)8ZsDo%z{Hx<&Z{vzW@!`}LMv*})*lEwCKp-Oi7pi=PyJk*<%l%x|7dnQu zTUL?cv9G2d>E<(uNPbH)rew6FSC8+wZdiNS*$?{t?fhGtqh_ z7tPs&0R8E%@eR3C^7paNr;^@e%g&)GTQVY)A=7})3Rm|F|L>YH!_;(5NdT5l{J^ii z4mD+i@vxcJi1Or}q4P>N1IG{VQeisWyybKO!Qpab0X1LY7-OM}(|pO7eHj=2vwv{> zKxQYAfWnDviDX6IK7Vy_30~K>We@?yi)l0o^A1Uh8r~09m#oVRScNclChh8qB}e|d zY%ifUu~`P2wreRV?TtKCChF8daC>}u5+;yho+j_U*F7MC2Z3;2}W|<|a9#56{p+63>IuD}Sf#2h281 zYWB#&>a%4LvWU7<>N7idhjVg(>04&fkP=B;;xMJq0is5fBgz!nMM-Q|hN@0{cAsS~ ze0rFz+jb|Wfa|TGVcgNTS)9%4x>qBkbE6Hzj6Y@LNHCSL&rD4TuD!g;_V7hB8gYEJ z+^DowTWvSP`?ZgX`WR=@o`0xBg47CwmeEBPo0kOJpyfq$il!K+V*(tK$3Lvptflw6 zaK`3o2rbgCQ)F6&Jt*e%LHqBNA#TXMJR94pk5?=h_)1+tMw(hOAACu8TJ(MjG`yHh zKPC|chrgfL!VW;u#l_KrSpjkw<14I+txr;|l7xpQm3vdnvl%dqZhu!Bs>oNPHo;bn zF@2UE9F3Glrr@Wr*t}>_BORaPP5l@@WmM``tmjklhekR;_U9szR6kR{;)hq(C)ml1 zC^{a-0`r6GXs3R)Fdd(;ig(flB&uqcbC}L^7q;-z<#@mU-y%4o}k<8EWz zAzu#LRY&-PCx9w|xqrfO+@TC8b$Y*;$--#B5_7)i<)x4h;S3=Ani@F7U8dHd<1`C7 zD*LFl^7f6Qaw~<=UR+@^+ zfEc&wktz10_ZeDi2ktt=90JUH6M{;g2MS!-VM=-LAMhrNy(KR|i8L%_>{GF>n0Mrfv?P1NJrfyi&$fvjA=Oeh_f|&V z;P^4I$KIffIe(MdgLLBecPr%>?VI(_yllquo7>RbYlY{6GRoonBd+wsd!O;OD}!`$ zU2MJmPF{kNe?IUUQG!>aKe4ORMXlGEcf)cN1!~w7Iqh5u3J>dj2<~cfp6CtRP%(|r z(+d4H7E0`+Da)#mTnMbAUBUM%?&PIaN$uZ~d)uvRV}H&w!qpcR^KecsgL)ofH@}09 z5)?cBQZ-qt7*U4LKZ})A47#+&6+y#%u(Jlqb6>s_*WIrV<#`}?VT&pXV^A$;dB(sK z4(3*Y;;TMf5qt3ojn1@aOHmrwqy)9G<&zPivg=aM^fzcc<$Yy`OQ;Mn=fLQ8Y|`MR zlch}f9DhSoO;Iy4R_5q1J!JV}C6hh~{&lW9?w%P*c8+^+Gu{2*M8Hr-gJ3q@^lVX{)J6FaDKuzb3$(?c}d)gMSO zhf;c+5n{Vul=w{w4_Ib@Lj413-oH^ssj9$516BHWGkEduo44>#D&5A zM`%|n#zS!+higJi6|F%y>iUbv=cB^cb|9+i?{VC0-}8o>9P`jOTr7{+IGyvU2sYUT z4?ESK|0;E(|2%k|LBQF?$EtaQs0zniAb;I0+^ap4V6;lDoCksiHmucSjp^~`#yO=_ z9AEo!(O+<7E)Fqd9>(Xc)GRdx4dZIK2;Z?TH3tWxc%%r_8URv2t-ne=PqS&RF@Lm} zZhYN1kQ}Pm*LMC;!o248msT!;osJ>COGF}Z`x@tW6N1QmxxT7wwz7kR7z-DKejVu%Nxdg6KQV(k6N}|6GvoGp%BD zLy`w`w=x4B4|pE{V_lA*HW9~`HgO3RBy4~0j)(N&iu)O4#yd%l`UFcW+?r&r`yja^ z&hLe7jH`XB>G`iHn+k-?fPRwsw*x32X=+xhUq zZyQD@_>QbGjoeiBrj|+EY-;stqrk~&NU7GFwILgc$%sfEOn>~|vteb`b+c|0=oo+L zDD79F4+{ZZMDoLfy`Lp1s9rwXA{aXS8jmfxG4!oWJQgx(FwxcVZ0edL46!j<5;yk7 z=aNhzrUY@QY_&ryv4&oPKN0KuzFHgB|Bf~)yJOOqMO(@E#SRmsm!9S`ob!#8!N_3! z#Rkq-Soz|#c(AK^DM%Ysy zT`Y5FaBF{>?0^>T0ORE(+o*rqg$Bn>le4PfQ4|I!-ng!q1Oc&#wLU&%J?GB69E5e1 zr#*UKGcXv#hH1G-Zb!3(Ns@B3ZRMAgo;s&=ijjxb^Xu&SamB!AjB08U>SybLhQ!X3 zrogy#iHBxp13_M8-J5j3QhFfQt zu=9VAkI{bbx6EKJzy*H_q6{Xm8aaGgbrVL}w%51=N!*`jdnq2KZ7Fl~JVUI7vKbZc z9rOnrl%vtk>nW=Dgb`iOTsAU{UMP1_Bmy@uu#@9aZQ`p0Q~;SBv_npd5Y38Fyqe14 znXDrRm*89aaEeFA^+ig#ER3tO`Ulzb@kEW(~LzM=lospzUVDeNmf=}=3D*H zoPcTC3k`XUTcUqXYqiic~Bla2+|2%>~>DWEQD} z)#Er1ckq}C4hC+~){CnuH?xQ-p^)xWSW;dAwnltx_ve2jT#1fmD$dYOgO~b+f{SD= zTf7k;7e~4tU+q&456yhOb^I{OVa>gx^f{hI=Pbem)rXihp^jlCY0L@By24I`3QW1@nz^`F z45zPzIT(K~i}#($wEVapEr{J9?D2y3y!}n?%t&)FV)9yuI(*LYuL|$5{pn_sqXUjj z+YdQ0-4iRXm}JaL(xy6XuSdywgj|He^ilgK+kULN7H@d#l5e31m+j(Bp+;1m>KhVM zSXniCaClufuUNI^MlLvg=I?S8M!>&)ax#+HqDCDk!SZqWq ztzUl%#iMEP?0+9F#QP}pYvALl{KiTfZy0GVyTZ4E>gd_<&wCOmGcVqTV;YprlO~bC z1@Pc0kSs9!Pso4{XZUO5T3hZ;(n^k`g1}HT8^Wt}+a*U($jWhVG-=xBAWcQc&kMX^ z?g9ala*=?xX`_X_}zeY!e}Dx-f?>W{z)k&Ds#sZ-h?WFTWtrMz6)>GJjZ z4PR%qKK~y2!j+~0eRQ$zX!AvHbj{FOq}Yuo$( zTT`Vdk(bFXA)f%_3C8ay_r(=-nICSMjjjoC53{|CPu=3=$|0<)PKPVU4EnMd)2)A% z`SE;nXQdiSBVRCCmay_IU8J+)AoS|%kc{Vw8)zF({cKuGYe_O`Ga^R{7J{6rGrw-v zU$4RlSBn+FNDx0LCbkD-S#2JPs|)A7T4jndL&u<7QL3u?ATzP_Brf*xEQ)MzHLuiI z?#2#J+6yXMCeT4z1l5DY`l$gToK%0s_i&WaWW95`EmlTa;?6aHZf{t}+l+BqZ@*TQ|ZLgXTJ{cdn$d61W~+RU06%BBgvRgNExno z+JvkOg{-O?E$!km&nHV#Z1oCljvD&(Z4dAex(>(`^+dO2k>S~9JNXHJz*{nLE5QC* zO!#<@gZ2P5%CJ;zih8qtvh{!X-RiD@?m*!<%@z&`p~%5Ssm!~NpcH1o@L2_BW|Ay4V=NYH(%{VmZwr!1n8+4JA$tkIZ3H9zMOFxa?9yaYQ2_uuS*R1+OPlsgyomC zJ_g2F(_l0GEI8JUKytR0IkY3?i)!r%i5>;W)~P)|C%^0wKFksap+0|g$g6@_RdrY8 znu3a#*;GZTx1G6yFQ)I<$RfkEzOwI`#(Ar$jXzc>HK;0om5JRa`?j&#IDRSoX_V+5 z7Q0{EXMcHb+w9U=nml-jNHGqFS1EBvf=Iq>xzMM3D z6E`a*&@uRsIlfN}O{JQf&dk(4zvVrO9*5~|;lKxUxdpr0X7oHPjL?t0WhLy0hbGqo z`H|1i6ie6wd@IkO7KJm$Q5a?FI7%dz!PB-CWoj|SE$Gkk1s;Ez9y-^g8|g-d=v)-w zo%)5)TFQbOl&+c?`KJ?#3*8L}Zi@ZQZ0k%}A3mUW(>d&Bj;1fC0v!mlu zWN%_>3Nkb{ATYN|W(57?5IHppFHB`_XLM*XAUQWSIhUcI1r!A_F)=kTmtlDVCx5kf zby(Eh);1~K9l}WG&?(*B-2)8G3^7QjbO;CnQX(y-bO;D2El77GQqmv|-{A9}qvu@j z`{#GfHS?Q$<-OKkd+!;>CtA9k(zZ}*umThU|%mZKn{RUc4PZ*~)$P=yxL+l+PU{<(`EY!`{!_nRW_7H=Qlk*|c zgSQM2paQaShI)HCI|4uuTYw5s9SG2Xdc%c|02U|&U=4Nvx!3`qb^txF5r05mS6)XK zps1s%ucgZhga_93a&v=v{0|pdT|Iq84uG7rhMqhCY`_6f)YsK}_|yYK;QZ}502+F5 z{X-tO;e)%nyq>h4v6eg!*Fy~e9>6oOho|F1+W+Fl1ZM{LLk;d~=K*#7odCe%0E4-S za&dWkdjsvgJYhhnhdt2E<$pInJqJfmfH%~`832EJfL*}98{-ABg|`#t0RAVy!|DLk z9Bsf5Pw<120`woHE4-C(C%7E;Z({I9z#d|{{9_LA1cU!0jRVN@x38L(mKwknC5jm)1~Lkk{4YRD;hRgi{>~ z?+yeA^MU=geuyV6rzQ#z6yXQ(z~2vZDi5)hg}S=JS$kqWv`@|v-X|E;!_9G_;6O(^F#Lh(?+JPa2EaVLz=8gMEB+H<@(2NJ9c^Inp@(lA z%-_+KA$CxJ$UkB@lYgoHXB=2~f$;sr3g5K0P>72!z!q$W$$zB*g~5k|<^Ru~`>$LI zUM?;gAXhNU|K8aDwg$O6y7>Mt_WxyP2!7zrq5<`A1-bl}&e2oB(Fbg+kn5Im;Z9|Vu7{s##IxHSHagn0p6I)8rN#nI@Xw!e1uv4dLPFW6BLbmU!TD`0HQa%>N?+lUIQIVF`ud zBg`H6AD%Y(R zg?}qAJ^01|s2`T4J`+;bPRF{f@+RN7gnulnCAozLpg*`HE7JPuGJ$vap5Qo*&aA9u z;Vr6n(F9K;v4;zx&kR0g8`aE68zSufk~lFQ)FnO3Hg1(4Njj+V$gT4e(Ral7BFURL zmhV~ITS)N^W{LM*Vz}(@-8nN=x3D0>{eNeL%8Q&s%3_#A2#AM|KLFR-wc<}Y_t7)# zC_)V5WYeyBd`}Rc5g^b%X2Kh=+0yfu=n!Qeg(t1wKl*r zzwNm|BerEmsEST^a1LFZ1a(bUWHwU?MG1G;JZXUT>$H1tWfsaV++@>Nc5=xnXeM-_qLXTXP&s zEE|AgTyC4svS>6q7*oP!g1Zd7UX&#QMK%ixkFqqe5K{8Gvt(B`@PouhP=9LAv}ih= zM7P{uBp_AWodjdAPD}tO*B~fv;cI&mg4PWPG3)wOceQDXO5GQNUxYej8dTTc`!0D^ zXn5!kcO2uorXq}G_06_&9jzq1K*Hl|+d(ggwMZHo6OMGT`*c9{6m)E$HZ+4ZqGV(T zYnDl#_i*@T>;RcHX1=}{;(ui@WOoQc^re-Wq@>4sLwGRb6r7Zo9)AqD;!&yNNxy>x zXwrT8q^5pe(R;AjQ>~Msx^llrMqJ)J{9M(z$@tuGgGznVG~;6o{Cw%YQhd8%7mr*|H53HXy-%&OQg5Pr4lP5R#QvC- z{yA%f8%JaO6~ptWrP`~?G_e?5fP7Z672)u}CuI!aR7wq1t)`u|TxA6p1=@#qr3A%e zQ<3zFN|nij+V{4rqkn3mEasMjA5o?Qzr3;tCr!-k*#R>4b$hSu>=->q#lfqOjZ2pe z$bP+uguacjIa?8k6Ol3PkDSsf`j$7NR?(|S5@^>|(zx(lT{SaDM+mJ~1!bs~^nSxh zfAs}c!qd9+_?07x?4xN^(t#)zkLY&Dc)su}GoBUkIapC9Fn=DmPJ~r&66H+G_$p)A z=W5;0lwLG|2<^7fA>ao?!wI93duTy;?ymK)awzx7<-D`i8G*VkqEx?TMQYCD;GHU+ z*2jBYdM;8^> zlfQId$-i{1u77zPM*Q082BJ~KILZ8@e5%vpF5hP(YhzMYKo39L``t7(>tzJ5`|hcX z*N;wJ-%-WrSUs^slI%&8lQ>Q07XGxjS8)YyO-k~f(42jw5j%c8Z&o?rP&al`PK1>8 zsFcK0X`disWsWBll~3*$a;T{Lt}@ELSA~%Qyj)2yCV$)34b-K^Lj4KI`)Q}vk+tKj z1IXnm2-Udejd%N2zLic0<-Qo+_4w{W7W*V}$+m)tDGgJ`w(&CED~z}CV8c_Le0jFp z&L~ZXN1B!Z?<>ue7cA?dZAn%~5KxYmU|eARx{CPL}!2>jScvT`fl(kEoLj5qm3V9$KC%jD9H^vmsoOI^ap_g6 z8(pwZC_35;{)v|f8ls^M*%$d!8ZRRjQ)wZMBMF0DRUevG8Z+1ZuR__0DtT+-y2f}Z z*8$%s1*bS1QCvfoN`$*Q#8FX@0(PPdu764@RE`B{s#S|x(DlSaZFyP(^IK+LStfys zvR|o-^hEPhBxA^M4D3QEOJW)dDz(xL?E_OPR~7Ce-sJk5)hv?zELi0<44wP3lgtafeXZ{uRHDnR_EEta(_A+ z6{a{gfBIjk7L{LhEB(-)#&?IL)c059U7Mw4`*2RtJ^t06VuNBl*fg z=?hLtvg%j#C*x)g5$d5|ksZBU^Ek-ND&C9hh~xydzg&d6Irj@PD7DTL9R~%!jdMM0 z4%|BuDE_Lf_E88Ur70^pvVWMR_I*x(Y>-@kE{ud_FMIu?uEGzAQ*+e-#7<#QE5sgE)2?$ z8){{i)8qdH0jkxWAz5Z`GsL1*VgTBQ4kAj2J%3saPZ+AFq#8<8$U|D2(OVYUr#?(sS<(F3T|6Dw^Tw$CZqn4N zoah>u`z4|+&#y;fc=_o`M&;PY?Pvljf*M{|>8|p|dYV=zdm1bJ_^;QCwJjBihy7Bt zwBj#)C!l+v82YKwo|6-cyB$QZBS2LcWD^2Y(;7N%EZCh`(t7Dkqw$_}UrYoU z#FP}6gxA*{7nk{J2fYgH@8zrBvGWZRs@~mmSJIH+bm%v-ZPJk_)8uyHYU~}g6V+F<1Ppa}_I|7z6#%;eaJMCXIN&ocF(0%5>CFX@j zn(QhIy;SY^P?bm!fz*fj{fd5r&=yPM)5nEC+syX!?tiR0%f~s7u;*yfc__OmxQrO| zx2|q*1_qdK?_SZ`Q$>A|z%+?x-T4x@Eh+H+S_~#s5l~UaNF5 zS+kr>8gPG}euNh6y)vl?(1*QP%X&MvJkar3x8U_%7tYr~5Md0rXpNkiZRBw~-<%+K z-^;{CH-D9?70whcE-~WC)wfILf;ebpiq4x9rmKb7&XJ~GJC4dJT5?Y_sTZgi=Qx!M zy!sCwD7yQBgaGkoap$% z)5~h2dF++J_>qnuc90S;Ui=alZ$7#25skziKc8Q-=DiTuN0*g4rc&0FN1Q#bo|*?j6ritk^h z`hTRpoD_L1((>UE-XPngTF-MuKB}G5B&nxeZ;nZfTL5FLDPc@A#=UYJwiR`0Ng`+8 z;^%+8N_7@atwszCH+ece>ZGy>z8FSlIDPzuH`?$Wg4{slj3k)qjHk zpl9FxK4=1%1w_Ly+ra1bm6Rv8JzLl%{_x74`HU*kM$frm06z+e1Ngxzk}kvAct|a^ z;2AxYyrFa~2}=tAUoS0?Y+dSQi~IXm=&FYjr(Qu7U(t|Dx3kTFJ+d~;9O2aLNGZwV zJ;ZAB_42d#GT1(?n|x1n8oG{}#eW#9R-z{FP&{Rkiz7%I96igLoE+{?JFsae0w=n0 ziINsb#b6}GrP>I@Pq!#S{W&X<>Hv@)n~%pGhtyZ$%Z3iN;+&>$>FzlB3`wyu&`}=Q>||rhksf$G|SL zUXV9>i1qE|qtWc<>lyv>R!*ykE`hJ%O(vW(PMi@x!28YOzRw4GISrT4UU7&y8GHv? z(NzYoS}JY0n?!Tj{vs z*@}1o&MRw*yIJ#i9AFo*I{{+TesG%nL6}SjpO50o7TtXvgrbe_K%-lWP(zNwxav6U zV)Y3x%4pyC zq2HaDbh`1IVSin5%xZNql9OwmR3lmqAzknLoLLe(v?-9GIkwM zQAO2H)%EfF9W+wimNq70Hj39f2;UdV8O*~><05HWD>h`wFZ#{t_JhIsQ5ooE)qaqD zgxD6h8B7j+839*?|8zm>4OMQqRBDk{auOc-&du^2?tf=BX`Z~Sm6;*}4_vLHVb&5r zR=zksCh2)Gn(dxSRB=dddZv>CxyI}VeGz9j9YNwxdU~HCj8JuAD8c%~THB9-8*kjc z9!PA$%x$eM6|PB;hX+Z1-FonWe3W`W?bbLxx6QI(_-tO2IH3iyUoOM@#EZKyVcB{j zote?M`+prd?{&Sxh$qs3a#^&MfrCe)PZb8|`md`~Stnb+GJsQ6U%xhj!Z%ljjvg+X zBPy&w?5~Mt-+t(rgd#QUuo<;y1jb~89--w{^8{BQ_fS@S2jp*Eh^Tnj+SDF2eO~Vs zj#0LqF_b;iX(kWwq)8UG``T7fhBe97>!j@eLVsTZ?et33GCt{Xq|mzOioaR*J2Zl$ zAj!hDAvwWjqUF*WP9U&8JYnLb%CK{<Y$tz<*v z%bCq4lP)Dg98cntG73yug6FODSPluvg8TKm6PnG+Jp6NJl`@1nKM5sP)_Rfp8{A>3 zV1HG~BI#2e`xmE<$SLncP>1^FWl&~_;D8J#T6}2X!6qxy$pYFLEhrX+x-Wp*j1k zjz+ePITx0%gl;xMsXD`V@bgD=<7KV-K7U4Hv(BkEDK{*^(rzA|o^q1(JpQp$E4rg& zA&%~pq;A2;f--5FA6}Zem$%hzu;bUYZA-b;(1I|vC{rSAp?TY8J~>T|(Y49#EQ(TN zhDREuf%;vcPE9fIm9*}N6bUtXRVYA~{elqYejC_SCQJ#Tdg3Xx0 zewsz++4)*hcdrjnLPs-$Q}*PQL@@8#ttR-$l07-R`5Z>~NmYm|FESj)9e*5oO9x!8 zFyIs0=tI=IcQ*d|fpgP*S!}d8P1@f(&wg`v=krqn1ZM~PKv7gYb9QeXDoOz@PdNy_ z5&FbUGP4;{U|@}KAlmsXGGR!I7waoW;Zwc|+|AO4V94=yrKq;A-toa2VWV-@hq4=r zHdJof2nE{&nfWcyrFPR-B!BxFF{DRFOj?*eowsDz8T+CAotb0UA=q=qqWU5D?(($W zloiZhl{-UBMpG*F`d`a|D$ceoYBzbTHp2H+#VL8-A+^UmvZp;PQs<_f;>tJc&tolz zh1GgQa6tqoMwV**f>u9MB3^NoIrWX@Bl1q!5L^Tio{npIQ9GOCaDS-e61xOzFp(U( zR%bmg_-R80{`}Tgqhx`UnP1K0o2xtpfSY@lBi*88vEqy_fG?YN-4MDQ_8IwpppMkR1L^DP&Fnae9YP6fSXF zh->U{^>7MzLD4WTf`7Mvvs(s)QM)27U4np&J$5VxNMXKqpdIAdkcLesa$bXDalC`AkZW&AhrZqo~ zT<+a4k{A+8nDucZhdq%^NLEqNrAZ*lr>Z=XSXj=EdL6?Qy;Vl4?T0;&etGeX3>k%A zo*pB{?GY#Zf2QfmBf5@x>?6c!*{|lh$J&KFUt|cJ7sIhA-@3?RieB$~3d6Qk*~&?T z%o^GWvyVylNq_rrgw^)-$$?TUd0anP#keh&(6kG=7($eBDfD-hqARQ}P47if2@M`S zm)@f%sN~5RcX}2SRlNOV$;#^)0Uk#}o~e;kCWWKnV<9bpODwHP%6<)tXBkk67|kyg z5eVPBE9gr{kjZ-MFkyETSHtg{mPHOZ8w}kf6D%X>>3_I1e_~cCKf}q#UC;-~txYTR zO{+Pg$EzQE zCV0C6SwDSjh)HhNk+!5fYS}b6Y0m7t#3eRiyhJ4D=*|30@{>URl-Dw0ZQCR5GLg{1 zH-p%FekxOptQd!;CDNzA(uGl>erlrfq&?@3JBC5g)kg1P4IM ztUDY~v%#dBFT=s?OC=qT+OeoyG22@Zb$@15EIMy>aBQCT@rA3d@-@{5H`!j;t%XY= z=DJ-84jnA|O4-EWhPEz9VLk#pT_v2kkPqfVSDU=zk6p6azQ%2{;W*iUUH@6l-F2c2&G#BwybYn|^cZ z5jq)Ipa|yprG4Bw$eR^E$ zh3Vn(GtEk!@k>r+UxhwVDBxmHmXQh)J8|z|;FIfBu1?qTj&J^`fKKoo$$z$1|Dl!0 zX(xF}KS>4R=}R0+hh7CRgv_V%=Z{$LEBBi)6t;m*;SD8^$lRoq_GP4=b;jXDFa|4u2*7&A8^Z^&j+yOWbFFBR@g5hX!{7mP7jEwF4whR_J68{n&-~j0HuR3nmF5Cx74FklYyCLZ!`sC( zTWsBB2(a#jc}%9>HF78|>;ksJFk2gralg(kP4gaGoSMqjkg)Yh4^7{Uc*sJAkyI%qJV-IteOqC+o<$q!!|1dLxbg5clWzg zZkRuhgm(#?sS#DLf~|H#-c|FA63UQ+OkA->^jDJwkGF{DlnC?LbQw;NH2Rcj=8DPO zTGe53aTnpM>h-jm5q}wPHh(h8j%0t6P^i5|?PCI_fHykRb?`XKLj$)ACPf@&P{^NH zGi;M1WU8hIyfSPllfc#^={`L)I=VyG@D_o+#dBZHh4pjSeYSw$SR%Z0RqYSTmy~r= zy=`zACfHHQ=%S)&^DVyqh@vBc7t&I98cu+)S0h+Fk$umtK7SqD2GOwq){T!#Pn#*e z$Xlcsi9xg>7)DmZ56Qk)ZRcr3-3w14i~0jSs6Bl~%F(9ia67dvxuqdfWv)1t{&x23vY=X zv&{k2n3)DyuC0XJzq5321UsMh>$l!$&E$}^B;j@GAh&ArbSZ-zV5Ivq*?1P~AFDE< zo}V)a-G6LzUHdnW)8e!ioeiEB{ICXrgrg(UKe4~TxTR?yZhORHC9366ZtDxp9#IH@7tkg9)gJq$AR>?#* zHj%+kupE1#poE1pd;e{RJu-#@mTX4jb2L>a2#wzR5|GGgm${`K`YYH4rWd%cR;cCw zo_{YX@!iNLIkdiaGMm;coXqXg7sr^wRC{vf&mtM$hTRdd-v=BYw!azWux8aS&vuAk z`KD!H5O;7MUJJ<|mC2Aakfku!_4^D#2ikg39AgvU1vy`5NFc}T67nOnUbcHu*`@?t zpT5X5QJrKZz~3y^ms#~!M{CM8`tc&}=znqd7J$@&WYGjv;jon1ETN(d@%Ed|8IZ+wU;MG`>k+^9VJH%w z7wBN0Kb*32VqX?^$MFdHI3f|nGkRrTjR zzW1A1&_ode#fu$&*^ee*`?89be*Xv^7q!2YsdT>^&s8pL}<{JJw9F6-_6NRy=U9(ed(L* zDV1$>jnJbgi#%LwMRt>%Q?H88qg_D)5>Tx?p{;sOIK|a|+VJjE+J8ia>=Lks!&?{y zqkG#ir|P8`{k1@1^)cQ0iL^=2k+p&Qg3rZZ#fJEsT3Z3-t|o;IJEF*rI?Ti$Vj%NJ z4Q6%RqB~m=ww~ar8B@J8WV^?L!yF4SekKr!+7>6>QC~HI^9UK*E@}+LoU_@WR`-~q zs}#KGMxXMDeeUHqV}FJQ!W=FbEORNOL+!SOZ^S$}<%(_{;eXG(#4V3#Ur~&!HXAly z6O(bq)2|fs)==-?=Mrr{N-D+Ce{bkKo;_&0Io!&s)#6PZ<+M6WRny#cX7YtH;w#R= z`=-r$e__=&Zt0c|34Ckyuzrhg8IL|n#s@DT{21rMlux* zKKta(#yk0o)z`jhc<Oyo%zFgr|`Fknt$&m=q) z%bL_!<4~PSV=Z~IMnQTQAi;%UQ0t_W zErvCJ7KOkX+qk*}50Hf?G|p+vg-tpK3{3&y+8l? zPG)aQo3Z;LbE6*5XY<9I^T|=y^nG4UR;zAt2JGeW>E-I#o7H61K_;x&Q~UF?RrBCM zqxOL{b}MZA5Gtft%RI0BX!}6lX(?lWO7&_nKYG)xns=Dz$$q2Wc9*MWv-G#Wop)To zakqzsovpgF)e?tPrr1-jyXE|1anvovC`i>o_hCADG{0=#af%2nOvS#(G?NAT;*4kE z8z2L7H9*D+GD(LPx_u4BUEYdfOs3_z9>;8SYp+0lD>IHyYhftHG3(S?KXT1~KG_z= zW+byOEzavmX6;)Zn^DXz>6TV^95a@Ya1cf@Tkj#j{7qGRpu#(xvo4>|@I?W*oDBGp#n|TdLz26&z~XP#uR_x1*W}+kC#JjzitQuWBZH zO*NCguWBYUUNv`lKV!0QVzwDZ%9r2Mltr zkXrBZIA#RIZ8Gk6J$J)nG?3R$L^P1;?s*&t5{0jnWA@kNnC*S#nB91PIp&*!&tlr* z-8g1v;Lcqd$83X#8Mh;u;iYmafvJcW6a@VdPDdgo}JAx z)H_Mda4pQt@cYWt!nda$%@-fKg^a1+tMAow^%xJ>;d`#)XvM^Vg)#L_fSokJ$E46o z*2U(w*a*Hi7l$j1seYKA{h}T|cpx*VhexaF{7k)3e|!C$pFf;`tXAjC|5ECBx;nWy zY>(!r>ci>mNL?Nu&gaY3AF;gqtkhKU3F?k|_4v{%=%4tBi$oNF$B235{ z|6m*^(IVx4z}67Cli)pgs!W+^%VP8xf#6&)wcj!i?YKhTdIJ1*qC#+BZ`*3MF{EtS zn;vKR)?_ zykXi*AGGMY>6eP-4YoYoglta+^N?8BDKqkl1CAcRT5GpsWP!BuY`H6k!`Xwpi;PzV zGH9(EkTS#;oahg@tej#4Ye9uI*0MPls-1I%+Aigg!;R%J)EJjH-M+N*sG<9>+gpZn z@3!oJblt7%HB0Cnot&kGS@(gt$vrjN(0cY?8n}8fm^~HE>*+wVhcr|ZE_7J$hH6li ztI|RvQfbxRUWaWp(Lx7f4(VE%w>s`=N&BxP%2nyGSCm~}r&(0@IyF^~`>!kqr>8XD zG}~f#?xwjn+n=XBe?RtSj%wuVyn>}M^3cP7-dlXBGSws8`J|;ok)vFPsH5ISjhuQK zQzzCuiXh=flh^}93z{n9UE+Wxl8PU5!S&vHtsGH{CqXb?xkJ(%3E!rzb>Q?o+o6PN z^ax{MYqVBIj?&7SY{jFXzl$`VF?f0n~X*)mdp zNscxvdU{j@odT&iE+7dE14WR15htc47NMWd@P!zs?#c6>BxXJcA6$YJx9BDISs#%i zlE@N8gx={9>!Hi5bc;$o&UC0zUZmo{IvEijT;6Z9K69l$rCyG%s*S`; zZkmxC_a8L(?D>WPCWMiRkBqr`$CgBYg`HJUoI$&-2{O356J#L3z~HWf1cDP>2M9XI zAVG%@5`rfLx1b5`?he7-LXbd^-~UXV_`PHbN zJdjRc>2hVp{^yj*Kf_J}@|msw+AA7#Tv<$Q+qsJRUaGdzPR?-oLHxuwJPy_bM%MZv zfyP$h1bad=+HX_ZO5{5YOHWr-2)MU)zF!6zP4pQw=0&x{&!4HX_4>!aHw*H`P8E3~ z5Dg%<*CaUe>&o^oI9?2aNjrdoqvtmG|HQkWR~yj??a)QwROE4M;naT;9X5_4spE3v zzF@71WTd8?%>D{qKpPPHQUKo|D?a8}Ea{gYpv$?QG`)Ok*FoE^vbake_oI@u^sdQX zg_m!Rs4197bkTeX@F3}&RK z{(1sza<6{}lfQ_Bq)$b;F{S_8=d%m8Q5=-SITQG4pxg^N_YzVy0Vzsdf*q3BPIFtU zjipd^&~esw?xE5MESu@jlv%`)SZ~LWd-7w~nIBEiE?VX5FK#mUp2hVlI6&vdEhv|!!u&TbqB&Zyx*W+_lq<+n`I*kqc%%E6Mma3N3)X zuSY{0o*TD0mCzTz$|FEEEo5t3dKk=w zz04f%o-L(04&d&{CX)&2%U@_nV`?}It8)+(ZSstrUZ808&K^(O%j2(A0os+`e$%&f zk__jK^nOVchD6H1nNEFzNdkPpAIWMmr;q;A{NP|C#?#$aH0DNgDN50R3W8KfD&rN- zBqz&Y7-)7>199+!A?~#%DxCbAj>EGi=lf2pK! z^W4m_*hH!Sa6%COS8Zx$*K$3I@Mw@hXNg6eNCHQ8!by{t+@m3Zk9ipHCr z6WB(Y>SPibKH)7)57DX}x(9P|MkW_`qV0&PbPImk74CG`$ADXopyDT@FCF-OZ`9u+ zS^{MFx8aW(@>1psjtddIyT;WkQ_uC9?t2_peJXG-O3(i!Tl!|^#sCb28I7U4Us|py z8ZX3*Agtu}vXwfr_)WF**8T`S-$W=6wmDpB^(>X3E7W zfk5fyP0xW&XGy>eB#vtKd_G9`hj~ZZr~K_fBX=O^@4j{%F;TXhARVM`__X#kc{{#c z=zbz>M`|?fqg;z>{Z+8Gj5*WkxPbJ2c7U|RLuV#aPWi6z6~!<3vTykNNu$K2i`|W- z({dOg_gZ=DY>wmN)!ZaX?A|vtAV|>UH-NRsJU6*tZKgS=>XUb0z&Gy>l*cUP*x;A} zm2Ee*SASMzt`B;89^I_}ZLDA3dVAVm_{Qzj5sSIc{ED3$&CI62IscdI)@B*rzl z>${8Xxvy@qU3UD7`KQ2Gytw~9pbWgP!&pLG+Se(kHD4pC%h{EcMIvP&-CimrQ07HQdz`s4Yq4Pp(XR*E9o?ga`-BWKX^gOtf{ z#Us62DDxxG_G$w1vrAEz8kTED-Ik?BGLC@0TuB$qirY{1eCd|p;=-F%{&lKq-S8fv zAJO!!!Ck;KWZUAtiu#*7OeSe{s%9*BBZc`31zypN4I3DvO26 zDxCf!UxB3NzjmOO9!MN;mP|*b1>ZKyPV2Pn2}o8_o`BD(52;I*&%3(rPpz@PdvMma z4$fbnz4sWuH#El`-8()8I)|Dcwzuo3-I$EI+JQ3p#HuZ3O9!1MW4GZ&AJQPbG2k=P?7WYZr)sg&y27 zg~_Huck8PK&hwzxHSMGN-bJ&m?XRwIy?BdKjoJtIt*wY%EUTWBv=Jb%qL=GlSv)pOC$ z@vp2G4ceYNHw1Lvx6ZG!#e|WoDCI|zpnt&CRqvf9>mwiyC!I)pORmsdO{GP#VVcYS zbOoI2B0nNQ6K^fgJx%T>bqW^Q4`qDvkX08<3JhnlnM<%1VAw`KC)DGM(uf5!Y#HXN zv1YxyN#WwN(00zp+~vW**??m$;{L+aCd|j!(t1OxBJ7pG^`d?%GSm@@rkLLl`Z*Nn zy%cAtNM0vUOT+X_U|&JnFsflzmO<3eaA~aBb^UDCQw%+96cSi?QpmewIp45mha|_=UH$N*PY?R_MfEU;_Il()z7eqx>=fcjDtXG8wU>eF)jK?O9L3 zgX@DjZ%p6HIO5JvRu31QfCO!UZAV}_A+>Tex3c6Xa)Z6&6K$p;<~BM{k43gK_%`eH z!sozq1yMdV8;dxmFSEWTvh7v;NVWE+QAQfcXS1x#hwmelRgB{-lb3eMS&ydqF9Ddk zXEicsqOvxjJueZdG74}?QHFnJ$R=MGW=rj`0#Nb3R+F-*`?FVj7H6@(AKlw%n6l-; zZ+-tobkv>sObjoI`+EMpzN)(`@UpM-!m9CwDnC_Xru$#6Ih)yF00zfMaMoimu{jZx9+@|2QGLkstQNu}8p^$R*yKl^I&B zLTRHI5k$IC2bt;#A0J;TqQ&8x-e|e5PE{!6rVk;}7+(SVw^rw)#lK4l+e>a+BW|4hflKk$bP6wzKwHV~-_-}X=w3vS)wFJx&nB}j*NRmJ zzLE@6-jZ3GOr?UJi4(iPI4`4JXYAj)(%2Q=XjYzQL#owuj1#r>WA^hfWzifXhtEL- z^Y-;zBSEIpGe}X5(ICNW>b=|SiJ9e~k5rnXRt7O$mR3yJOkztxg{-zsNIJ>RDRg_n z8bZM~p0kAP=!oze%_b{@fU8c+uA4j5?2!W=lpwTCoEG_EHWdjhd$iq*({9x%#1n#N z2ThhPyfmkSq?k0^&QD2F#!rq}BEG&wXx)>9udPnEeOjaRNa#ao)I})7f3U_Y1o*+P zvLX*M7zp}(PD{&zET+oZ`}8^)FdhBnM)$KlQXdkh{qN;}Ey?|ky(-_wGqBBa_`H(S zV{Yzak{0nqU33u0?PVt@@cuzP+jJ4h1jCoi>r%)GD6N3@D8gFv7jFv5h*T7Q9;q&gnG3r=0zh_;CnH^0UiXA2^N%v&$bR7T8`iJ}C+<74H;3q$e-UdI|F^gvEj+C?Z?0bfdMlNoa2uN5*F7Oe^@|oN zlYp|HNOj%8=K1GyO6WA!m<)ynX7`q82&1|A#Scr&Seb?joJ==cA&YNcmOzci5KT-- zjeu~0qT>qYRD!Gg!J3@X6u~k9L_@@>=}ARkcwAl-O@>g`RAT}nhiJ{pDySH{po#d= zkR%)Cvzq2D{dEXEash;6606g^$R^z?8MrK0MQYgZ8pdgM8SASzWQKO>tK)ZeeFp1# zqT}AZSlOfzftHSs=T=J?(v`)WV{s&gmO+R%K1%<_X1V+7@B@GFi@8FAE$Q-rQmm;U zRVROYa>~sYq8qc^kyWdp_Ftk>0_}!QJAmMpHt1r{b^RtjypS%Bl6Sq5@%u?Ju$q&; z8NR?B*j$2&c8=6=!!Wtx{7I7@ARxHc7+QXF;}*-iL^Hu&&VI`=#5V&Mjc~oN3uUek z@Q~?4J-HwGnIk*y4iCC2ODH9aFMZo*_KVnnLT%om3l(>|e-v7?#w=cT*ek8i2$V$= z^2gsEQ#O~n8p=H*aocau=m<}qpO!hmIx&SouMX?n3d}1UEE>uyX}W(jE0X`+z)JeT zpap6(;y`Hoh7eX=?|d0kkvnH(rC&ruyv$KdMW3qXrobWD)nW0qW+^o+I62p~RC#no zHZ3lt$IEWYQy_c;3b(@s_V74BkNQKBagwk(a+fDAv|_>BsUiZ)&2UeT2QPf57YkEW z4W^pQZubIB!u;JDu5pG-BUQ>bf4zPAg44YzWa!o|h>u-W%;Twq8U}SjVi+7V*TBl{ z^NbWv)@3ofr~Z6$0oYUjZmv3sMNDldOv#OI7#vT(Jf#3sn8n~5SvdIsXfRLnGzqd{ zB;jB@!y;PTRx^q8y57@0JjTj|Mnaw?{~1%_7f+?$Qy6uBu!fH%DXGy273h#DJ`Nq* zC%F}(%<&2SGZC85U%zIq^4wmR>MF&`^-z`d&@q*7p(A!O+RY1kMepqKl!;@=t4yjY z<9&`5qW~;dt{&k;Ng(O35fqmr%|0qp7qV#n%qO=uNm#Hx!o;4g89IulzZ_)Pj9AoH^dI`9zYBG>uej@{YR=A%N^EX1>O#VD&3gCbU;6gNj?4 z{faDAOxWm8t!vd{8$&49V_ZiDc&&>(S0e|s^nEd` zgKUlQvrhIQew}rLY8&)4ap!|#0!$H@2ZaYJ6~1(>tHmh&O(X(kyNUEbd2I@mb@nvW zu8B#I0)e~qS$0{?s5<-xF*H@1S20xVhF{f=GIL^KoAkM{FpOq1Llj%ms`rgOjk9mx zA5Gl>1l)G3w4uCePXMv4S|*M+5~qRmFZ4_EcVZ(_%=6(TX=kT!7>=rpdm z*a)tpJDOV!vN7%SNEq~CQ=Q>$Sf2f%#v8`yL5Upen(fu=_r`SJSF1zO80*lr55#}3 zM>J;Pi~gkdjtbB>zVgr=G??p9n>%_wXC|RL4yG4FY>c+nTe|-|&>-=7)G5F(=MozM zkM=+`i{%uTn@W`f=X#$UG9MSR%Ot+<%i|wp={lQ6MV!Yj|3VIKlViJn7dXfB9k>>i z|ME4_K1y(9_s{C-Ll6bjW@KmNJRzD9gn;`jFJ;#EW)+eh66TpBPu&2|%;~uE!ysL% zYZ^JGPug{-*;XHl3-%r`a+;K0K3Wr{dcttrbd`h@lq3{|gvBLb$|6EAK?wyBWhhih zSpg;@ru=wqmt~Ru|CU(%-{2`0F(HXm0VSMg0OMSnHia%l$f<;{gT3k;R%%ghNNcrp z%2goTKaz`u<*$5_n4@i8O?CB@QpwyF`{QR`8i*?|f}o~pMlQto+j6PESCuQN@i?z^ zDWNtYtA%svlbmyD8SF%8GUsphKbs#0*?)4#ey50#m#&&hA+niBtj591199ilMQ`BT z06~T9FBKpUJYhQRf5qA?F|l()Agbv^y-&pnaEX#4GXvcmzNN4tV{#BqV(N+lN%GGe zk_3_kmTJjHHfXR%^3Y53^j3D>SWG_lL4cPX^fBfl+g9Obhk>6$5FZ(G-ip-T0F3xTkvV75Ef@sqm$`GL9_nQwP#_A2W2(Rj zwED!&LvWF8ZLdS?V;c4ge~O?@dnbW(Z#s%CgYqA56PodLD>q}NZs76ua`Q$7 z*12a^_4pWvQ?k}KFR=fSaWY$gV~E1>8j7cBCa>-|4>iCrGcyVA9x0|d6%3F zb=Zi@%Qg-ZNA=seQbrE;|4<<3ib*%~WGM2-{gP;6dLugFv;Tz?^Hk1E%uLML9}ynl zehRyd>-v9jw}+`c-nlHCD5DlIO`Gy3(jom&Vp3qKj#DPf!Rh3!XVU*;%o{gSB}yfA zfEjp|e{+KFL2- UTos2IPgq18kAp*5TLtgG0H?ofSO5S3 diff --git a/doc/src/DecisionTrees/DecisionTrees.do.txt b/doc/src/DecisionTrees/DecisionTrees.do.txt index 583e64200..1f829c9ec 100644 --- a/doc/src/DecisionTrees/DecisionTrees.do.txt +++ b/doc/src/DecisionTrees/DecisionTrees.do.txt @@ -1691,7 +1691,7 @@ o Establish a cost function, here ${\cal C}(\bm{y},\bm{f}) = \frac{1}{n} \sum_{i 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 optimial values $\beta_m$ and $\gamma_m$ + 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 @@ -1700,13 +1700,13 @@ at the internal nodes, and the predictions at the terminal nodes. !split -===== Squared Error Exampe and Iterative Fitting ===== +===== 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, we need to optmize +This means that for every iteration, we need to optimize !bt \[ @@ -1727,7 +1727,7 @@ and \frac{\partial {\cal C}}{\partial \gamma} =-2\sum_{i}\beta x_i(y_i-\beta(1+\gamma x_i))=0. \] !et -We can then rewrite these equations as (defining $w_i=1+\gamma x_i)$ +We can then rewrite these equations as (defining $\bm{w}=\bm{e}+\gamma x_i)$ with $\bm{e}$ being the unit vector) !bt \[ \gamma \bm{w}^T(\bm{y}-\beta\gamma \bm{w})=0, @@ -1740,7 +1740,7 @@ which gives us $\beta = \bm{w}^T\bm{y}/(\bm{w}^T\bm{w})$. Similarly we have \] !et -which leads $\gamma =(\bm{x}^T\bm{y}-\beta\bm{x}^T\em{e})/(\beta\bm{x}^T\bm{x})$. Inserting +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 $\gamma$. 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 @@ -1749,7 +1749,7 @@ $f_1(x) = \beta_1(1+\gamma_1x)$. Doing this $M$ times results in our final estim !split -===== Iterative Fitting, Classification, AdaBoost ===== +===== 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 any of the two values @@ -1940,7 +1940,7 @@ plt.show() !split ===== Gradient boosting: Basics ===== -Gradient boosting is again a similar technique to Adapative boosting, +Gradient boosting is again a similar technique to Adaptive boosting, it combines so-called weak classifiers or regressors into a strong method via a series of iterations.