Morten Hjorth-Jensen
2983 lines
153 KiB
HTML
2983 lines
153 KiB
HTML
<!--
|
|
Automatically generated HTML file from DocOnce source
|
|
(https://github.com/doconce/doconce/)
|
|
-->
|
|
<html>
|
|
<head>
|
|
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
|
|
<meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
|
|
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
|
<meta name="description" content="Week 35: From Ordinary Linear Regression to Ridge and Lasso Regression">
|
|
|
|
<title>Week 35: From Ordinary Linear Regression to Ridge and Lasso Regression</title>
|
|
|
|
|
|
<link href="https://cdn.rawgit.com/doconce/doconce/master/bundled/html_styles/style_solarized_box/css/solarized_light_code.css" rel="stylesheet" type="text/css" title="light"/>
|
|
<script src="https://cdn.rawgit.com/doconce/doconce/master/bundled/html_styles/style_solarized_box/js/highlight.pack.js"></script>
|
|
<script>hljs.initHighlightingOnLoad();</script>
|
|
|
|
<link href="https://thomasf.github.io/solarized-css/solarized-light.min.css" rel="stylesheet">
|
|
<style type="text/css">
|
|
h1 {color: #b58900;} /* yellow */
|
|
/* h1 {color: #cb4b16;} orange */
|
|
/* h1 {color: #d33682;} magenta, the original choice of thomasf */
|
|
code { padding: 0px; background-color: inherit; }
|
|
pre {
|
|
border: 0pt solid #93a1a1;
|
|
box-shadow: none;
|
|
}
|
|
.alert-text-small { font-size: 80%; }
|
|
.alert-text-large { font-size: 130%; }
|
|
.alert-text-normal { font-size: 90%; }
|
|
.alert {
|
|
padding:8px 35px 8px 14px; margin-bottom:18px;
|
|
text-shadow:0 1px 0 rgba(255,255,255,0.5);
|
|
border:1px solid #93a1a1;
|
|
border-radius: 4px;
|
|
-webkit-border-radius: 4px;
|
|
-moz-border-radius: 4px;
|
|
color: #555;
|
|
background-color: #eee8d5;
|
|
background-position: 10px 5px;
|
|
background-repeat: no-repeat;
|
|
background-size: 38px;
|
|
padding-left: 55px;
|
|
width: 75%;
|
|
}
|
|
.alert-block {padding-top:14px; padding-bottom:14px}
|
|
.alert-block > p, .alert-block > ul {margin-bottom:1em}
|
|
.alert li {margin-top: 1em}
|
|
.alert-block p+p {margin-top:5px}
|
|
.alert-notice { background-image: url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_yellow_notice.png); }
|
|
.alert-summary { background-image:url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_yellow_summary.png); }
|
|
.alert-warning { background-image: url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_yellow_warning.png); }
|
|
.alert-question {background-image:url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_yellow_question.png); }
|
|
|
|
div { text-align: justify; text-justify: inter-word; }
|
|
</style>
|
|
|
|
|
|
</head>
|
|
|
|
<!-- tocinfo
|
|
{'highest level': 2,
|
|
'sections': [('Plans for week 35, August 24-28',
|
|
2,
|
|
None,
|
|
'plans-for-week-35-august-24-28'),
|
|
('Thursday August 27', 2, None, 'thursday-august-27'),
|
|
('Why Linear Regression (aka Ordinary Least Squares and family), '
|
|
'repeat from last week',
|
|
2,
|
|
None,
|
|
'why-linear-regression-aka-ordinary-least-squares-and-family-repeat-from-last-week'),
|
|
('Regression analysis, overarching aims',
|
|
2,
|
|
None,
|
|
'regression-analysis-overarching-aims'),
|
|
('Regression analysis, overarching aims II',
|
|
2,
|
|
None,
|
|
'regression-analysis-overarching-aims-ii'),
|
|
('Examples', 2, None, 'examples'),
|
|
('General linear models', 2, None, 'general-linear-models'),
|
|
('Rewriting the fitting procedure as a linear algebra problem',
|
|
2,
|
|
None,
|
|
'rewriting-the-fitting-procedure-as-a-linear-algebra-problem'),
|
|
('Rewriting the fitting procedure as a linear algebra problem, '
|
|
'more details',
|
|
2,
|
|
None,
|
|
'rewriting-the-fitting-procedure-as-a-linear-algebra-problem-more-details'),
|
|
('Generalizing the fitting procedure as a linear algebra problem',
|
|
2,
|
|
None,
|
|
'generalizing-the-fitting-procedure-as-a-linear-algebra-problem'),
|
|
('Generalizing the fitting procedure as a linear algebra problem',
|
|
2,
|
|
None,
|
|
'generalizing-the-fitting-procedure-as-a-linear-algebra-problem'),
|
|
('Optimizing our parameters',
|
|
2,
|
|
None,
|
|
'optimizing-our-parameters'),
|
|
('Our model for the nuclear binding energies',
|
|
2,
|
|
None,
|
|
'our-model-for-the-nuclear-binding-energies'),
|
|
('Optimizing our parameters, more details',
|
|
2,
|
|
None,
|
|
'optimizing-our-parameters-more-details'),
|
|
('Interpretations and optimizing our parameters',
|
|
2,
|
|
None,
|
|
'interpretations-and-optimizing-our-parameters'),
|
|
('Interpretations and optimizing our parameters',
|
|
2,
|
|
None,
|
|
'interpretations-and-optimizing-our-parameters'),
|
|
('Some useful matrix and vector expressions',
|
|
2,
|
|
None,
|
|
'some-useful-matrix-and-vector-expressions'),
|
|
('Meet the Hessian Matrix', 2, None, 'meet-the-hessian-matrix'),
|
|
('Interpretations and optimizing our parameters',
|
|
2,
|
|
None,
|
|
'interpretations-and-optimizing-our-parameters'),
|
|
('Own code for Ordinary Least Squares',
|
|
2,
|
|
None,
|
|
'own-code-for-ordinary-least-squares'),
|
|
('Adding error analysis and training set up',
|
|
2,
|
|
None,
|
|
'adding-error-analysis-and-training-set-up'),
|
|
('Splitting our Data in Training and Test data',
|
|
2,
|
|
None,
|
|
'splitting-our-data-in-training-and-test-data'),
|
|
('Examples', 2, None, 'examples'),
|
|
('Making your own test-train splitting',
|
|
2,
|
|
None,
|
|
'making-your-own-test-train-splitting'),
|
|
('The Boston housing data example',
|
|
2,
|
|
None,
|
|
'the-boston-housing-data-example'),
|
|
('Housing data, the code', 2, None, 'housing-data-the-code'),
|
|
('Reducing the number of degrees of freedom, overarching view',
|
|
2,
|
|
None,
|
|
'reducing-the-number-of-degrees-of-freedom-overarching-view'),
|
|
('Preprocessing our data', 2, None, 'preprocessing-our-data'),
|
|
('Functionality in Scikit-Learn',
|
|
2,
|
|
None,
|
|
'functionality-in-scikit-learn'),
|
|
('More preprocessing', 2, None, 'more-preprocessing'),
|
|
('Frequently used scaling functions',
|
|
2,
|
|
None,
|
|
'frequently-used-scaling-functions'),
|
|
('Example of own Standard scaling',
|
|
2,
|
|
None,
|
|
'example-of-own-standard-scaling'),
|
|
('Min-Max Scaling', 2, None, 'min-max-scaling'),
|
|
('Testing the Means Squared Error as function of Complexity',
|
|
2,
|
|
None,
|
|
'testing-the-means-squared-error-as-function-of-complexity'),
|
|
('More preprocessing examples, Franke function and regression',
|
|
2,
|
|
None,
|
|
'more-preprocessing-examples-franke-function-and-regression'),
|
|
('Mathematical Interpretation of Ordinary Least Squares',
|
|
2,
|
|
None,
|
|
'mathematical-interpretation-of-ordinary-least-squares'),
|
|
('Residual Error', 2, None, 'residual-error'),
|
|
('Simple case', 2, None, 'simple-case'),
|
|
('The singular value decomposition',
|
|
2,
|
|
None,
|
|
'the-singular-value-decomposition'),
|
|
('Linear Regression Problems',
|
|
2,
|
|
None,
|
|
'linear-regression-problems'),
|
|
('Fixing the singularity', 2, None, 'fixing-the-singularity'),
|
|
('Basic math of the SVD', 2, None, 'basic-math-of-the-svd'),
|
|
('The SVD, a Fantastic Algorithm',
|
|
2,
|
|
None,
|
|
'the-svd-a-fantastic-algorithm'),
|
|
('Economy-size SVD', 2, None, 'economy-size-svd'),
|
|
('Codes for the SVD', 2, None, 'codes-for-the-svd'),
|
|
('Note about SVD Calculations',
|
|
2,
|
|
None,
|
|
'note-about-svd-calculations'),
|
|
('Mathematical Properties', 2, None, 'mathematical-properties'),
|
|
('Friday September 3', 2, None, 'friday-september-3'),
|
|
('Ridge and LASSO Regression',
|
|
2,
|
|
None,
|
|
'ridge-and-lasso-regression'),
|
|
('More on Ridge Regression', 2, None, 'more-on-ridge-regression'),
|
|
('Interpreting the Ridge results',
|
|
2,
|
|
None,
|
|
'interpreting-the-ridge-results'),
|
|
('More interpretations', 2, None, 'more-interpretations'),
|
|
('A better understanding of regularization',
|
|
2,
|
|
None,
|
|
'a-better-understanding-of-regularization'),
|
|
('Decomposing the OLS and Ridge expressions',
|
|
2,
|
|
None,
|
|
'decomposing-the-ols-and-ridge-expressions'),
|
|
('Introducing the Covariance and Correlation functions',
|
|
2,
|
|
None,
|
|
'introducing-the-covariance-and-correlation-functions'),
|
|
('Correlation Function and Design/Feature Matrix',
|
|
2,
|
|
None,
|
|
'correlation-function-and-design-feature-matrix'),
|
|
('Covariance Matrix Examples',
|
|
2,
|
|
None,
|
|
'covariance-matrix-examples'),
|
|
('Correlation Matrix', 2, None, 'correlation-matrix'),
|
|
('Correlation Matrix with Pandas',
|
|
2,
|
|
None,
|
|
'correlation-matrix-with-pandas'),
|
|
('Correlation Matrix with Pandas and the Franke function',
|
|
2,
|
|
None,
|
|
'correlation-matrix-with-pandas-and-the-franke-function'),
|
|
('Rewriting the Covariance and/or Correlation Matrix',
|
|
2,
|
|
None,
|
|
'rewriting-the-covariance-and-or-correlation-matrix'),
|
|
('Linking with SVD', 2, None, 'linking-with-svd'),
|
|
('Exercises for week 37, September 6-10',
|
|
2,
|
|
None,
|
|
'exercises-for-week-37-september-6-10'),
|
|
('Exercise 1: Adding Ridge and Lasso Regression',
|
|
2,
|
|
None,
|
|
'exercise-1-adding-ridge-and-lasso-regression'),
|
|
('Exercise: Linear Regression for a two-dimensional function',
|
|
3,
|
|
None,
|
|
'exercise-linear-regression-for-a-two-dimensional-function')]}
|
|
end of tocinfo -->
|
|
|
|
<body>
|
|
|
|
|
|
|
|
<script type="text/x-mathjax-config">
|
|
MathJax.Hub.Config({
|
|
TeX: {
|
|
equationNumbers: { autoNumber: "AMS" },
|
|
extensions: ["AMSmath.js", "AMSsymbols.js", "autobold.js", "color.js"]
|
|
}
|
|
});
|
|
</script>
|
|
<script type="text/javascript" async
|
|
src="https://cdnjs.cloudflare.com/ajax/libs/mathjax/2.7.1/MathJax.js?config=TeX-AMS-MML_HTMLorMML">
|
|
</script>
|
|
|
|
|
|
|
|
|
|
<!-- ------------------- main content ---------------------- -->
|
|
|
|
|
|
|
|
<center><h1>Week 35: From Ordinary Linear Regression to Ridge and Lasso Regression</h1></center> <!-- document title -->
|
|
|
|
<p>
|
|
<!-- author(s): Morten Hjorth-Jensen -->
|
|
|
|
<center>
|
|
<b>Morten Hjorth-Jensen</b> [1, 2]
|
|
</center>
|
|
|
|
<p>
|
|
<!-- institution(s) -->
|
|
|
|
<center>[1] <b>Department of Physics, University of Oslo</b></center>
|
|
<center>[2] <b>Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University</b></center>
|
|
<br>
|
|
<p>
|
|
<center><h4>Sep 2, 2021</h4></center> <!-- date -->
|
|
<br>
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="plans-for-week-35-august-24-28">Plans for week 35, August 24-28 </h2>
|
|
|
|
<ul>
|
|
<li> Thursday: Review of ordinary Least Squares with applications and discussion of Ridge Regression</li>
|
|
<li> Friday: Analysis of Ridge and Lasso Regression</li>
|
|
</ul>
|
|
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="thursday-august-27">Thursday August 27 </h2>
|
|
|
|
<p>
|
|
The main topics on Thursday are:
|
|
|
|
<ol>
|
|
<li> Repetition from last week on linear regression</li>
|
|
<li> Discussion of how to prepare data and examples of applications of linear regression</li>
|
|
<li> Mathematical interpretations of Linear Regression</li>
|
|
<li> Start discussing Ridge and Lasso regression and Singular Value Decomposition</li>
|
|
</ol>
|
|
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="why-linear-regression-aka-ordinary-least-squares-and-family-repeat-from-last-week">Why Linear Regression (aka Ordinary Least Squares and family), repeat from last week </h2>
|
|
|
|
<p>
|
|
We need first a reminder from last week about linear regression.
|
|
|
|
<p>
|
|
Fitting a continuous function with linear parameterization in terms of the parameters \( \boldsymbol{\beta} \).
|
|
|
|
<ul>
|
|
<li> Method of choice for fitting a continuous function!</li>
|
|
<li> Gives an excellent introduction to central Machine Learning features with <b>understandable pedagogical</b> links to other methods like <b>Neural Networks</b>, <b>Support Vector Machines</b> etc</li>
|
|
<li> Analytical expression for the fitting parameters \( \boldsymbol{\beta} \)</li>
|
|
<li> Analytical expressions for statistical propertiers like mean values, variances, confidence intervals and more</li>
|
|
<li> Analytical relation with probabilistic interpretations</li>
|
|
<li> Easy to introduce basic concepts like bias-variance tradeoff, cross-validation, resampling and regularization techniques and many other ML topics</li>
|
|
<li> Easy to code! And links well with classification problems and logistic regression and neural networks</li>
|
|
<li> Allows for <b>easy</b> hands-on understanding of gradient descent methods</li>
|
|
<li> and many more features</li>
|
|
</ul>
|
|
|
|
For more discussions of Ridge and Lasso regression, <a href="https://arxiv.org/abs/1509.09169" target="_blank">Wessel van Wieringen's</a> article is highly recommended.
|
|
Similarly, <a href="https://arxiv.org/abs/1803.08823" target="_blank">Mehta et al's article</a> is also recommended.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="regression-analysis-overarching-aims">Regression analysis, overarching aims </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
|
|
<p>
|
|
Regression modeling deals with the description of the sampling distribution of a given random variable \( y \) and how it varies as function of another variable or a set of such variables \( \boldsymbol{x} =[x_0, x_1,\dots, x_{n-1}]^T \).
|
|
The first variable is called the <b>dependent</b>, the <b>outcome</b> or the <b>response</b> variable while the set of variables \( \boldsymbol{x} \) is called the independent variable, or the predictor variable or the explanatory variable.
|
|
|
|
<p>
|
|
A regression model aims at finding a likelihood function \( p(\boldsymbol{y}\vert \boldsymbol{x}) \), that is the conditional distribution for \( \boldsymbol{y} \) with a given \( \boldsymbol{x} \). The estimation of \( p(\boldsymbol{y}\vert \boldsymbol{x}) \) is made using a data set with
|
|
|
|
<ul>
|
|
<li> \( n \) cases \( i = 0, 1, 2, \dots, n-1 \)</li>
|
|
<li> Response (target, dependent or outcome) variable \( y_i \) with \( i = 0, 1, 2, \dots, n-1 \)</li>
|
|
<li> \( p \) so-called explanatory (independent or predictor) variables \( \boldsymbol{x}_i=[x_{i0}, x_{i1}, \dots, x_{ip-1}] \) with \( i = 0, 1, 2, \dots, n-1 \) and explanatory variables running from \( 0 \) to \( p-1 \). See below for more explicit examples.</li>
|
|
</ul>
|
|
|
|
The goal of the regression analysis is to extract/exploit relationship between \( \boldsymbol{y} \) and \( \boldsymbol{x} \) in or to infer causal dependencies, approximations to the likelihood functions, functional relationships and to make predictions, making fits and many other things.
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="regression-analysis-overarching-aims-ii">Regression analysis, overarching aims II </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
|
|
<p>
|
|
Consider an experiment in which \( p \) characteristics of \( n \) samples are
|
|
measured. The data from this experiment, for various explanatory variables \( p \) are normally represented by a matrix
|
|
\( \mathbf{X} \).
|
|
|
|
<p>
|
|
The matrix \( \mathbf{X} \) is called the <em>design
|
|
matrix</em>. Additional information of the samples is available in the
|
|
form of \( \boldsymbol{y} \) (also as above). The variable \( \boldsymbol{y} \) is
|
|
generally referred to as the <em>response variable</em>. The aim of
|
|
regression analysis is to explain \( \boldsymbol{y} \) in terms of
|
|
\( \boldsymbol{X} \) through a functional relationship like \( y_i =
|
|
f(\mathbf{X}_{i,\ast}) \). When no prior knowledge on the form of
|
|
\( f(\cdot) \) is available, it is common to assume a linear relationship
|
|
between \( \boldsymbol{X} \) and \( \boldsymbol{y} \). This assumption gives rise to
|
|
the <em>linear regression model</em> where \( \boldsymbol{\beta} = [\beta_0, \ldots,
|
|
\beta_{p-1}]^{T} \) are the <em>regression parameters</em>.
|
|
|
|
<p>
|
|
Linear regression gives us a set of analytical equations for the parameters \( \beta_j \).
|
|
|
|
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="examples">Examples </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
In order to understand the relation among the predictors \( p \), the set of data \( n \) and the target (outcome, output etc) \( \boldsymbol{y} \),
|
|
consider the model we discussed for describing nuclear binding energies.
|
|
|
|
<p>
|
|
There we assumed that we could parametrize the data using a polynomial approximation based on the liquid drop model.
|
|
Assuming
|
|
$$
|
|
BE(A) = a_0+a_1A+a_2A^{2/3}+a_3A^{-1/3}+a_4A^{-1},
|
|
$$
|
|
|
|
we have five predictors, that is the intercept, the \( A \) dependent term, the \( A^{2/3} \) term and the \( A^{-1/3} \) and \( A^{-1} \) terms.
|
|
This gives \( p=0,1,2,3,4 \). Furthermore we have \( n \) entries for each predictor. It means that our design matrix is a
|
|
\( p\times n \) matrix \( \boldsymbol{X} \).
|
|
|
|
<p>
|
|
Here the predictors are based on a model we have made. A popular data set which is widely encountered in ML applications is the
|
|
so-called <a href="https://www.sciencedirect.com/science/article/pii/S0957417407006719?via%3Dihub" target="_blank">credit card default data from Taiwan</a>. The data set contains data on \( n=30000 \) credit card holders with predictors like gender, marital status, age, profession, education, etc. In total there are \( 24 \) such predictors or attributes leading to a design matrix of dimensionality \( 24 \times 30000 \). This is however a classification problem and we will come back to it when we discuss Logistic Regression.
|
|
|
|
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="general-linear-models">General linear models </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
Before we proceed let us study a case from linear algebra where we aim at fitting a set of data \( \boldsymbol{y}=[y_0,y_1,\dots,y_{n-1}] \). We could think of these data as a result of an experiment or a complicated numerical experiment. These data are functions of a series of variables \( \boldsymbol{x}=[x_0,x_1,\dots,x_{n-1}] \), that is \( y_i = y(x_i) \) with \( i=0,1,2,\dots,n-1 \). The variables \( x_i \) could represent physical quantities like time, temperature, position etc. We assume that \( y(x) \) is a smooth function.
|
|
|
|
<p>
|
|
Since obtaining these data points may not be trivial, we want to use these data to fit a function which can allow us to make predictions for values of \( y \) which are not in the present set. The perhaps simplest approach is to assume we can parametrize our function in terms of a polynomial of degree \( n-1 \) with \( n \) points, that is
|
|
$$
|
|
y=y(x) \rightarrow y(x_i)=\tilde{y}_i+\epsilon_i=\sum_{j=0}^{n-1} \beta_j x_i^j+\epsilon_i,
|
|
$$
|
|
|
|
where \( \epsilon_i \) is the error in our approximation.
|
|
|
|
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="rewriting-the-fitting-procedure-as-a-linear-algebra-problem">Rewriting the fitting procedure as a linear algebra problem </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
For every set of values \( y_i,x_i \) we have thus the corresponding set of equations
|
|
$$
|
|
\begin{align*}
|
|
y_0&=\beta_0+\beta_1x_0^1+\beta_2x_0^2+\dots+\beta_{n-1}x_0^{n-1}+\epsilon_0\\
|
|
y_1&=\beta_0+\beta_1x_1^1+\beta_2x_1^2+\dots+\beta_{n-1}x_1^{n-1}+\epsilon_1\\
|
|
y_2&=\beta_0+\beta_1x_2^1+\beta_2x_2^2+\dots+\beta_{n-1}x_2^{n-1}+\epsilon_2\\
|
|
\dots & \dots \\
|
|
y_{n-1}&=\beta_0+\beta_1x_{n-1}^1+\beta_2x_{n-1}^2+\dots+\beta_{n-1}x_{n-1}^{n-1}+\epsilon_{n-1}.\\
|
|
\end{align*}
|
|
$$
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="rewriting-the-fitting-procedure-as-a-linear-algebra-problem-more-details">Rewriting the fitting procedure as a linear algebra problem, more details </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
Defining the vectors
|
|
$$
|
|
\boldsymbol{y} = [y_0,y_1, y_2,\dots, y_{n-1}]^T,
|
|
$$
|
|
|
|
and
|
|
$$
|
|
\boldsymbol{\beta} = [\beta_0,\beta_1, \beta_2,\dots, \beta_{n-1}]^T,
|
|
$$
|
|
|
|
and
|
|
$$
|
|
\boldsymbol{\epsilon} = [\epsilon_0,\epsilon_1, \epsilon_2,\dots, \epsilon_{n-1}]^T,
|
|
$$
|
|
|
|
and the design matrix
|
|
$$
|
|
\boldsymbol{X}=
|
|
\begin{bmatrix}
|
|
1& x_{0}^1 &x_{0}^2& \dots & \dots &x_{0}^{n-1}\\
|
|
1& x_{1}^1 &x_{1}^2& \dots & \dots &x_{1}^{n-1}\\
|
|
1& x_{2}^1 &x_{2}^2& \dots & \dots &x_{2}^{n-1}\\
|
|
\dots& \dots &\dots& \dots & \dots &\dots\\
|
|
1& x_{n-1}^1 &x_{n-1}^2& \dots & \dots &x_{n-1}^{n-1}\\
|
|
\end{bmatrix}
|
|
$$
|
|
|
|
we can rewrite our equations as
|
|
$$
|
|
\boldsymbol{y} = \boldsymbol{X}\boldsymbol{\beta}+\boldsymbol{\epsilon}.
|
|
$$
|
|
|
|
The above design matrix is called a <a href="https://en.wikipedia.org/wiki/Vandermonde_matrix" target="_blank">Vandermonde matrix</a>.
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="generalizing-the-fitting-procedure-as-a-linear-algebra-problem">Generalizing the fitting procedure as a linear algebra problem </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
|
|
<p>
|
|
We are obviously not limited to the above polynomial expansions. We
|
|
could replace the various powers of \( x \) with elements of Fourier
|
|
series or instead of \( x_i^j \) we could have \( \cos{(j x_i)} \) or \( \sin{(j
|
|
x_i)} \), or time series or other orthogonal functions. For every set
|
|
of values \( y_i,x_i \) we can then generalize the equations to
|
|
|
|
$$
|
|
\begin{align*}
|
|
y_0&=\beta_0x_{00}+\beta_1x_{01}+\beta_2x_{02}+\dots+\beta_{n-1}x_{0n-1}+\epsilon_0\\
|
|
y_1&=\beta_0x_{10}+\beta_1x_{11}+\beta_2x_{12}+\dots+\beta_{n-1}x_{1n-1}+\epsilon_1\\
|
|
y_2&=\beta_0x_{20}+\beta_1x_{21}+\beta_2x_{22}+\dots+\beta_{n-1}x_{2n-1}+\epsilon_2\\
|
|
\dots & \dots \\
|
|
y_{i}&=\beta_0x_{i0}+\beta_1x_{i1}+\beta_2x_{i2}+\dots+\beta_{n-1}x_{in-1}+\epsilon_i\\
|
|
\dots & \dots \\
|
|
y_{n-1}&=\beta_0x_{n-1,0}+\beta_1x_{n-1,2}+\beta_2x_{n-1,2}+\dots+\beta_{n-1}x_{n-1,n-1}+\epsilon_{n-1}.\\
|
|
\end{align*}
|
|
$$
|
|
|
|
<p>
|
|
<b>Note that we have \( p=n \) here. The matrix is symmetric. This is generally not the case!</b>
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="generalizing-the-fitting-procedure-as-a-linear-algebra-problem">Generalizing the fitting procedure as a linear algebra problem </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
We redefine in turn the matrix \( \boldsymbol{X} \) as
|
|
$$
|
|
\boldsymbol{X}=
|
|
\begin{bmatrix}
|
|
x_{00}& x_{01} &x_{02}& \dots & \dots &x_{0,n-1}\\
|
|
x_{10}& x_{11} &x_{12}& \dots & \dots &x_{1,n-1}\\
|
|
x_{20}& x_{21} &x_{22}& \dots & \dots &x_{2,n-1}\\
|
|
\dots& \dots &\dots& \dots & \dots &\dots\\
|
|
x_{n-1,0}& x_{n-1,1} &x_{n-1,2}& \dots & \dots &x_{n-1,n-1}\\
|
|
\end{bmatrix}
|
|
$$
|
|
|
|
and without loss of generality we rewrite again our equations as
|
|
$$
|
|
\boldsymbol{y} = \boldsymbol{X}\boldsymbol{\beta}+\boldsymbol{\epsilon}.
|
|
$$
|
|
|
|
The left-hand side of this equation is kwown. Our error vector \( \boldsymbol{\epsilon} \) and the parameter vector \( \boldsymbol{\beta} \) are our unknow quantities. How can we obtain the optimal set of \( \beta_i \) values?
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="optimizing-our-parameters">Optimizing our parameters </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
We have defined the matrix \( \boldsymbol{X} \) via the equations
|
|
$$
|
|
\begin{align*}
|
|
y_0&=\beta_0x_{00}+\beta_1x_{01}+\beta_2x_{02}+\dots+\beta_{n-1}x_{0n-1}+\epsilon_0\\
|
|
y_1&=\beta_0x_{10}+\beta_1x_{11}+\beta_2x_{12}+\dots+\beta_{n-1}x_{1n-1}+\epsilon_1\\
|
|
y_2&=\beta_0x_{20}+\beta_1x_{21}+\beta_2x_{22}+\dots+\beta_{n-1}x_{2n-1}+\epsilon_1\\
|
|
\dots & \dots \\
|
|
y_{i}&=\beta_0x_{i0}+\beta_1x_{i1}+\beta_2x_{i2}+\dots+\beta_{n-1}x_{in-1}+\epsilon_1\\
|
|
\dots & \dots \\
|
|
y_{n-1}&=\beta_0x_{n-1,0}+\beta_1x_{n-1,2}+\beta_2x_{n-1,2}+\dots+\beta_{n-1}x_{n-1,n-1}+\epsilon_{n-1}.\\
|
|
\end{align*}
|
|
$$
|
|
|
|
<p>
|
|
As we noted above, we stayed with a system with the design matrix
|
|
\( \boldsymbol{X}\in {\mathbb{R}}^{n\times n} \), that is we have \( p=n \). For reasons to come later (algorithmic arguments) we will hereafter define
|
|
our matrix as \( \boldsymbol{X}\in {\mathbb{R}}^{n\times p} \), with the predictors refering to the column numbers and the entries \( n \) being the row elements.
|
|
|
|
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="our-model-for-the-nuclear-binding-energies">Our model for the nuclear binding energies </h2>
|
|
|
|
<p>
|
|
In our <a href="https://compphysics.github.io/MachineLearning/doc/pub/How2ReadData/html/How2ReadData.html" target="_blank">introductory notes</a> we looked at the so-called <a href="https://en.wikipedia.org/wiki/Semi-empirical_mass_formula" target="_blank">liquid drop model</a>. Let us remind ourselves about what we did by looking at the code.
|
|
|
|
<p>
|
|
We restate the parts of the code we are most interested in.
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># Common imports</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> display
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
|
|
|
|
<span style="color: #228B22"># Where to save the figures and data files</span>
|
|
PROJECT_ROOT_DIR = <span style="color: #CD5555">"Results"</span>
|
|
FIGURE_ID = <span style="color: #CD5555">"Results/FigureFiles"</span>
|
|
DATA_ID = <span style="color: #CD5555">"DataFiles/"</span>
|
|
|
|
<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(PROJECT_ROOT_DIR):
|
|
os.mkdir(PROJECT_ROOT_DIR)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(FIGURE_ID):
|
|
os.makedirs(FIGURE_ID)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(DATA_ID):
|
|
os.makedirs(DATA_ID)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">image_path</span>(fig_id):
|
|
<span style="color: #8B008B; font-weight: bold">return</span> os.path.join(FIGURE_ID, fig_id)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">data_path</span>(dat_id):
|
|
<span style="color: #8B008B; font-weight: bold">return</span> os.path.join(DATA_ID, dat_id)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">save_fig</span>(fig_id):
|
|
plt.savefig(image_path(fig_id) + <span style="color: #CD5555">".png"</span>, <span style="color: #658b00">format</span>=<span style="color: #CD5555">'png'</span>)
|
|
|
|
infile = <span style="color: #658b00">open</span>(data_path(<span style="color: #CD5555">"MassEval2016.dat"</span>),<span style="color: #CD5555">'r'</span>)
|
|
|
|
|
|
<span style="color: #228B22"># Read the experimental data with Pandas</span>
|
|
Masses = pd.read_fwf(infile, usecols=(<span style="color: #B452CD">2</span>,<span style="color: #B452CD">3</span>,<span style="color: #B452CD">4</span>,<span style="color: #B452CD">6</span>,<span style="color: #B452CD">11</span>),
|
|
names=(<span style="color: #CD5555">'N'</span>, <span style="color: #CD5555">'Z'</span>, <span style="color: #CD5555">'A'</span>, <span style="color: #CD5555">'Element'</span>, <span style="color: #CD5555">'Ebinding'</span>),
|
|
widths=(<span style="color: #B452CD">1</span>,<span style="color: #B452CD">3</span>,<span style="color: #B452CD">5</span>,<span style="color: #B452CD">5</span>,<span style="color: #B452CD">5</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">3</span>,<span style="color: #B452CD">4</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">13</span>,<span style="color: #B452CD">11</span>,<span style="color: #B452CD">11</span>,<span style="color: #B452CD">9</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">2</span>,<span style="color: #B452CD">11</span>,<span style="color: #B452CD">9</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">3</span>,<span style="color: #B452CD">1</span>,<span style="color: #B452CD">12</span>,<span style="color: #B452CD">11</span>,<span style="color: #B452CD">1</span>),
|
|
header=<span style="color: #B452CD">39</span>,
|
|
index_col=<span style="color: #8B008B; font-weight: bold">False</span>)
|
|
|
|
<span style="color: #228B22"># Extrapolated values are indicated by '#' in place of the decimal place, so</span>
|
|
<span style="color: #228B22"># the Ebinding column won't be numeric. Coerce to float and drop these entries.</span>
|
|
Masses[<span style="color: #CD5555">'Ebinding'</span>] = pd.to_numeric(Masses[<span style="color: #CD5555">'Ebinding'</span>], errors=<span style="color: #CD5555">'coerce'</span>)
|
|
Masses = Masses.dropna()
|
|
<span style="color: #228B22"># Convert from keV to MeV.</span>
|
|
Masses[<span style="color: #CD5555">'Ebinding'</span>] /= <span style="color: #B452CD">1000</span>
|
|
|
|
<span style="color: #228B22"># Group the DataFrame by nucleon number, A.</span>
|
|
Masses = Masses.groupby(<span style="color: #CD5555">'A'</span>)
|
|
<span style="color: #228B22"># Find the rows of the grouped DataFrame with the maximum binding energy.</span>
|
|
Masses = Masses.apply(<span style="color: #8B008B; font-weight: bold">lambda</span> t: t[t.Ebinding==t.Ebinding.max()])
|
|
A = Masses[<span style="color: #CD5555">'A'</span>]
|
|
Z = Masses[<span style="color: #CD5555">'Z'</span>]
|
|
N = Masses[<span style="color: #CD5555">'N'</span>]
|
|
Element = Masses[<span style="color: #CD5555">'Element'</span>]
|
|
Energies = Masses[<span style="color: #CD5555">'Ebinding'</span>]
|
|
|
|
<span style="color: #228B22"># Now we set up the design matrix X</span>
|
|
X = np.zeros((<span style="color: #658b00">len</span>(A),<span style="color: #B452CD">5</span>))
|
|
X[:,<span style="color: #B452CD">0</span>] = <span style="color: #B452CD">1</span>
|
|
X[:,<span style="color: #B452CD">1</span>] = A
|
|
X[:,<span style="color: #B452CD">2</span>] = A**(<span style="color: #B452CD">2.0</span>/<span style="color: #B452CD">3.0</span>)
|
|
X[:,<span style="color: #B452CD">3</span>] = A**(-<span style="color: #B452CD">1.0</span>/<span style="color: #B452CD">3.0</span>)
|
|
X[:,<span style="color: #B452CD">4</span>] = A**(-<span style="color: #B452CD">1.0</span>)
|
|
<span style="color: #228B22"># Then nice printout using pandas</span>
|
|
DesignMatrix = pd.DataFrame(X)
|
|
DesignMatrix.index = A
|
|
DesignMatrix.columns = [<span style="color: #CD5555">'1'</span>, <span style="color: #CD5555">'A'</span>, <span style="color: #CD5555">'A^(2/3)'</span>, <span style="color: #CD5555">'A^(-1/3)'</span>, <span style="color: #CD5555">'1/A'</span>]
|
|
display(DesignMatrix)
|
|
</pre></div>
|
|
<p>
|
|
With \( \boldsymbol{\beta}\in {\mathbb{R}}^{p\times 1} \), it means that we will hereafter write our equations for the approximation as
|
|
$$
|
|
\boldsymbol{\tilde{y}}= \boldsymbol{X}\boldsymbol{\beta},
|
|
$$
|
|
|
|
throughout these lectures.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="optimizing-our-parameters-more-details">Optimizing our parameters, more details </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
With the above we use the design matrix to define the approximation \( \boldsymbol{\tilde{y}} \) via the unknown quantity \( \boldsymbol{\beta} \) as
|
|
$$
|
|
\boldsymbol{\tilde{y}}= \boldsymbol{X}\boldsymbol{\beta},
|
|
$$
|
|
|
|
and in order to find the optimal parameters \( \beta_i \) instead of solving the above linear algebra problem, we define a function which gives a measure of the spread between the values \( y_i \) (which represent hopefully the exact values) and the parameterized values \( \tilde{y}_i \), namely
|
|
$$
|
|
C(\boldsymbol{\beta})=\frac{1}{n}\sum_{i=0}^{n-1}\left(y_i-\tilde{y}_i\right)^2=\frac{1}{n}\left\{\left(\boldsymbol{y}-\boldsymbol{\tilde{y}}\right)^T\left(\boldsymbol{y}-\boldsymbol{\tilde{y}}\right)\right\},
|
|
$$
|
|
|
|
or using the matrix \( \boldsymbol{X} \) and in a more compact matrix-vector notation as
|
|
$$
|
|
C(\boldsymbol{\beta})=\frac{1}{n}\left\{\left(\boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right)^T\left(\boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right)\right\}.
|
|
$$
|
|
|
|
This function is one possible way to define the so-called cost function.
|
|
|
|
<p>
|
|
It is also common to define
|
|
the function \( C \) as
|
|
|
|
$$
|
|
C(\boldsymbol{\beta})=\frac{1}{2n}\sum_{i=0}^{n-1}\left(y_i-\tilde{y}_i\right)^2,
|
|
$$
|
|
|
|
since when taking the first derivative with respect to the unknown parameters \( \beta \), the factor of \( 2 \) cancels out.
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="interpretations-and-optimizing-our-parameters">Interpretations and optimizing our parameters </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
|
|
<p>
|
|
The function
|
|
$$
|
|
C(\boldsymbol{\beta})=\frac{1}{n}\left\{\left(\boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right)^T\left(\boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right)\right\},
|
|
$$
|
|
|
|
can be linked to the variance of the quantity \( y_i \) if we interpret the latter as the mean value.
|
|
When linking (see the discussion below) with the maximum likelihood approach below, we will indeed interpret \( y_i \) as a mean value
|
|
$$
|
|
y_{i}=\langle y_i \rangle = \beta_0x_{i,0}+\beta_1x_{i,1}+\beta_2x_{i,2}+\dots+\beta_{n-1}x_{i,n-1}+\epsilon_i,
|
|
$$
|
|
|
|
<p>
|
|
where \( \langle y_i \rangle \) is the mean value. Keep in mind also that
|
|
till now we have treated \( y_i \) as the exact value. Normally, the
|
|
response (dependent or outcome) variable \( y_i \) the outcome of a
|
|
numerical experiment or another type of experiment and is thus only an
|
|
approximation to the true value. It is then always accompanied by an
|
|
error estimate, often limited to a statistical error estimate given by
|
|
the standard deviation discussed earlier. In the discussion here we
|
|
will treat \( y_i \) as our exact value for the response variable.
|
|
|
|
<p>
|
|
In order to find the parameters \( \beta_i \) we will then minimize the spread of \( C(\boldsymbol{\beta}) \), that is we are going to solve the problem
|
|
$$
|
|
{\displaystyle \min_{\boldsymbol{\beta}\in
|
|
{\mathbb{R}}^{p}}}\frac{1}{n}\left\{\left(\boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right)^T\left(\boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right)\right\}.
|
|
$$
|
|
|
|
In practical terms it means we will require
|
|
$$
|
|
\frac{\partial C(\boldsymbol{\beta})}{\partial \beta_j} = \frac{\partial }{\partial \beta_j}\left[ \frac{1}{n}\sum_{i=0}^{n-1}\left(y_i-\beta_0x_{i,0}-\beta_1x_{i,1}-\beta_2x_{i,2}-\dots-\beta_{n-1}x_{i,n-1}\right)^2\right]=0,
|
|
$$
|
|
|
|
which results in
|
|
$$
|
|
\frac{\partial C(\boldsymbol{\beta})}{\partial \beta_j} = -\frac{2}{n}\left[ \sum_{i=0}^{n-1}x_{ij}\left(y_i-\beta_0x_{i,0}-\beta_1x_{i,1}-\beta_2x_{i,2}-\dots-\beta_{n-1}x_{i,n-1}\right)\right]=0,
|
|
$$
|
|
|
|
or in a matrix-vector form as
|
|
$$
|
|
\frac{\partial C(\boldsymbol{\beta})}{\partial \boldsymbol{\beta}} = 0 = \boldsymbol{X}^T\left( \boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right).
|
|
$$
|
|
|
|
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="interpretations-and-optimizing-our-parameters">Interpretations and optimizing our parameters </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
We can rewrite
|
|
$$
|
|
\frac{\partial C(\boldsymbol{\beta})}{\partial \boldsymbol{\beta}} = 0 = \boldsymbol{X}^T\left( \boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right),
|
|
$$
|
|
|
|
as
|
|
$$
|
|
\boldsymbol{X}^T\boldsymbol{y} = \boldsymbol{X}^T\boldsymbol{X}\boldsymbol{\beta},
|
|
$$
|
|
|
|
and if the matrix \( \boldsymbol{X}^T\boldsymbol{X} \) is invertible we have the solution
|
|
$$
|
|
\boldsymbol{\beta} =\left(\boldsymbol{X}^T\boldsymbol{X}\right)^{-1}\boldsymbol{X}^T\boldsymbol{y}.
|
|
$$
|
|
|
|
<p>
|
|
We note also that since our design matrix is defined as \( \boldsymbol{X}\in
|
|
{\mathbb{R}}^{n\times p} \), the product \( \boldsymbol{X}^T\boldsymbol{X} \in
|
|
{\mathbb{R}}^{p\times p} \). In the above case we have that \( p \ll n \),
|
|
in our case \( p=5 \) meaning that we end up with inverting a small
|
|
\( 5\times 5 \) matrix. This is a rather common situation, in many cases we end up with low-dimensional
|
|
matrices to invert. The methods discussed here and for many other
|
|
supervised learning algorithms like classification with logistic
|
|
regression or support vector machines, exhibit dimensionalities which
|
|
allow for the usage of direct linear algebra methods such as <b>LU</b> decomposition or <b>Singular Value Decomposition</b> (SVD) for finding the inverse of the matrix
|
|
\( \boldsymbol{X}^T\boldsymbol{X} \).
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
<b>Small question</b>: Do you think the example we have at hand here (the nuclear binding energies) can lead to problems in inverting the matrix \( \boldsymbol{X}^T\boldsymbol{X} \)? What kind of problems can we expect?
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="some-useful-matrix-and-vector-expressions">Some useful matrix and vector expressions </h2>
|
|
|
|
<p>
|
|
The following matrix and vector relation will be useful here and for the rest of the course. Vectors are always written as boldfaced lower case letters and
|
|
matrices as upper case boldfaced letters.
|
|
|
|
$$
|
|
\frac{\partial (\boldsymbol{b}^T\boldsymbol{a})}{\partial \boldsymbol{a}} = \boldsymbol{b},
|
|
$$
|
|
|
|
$$
|
|
\frac{\partial (\boldsymbol{a}^T\boldsymbol{A}\boldsymbol{a})}{\partial \boldsymbol{a}} = (\boldsymbol{A}+\boldsymbol{A}^T)\boldsymbol{a},
|
|
$$
|
|
|
|
$$
|
|
\frac{\partial tr(\boldsymbol{B}\boldsymbol{A})}{\partial \boldsymbol{A}} = \boldsymbol{B}^T,
|
|
$$
|
|
|
|
$$
|
|
\frac{\partial \log{\vert\boldsymbol{A}\vert}}{\partial \boldsymbol{A}} = (\boldsymbol{A}^{-1})^T.
|
|
$$
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="meet-the-hessian-matrix">Meet the Hessian Matrix </h2>
|
|
|
|
<p>
|
|
A very important matrix we will meet again and again in Machine Learning is the Hessian.
|
|
It is given by the second derivative of the cost function with respect to the parameter \( \beta \). Using the above expression for derivatives of vectors and matrices, we find that the second derivative of the cost function is,
|
|
|
|
$$
|
|
\frac{\partial}{\partial \boldsymbol{\beta}^T}\frac{\partial C(\boldsymbol{\beta})}{\partial \boldsymbol{\beta}} =\frac{\partial}{\partial \boldsymbol{\beta}^T}\left[-\frac{2}{n}\boldsymbol{X}^T\left( \boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right)\right]=\frac{2}{n}\boldsymbol{X}^T\boldsymbol{X}.
|
|
$$
|
|
|
|
<p>
|
|
The Hessian matrix plays an important role and is defined here as
|
|
|
|
$$
|
|
\boldsymbol{H}=\boldsymbol{X}^T\boldsymbol{X}.
|
|
$$
|
|
|
|
<p>
|
|
For ordinary least squares, it is inversely proportional (derivation next week) with the variance of the optimal parameters
|
|
\( \hat{\boldsymbol{\beta}} \). Furthermore, we will see later this week that is (beside \( 1/n \)) equal to the covariance matrix. It plays also a very important role in optmization algorithms and Principal Component Analysis as a way to reduce the dimensionality of a machine learning problem.
|
|
|
|
<p>
|
|
<b>Linear algebra question:</b> Can we use the Hessian matrix to say something about properties of the cost function (our optmization problem)? (hint: think about convex or concave problems and how to relate these to a matrix!).
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="interpretations-and-optimizing-our-parameters">Interpretations and optimizing our parameters </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
The residuals \( \boldsymbol{\epsilon} \) are in turn given by
|
|
$$
|
|
\boldsymbol{\epsilon} = \boldsymbol{y}-\boldsymbol{\tilde{y}} = \boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta},
|
|
$$
|
|
|
|
and with
|
|
$$
|
|
\boldsymbol{X}^T\left( \boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right)= 0,
|
|
$$
|
|
|
|
we have
|
|
$$
|
|
\boldsymbol{X}^T\boldsymbol{\epsilon}=\boldsymbol{X}^T\left( \boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right)= 0,
|
|
$$
|
|
|
|
meaning that the solution for \( \boldsymbol{\beta} \) is the one which minimizes the residuals. Later we will link this with the maximum likelihood approach.
|
|
|
|
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="own-code-for-ordinary-least-squares">Own code for Ordinary Least Squares </h2>
|
|
|
|
<p>
|
|
It is rather straightforward to implement the matrix inversion and obtain the parameters \( \boldsymbol{\beta} \). After having defined the matrix \( \boldsymbol{X} \) we simply need to
|
|
write
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># matrix inversion to find beta</span>
|
|
beta = np.linalg.inv(X.T.dot(X)).dot(X.T).dot(Energies)
|
|
<span style="color: #228B22"># and then make the prediction</span>
|
|
ytilde = X @ beta
|
|
</pre></div>
|
|
<p>
|
|
Alternatively, you can use the least squares functionality in <b>Numpy</b> as
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>fit = np.linalg.lstsq(X, Energies, rcond =<span style="color: #8B008B; font-weight: bold">None</span>)[<span style="color: #B452CD">0</span>]
|
|
ytildenp = np.dot(fit,X.T)
|
|
</pre></div>
|
|
<p>
|
|
And finally we plot our fit with and compare with data
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>Masses[<span style="color: #CD5555">'Eapprox'</span>] = ytilde
|
|
<span style="color: #228B22"># Generate a plot comparing the experimental with the fitted values values.</span>
|
|
fig, ax = plt.subplots()
|
|
ax.set_xlabel(<span style="color: #CD5555">r'$A = N + Z$'</span>)
|
|
ax.set_ylabel(<span style="color: #CD5555">r'$E_\mathrm{bind}\,/\mathrm{MeV}$'</span>)
|
|
ax.plot(Masses[<span style="color: #CD5555">'A'</span>], Masses[<span style="color: #CD5555">'Ebinding'</span>], alpha=<span style="color: #B452CD">0.7</span>, lw=<span style="color: #B452CD">2</span>,
|
|
label=<span style="color: #CD5555">'Ame2016'</span>)
|
|
ax.plot(Masses[<span style="color: #CD5555">'A'</span>], Masses[<span style="color: #CD5555">'Eapprox'</span>], alpha=<span style="color: #B452CD">0.7</span>, lw=<span style="color: #B452CD">2</span>, c=<span style="color: #CD5555">'m'</span>,
|
|
label=<span style="color: #CD5555">'Fit'</span>)
|
|
ax.legend()
|
|
save_fig(<span style="color: #CD5555">"Masses2016OLS"</span>)
|
|
plt.show()
|
|
</pre></div>
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="adding-error-analysis-and-training-set-up">Adding error analysis and training set up </h2>
|
|
|
|
<p>
|
|
We can easily test our fit by computing the \( R2 \) score that we discussed in connection with the functionality of <b>Scikit-Learn</b> in the introductory slides.
|
|
Since we are not using <b>Scikit-Learn</b> here we can define our own \( R2 \) function as
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">R2</span>(y_data, y_model):
|
|
<span style="color: #8B008B; font-weight: bold">return</span> <span style="color: #B452CD">1</span> - np.sum((y_data - y_model) ** <span style="color: #B452CD">2</span>) / np.sum((y_data - np.mean(y_data)) ** <span style="color: #B452CD">2</span>)
|
|
</pre></div>
|
|
<p>
|
|
and we would be using it as
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #658b00">print</span>(R2(Energies,ytilde))
|
|
</pre></div>
|
|
<p>
|
|
We can easily add our <b>MSE</b> score as
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">MSE</span>(y_data,y_model):
|
|
n = np.size(y_model)
|
|
<span style="color: #8B008B; font-weight: bold">return</span> np.sum((y_data-y_model)**<span style="color: #B452CD">2</span>)/n
|
|
|
|
<span style="color: #658b00">print</span>(MSE(Energies,ytilde))
|
|
</pre></div>
|
|
<p>
|
|
and finally the relative error as
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">RelativeError</span>(y_data,y_model):
|
|
<span style="color: #8B008B; font-weight: bold">return</span> <span style="color: #658b00">abs</span>((y_data-y_model)/y_data)
|
|
<span style="color: #658b00">print</span>(RelativeError(Energies, ytilde))
|
|
</pre></div>
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="splitting-our-data-in-training-and-test-data">Splitting our Data in Training and Test data </h2>
|
|
|
|
<p>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
|
|
<p>
|
|
It is normal in essentially all Machine Learning studies to split the
|
|
data in a training set and a test set (sometimes also an additional
|
|
validation set). <b>Scikit-Learn</b> has an own function for this. There
|
|
is no explicit recipe for how much data should be included as training
|
|
data and say test data. An accepted rule of thumb is to use
|
|
approximately \( 2/3 \) to \( 4/5 \) of the data as training data. We will
|
|
postpone a discussion of this splitting to the end of these notes and
|
|
our discussion of the so-called <b>bias-variance</b> tradeoff. Here we
|
|
limit ourselves to repeat the above equation of state fitting example
|
|
but now splitting the data into a training set and a test set.
|
|
|
|
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="examples">Examples </h2>
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
|
|
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">R2</span>(y_data, y_model):
|
|
<span style="color: #8B008B; font-weight: bold">return</span> <span style="color: #B452CD">1</span> - np.sum((y_data - y_model) ** <span style="color: #B452CD">2</span>) / np.sum((y_data - np.mean(y_data)) ** <span style="color: #B452CD">2</span>)
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">MSE</span>(y_data,y_model):
|
|
n = np.size(y_model)
|
|
<span style="color: #8B008B; font-weight: bold">return</span> np.sum((y_data-y_model)**<span style="color: #B452CD">2</span>)/n
|
|
|
|
x = np.random.rand(<span style="color: #B452CD">100</span>)
|
|
y = <span style="color: #B452CD">2.0</span>+<span style="color: #B452CD">5</span>*x*x+<span style="color: #B452CD">0.1</span>*np.random.randn(<span style="color: #B452CD">100</span>)
|
|
|
|
|
|
<span style="color: #228B22"># The design matrix now as function of a given polynomial</span>
|
|
X = np.zeros((<span style="color: #658b00">len</span>(x),<span style="color: #B452CD">3</span>))
|
|
X[:,<span style="color: #B452CD">0</span>] = <span style="color: #B452CD">1.0</span>
|
|
X[:,<span style="color: #B452CD">1</span>] = x
|
|
X[:,<span style="color: #B452CD">2</span>] = x**<span style="color: #B452CD">2</span>
|
|
<span style="color: #228B22"># We split the data in test and training data</span>
|
|
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=<span style="color: #B452CD">0.2</span>)
|
|
<span style="color: #228B22"># matrix inversion to find beta</span>
|
|
beta = np.linalg.inv(X_train.T @ X_train) @ X_train.T @ y_train
|
|
<span style="color: #658b00">print</span>(beta)
|
|
<span style="color: #228B22"># and then make the prediction</span>
|
|
ytilde = X_train @ beta
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Training R2"</span>)
|
|
<span style="color: #658b00">print</span>(R2(y_train,ytilde))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Training MSE"</span>)
|
|
<span style="color: #658b00">print</span>(MSE(y_train,ytilde))
|
|
ypredict = X_test @ beta
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Test R2"</span>)
|
|
<span style="color: #658b00">print</span>(R2(y_test,ypredict))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Test MSE"</span>)
|
|
<span style="color: #658b00">print</span>(MSE(y_test,ypredict))
|
|
</pre></div>
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="making-your-own-test-train-splitting">Making your own test-train splitting </h2>
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># equivalently in numpy</span>
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">train_test_split_numpy</span>(inputs, labels, train_size, test_size):
|
|
n_inputs = <span style="color: #658b00">len</span>(inputs)
|
|
inputs_shuffled = inputs.copy()
|
|
labels_shuffled = labels.copy()
|
|
|
|
np.random.shuffle(inputs_shuffled)
|
|
np.random.shuffle(labels_shuffled)
|
|
|
|
train_end = <span style="color: #658b00">int</span>(n_inputs*train_size)
|
|
X_train, X_test = inputs_shuffled[:train_end], inputs_shuffled[train_end:]
|
|
Y_train, Y_test = labels_shuffled[:train_end], labels_shuffled[train_end:]
|
|
|
|
<span style="color: #8B008B; font-weight: bold">return</span> X_train, X_test, Y_train, Y_test
|
|
</pre></div>
|
|
<p>
|
|
But since <b>scikit-learn</b> has its own function for doing this and since
|
|
it interfaces easily with <b>tensorflow</b> and other libraries, we
|
|
normally recommend using the latter functionality.
|
|
|
|
<p>
|
|
<!-- !split -->
|
|
|
|
<h2 id="the-boston-housing-data-example">The Boston housing data example </h2>
|
|
|
|
<p>
|
|
The Boston housing
|
|
data set was originally a part of UCI Machine Learning Repository
|
|
and has been removed now. The data set is now included in <b>Scikit-Learn</b>'s
|
|
library. There are 506 samples and 13 feature (predictor) variables
|
|
in this data set. The objective is to predict the value of prices of
|
|
the house using the features (predictors) listed here.
|
|
|
|
<p>
|
|
The features/predictors are
|
|
|
|
<ol>
|
|
<li> CRIM: Per capita crime rate by town</li>
|
|
<li> ZN: Proportion of residential land zoned for lots over 25000 square feet</li>
|
|
<li> INDUS: Proportion of non-retail business acres per town</li>
|
|
<li> CHAS: Charles River dummy variable (= 1 if tract bounds river; 0 otherwise)</li>
|
|
<li> NOX: Nitric oxide concentration (parts per 10 million)</li>
|
|
<li> RM: Average number of rooms per dwelling</li>
|
|
<li> AGE: Proportion of owner-occupied units built prior to 1940</li>
|
|
<li> DIS: Weighted distances to five Boston employment centers</li>
|
|
<li> RAD: Index of accessibility to radial highways</li>
|
|
<li> TAX: Full-value property tax rate per USD10000</li>
|
|
<li> B: \( 1000(Bk - 0.63)^2 \), where \( Bk \) is the proportion of [people of African American descent] by town</li>
|
|
<li> LSTAT: Percentage of lower status of the population</li>
|
|
<li> MEDV: Median value of owner-occupied homes in USD 1000s</li>
|
|
</ol>
|
|
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="housing-data-the-code">Housing data, the code </h2>
|
|
We start by importing the libraries
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
|
|
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">seaborn</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">sns</span>
|
|
</pre></div>
|
|
<p>
|
|
and load the Boston Housing DataSet from <b>Scikit-Learn</b>
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.datasets</span> <span style="color: #8B008B; font-weight: bold">import</span> load_boston
|
|
|
|
boston_dataset = load_boston()
|
|
|
|
<span style="color: #228B22"># boston_dataset is a dictionary</span>
|
|
<span style="color: #228B22"># let's check what it contains</span>
|
|
boston_dataset.keys()
|
|
</pre></div>
|
|
<p>
|
|
Then we invoke Pandas
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>boston = pd.DataFrame(boston_dataset.data, columns=boston_dataset.feature_names)
|
|
boston.head()
|
|
boston[<span style="color: #CD5555">'MEDV'</span>] = boston_dataset.target
|
|
</pre></div>
|
|
<p>
|
|
and preprocess the data
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># check for missing values in all the columns</span>
|
|
boston.isnull().sum()
|
|
</pre></div>
|
|
<p>
|
|
We can then visualize the data
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># set the size of the figure</span>
|
|
sns.set(rc={<span style="color: #CD5555">'figure.figsize'</span>:(<span style="color: #B452CD">11.7</span>,<span style="color: #B452CD">8.27</span>)})
|
|
|
|
<span style="color: #228B22"># plot a histogram showing the distribution of the target values</span>
|
|
sns.distplot(boston[<span style="color: #CD5555">'MEDV'</span>], bins=<span style="color: #B452CD">30</span>)
|
|
plt.show()
|
|
</pre></div>
|
|
<p>
|
|
It is now useful to look at the correlation matrix
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># compute the pair wise correlation for all columns </span>
|
|
correlation_matrix = boston.corr().round(<span style="color: #B452CD">2</span>)
|
|
<span style="color: #228B22"># use the heatmap function from seaborn to plot the correlation matrix</span>
|
|
<span style="color: #228B22"># annot = True to print the values inside the square</span>
|
|
sns.heatmap(data=correlation_matrix, annot=<span style="color: #8B008B; font-weight: bold">True</span>)
|
|
</pre></div>
|
|
<p>
|
|
From the above coorelation plot we can see that <b>MEDV</b> is strongly correlated to <b>LSTAT</b> and <b>RM</b>. We see also that <b>RAD</b> and <b>TAX</b> are stronly correlated, but we don't include this in our features together to avoid multi-colinearity
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>plt.figure(figsize=(<span style="color: #B452CD">20</span>, <span style="color: #B452CD">5</span>))
|
|
|
|
features = [<span style="color: #CD5555">'LSTAT'</span>, <span style="color: #CD5555">'RM'</span>]
|
|
target = boston[<span style="color: #CD5555">'MEDV'</span>]
|
|
|
|
<span style="color: #8B008B; font-weight: bold">for</span> i, col <span style="color: #8B008B">in</span> <span style="color: #658b00">enumerate</span>(features):
|
|
plt.subplot(<span style="color: #B452CD">1</span>, <span style="color: #658b00">len</span>(features) , i+<span style="color: #B452CD">1</span>)
|
|
x = boston[col]
|
|
y = target
|
|
plt.scatter(x, y, marker=<span style="color: #CD5555">'o'</span>)
|
|
plt.title(col)
|
|
plt.xlabel(col)
|
|
plt.ylabel(<span style="color: #CD5555">'MEDV'</span>)
|
|
</pre></div>
|
|
<p>
|
|
Now we start training our model
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>X = pd.DataFrame(np.c_[boston[<span style="color: #CD5555">'LSTAT'</span>], boston[<span style="color: #CD5555">'RM'</span>]], columns = [<span style="color: #CD5555">'LSTAT'</span>,<span style="color: #CD5555">'RM'</span>])
|
|
Y = boston[<span style="color: #CD5555">'MEDV'</span>]
|
|
</pre></div>
|
|
<p>
|
|
We split the data into training and test sets
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
|
|
|
|
<span style="color: #228B22"># splits the training and test data set in 80% : 20%</span>
|
|
<span style="color: #228B22"># assign random_state to any value.This ensures consistency.</span>
|
|
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size = <span style="color: #B452CD">0.2</span>, random_state=<span style="color: #B452CD">5</span>)
|
|
<span style="color: #658b00">print</span>(X_train.shape)
|
|
<span style="color: #658b00">print</span>(X_test.shape)
|
|
<span style="color: #658b00">print</span>(Y_train.shape)
|
|
<span style="color: #658b00">print</span>(Y_test.shape)
|
|
</pre></div>
|
|
<p>
|
|
Then we use the linear regression functionality from <b>Scikit-Learn</b>
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> LinearRegression
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> mean_squared_error, r2_score
|
|
|
|
lin_model = LinearRegression()
|
|
lin_model.fit(X_train, Y_train)
|
|
|
|
<span style="color: #228B22"># model evaluation for training set</span>
|
|
|
|
y_train_predict = lin_model.predict(X_train)
|
|
rmse = (np.sqrt(mean_squared_error(Y_train, y_train_predict)))
|
|
r2 = r2_score(Y_train, y_train_predict)
|
|
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"The model performance for training set"</span>)
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"--------------------------------------"</span>)
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">'RMSE is {}'</span>.format(rmse))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">'R2 score is {}'</span>.format(r2))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"\n"</span>)
|
|
|
|
<span style="color: #228B22"># model evaluation for testing set</span>
|
|
|
|
y_test_predict = lin_model.predict(X_test)
|
|
<span style="color: #228B22"># root mean square error of the model</span>
|
|
rmse = (np.sqrt(mean_squared_error(Y_test, y_test_predict)))
|
|
|
|
<span style="color: #228B22"># r-squared score of the model</span>
|
|
r2 = r2_score(Y_test, y_test_predict)
|
|
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"The model performance for testing set"</span>)
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"--------------------------------------"</span>)
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">'RMSE is {}'</span>.format(rmse))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">'R2 score is {}'</span>.format(r2))
|
|
</pre></div>
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># plotting the y_test vs y_pred</span>
|
|
<span style="color: #228B22"># ideally should have been a straight line</span>
|
|
plt.scatter(Y_test, y_test_predict)
|
|
plt.show()
|
|
</pre></div>
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="reducing-the-number-of-degrees-of-freedom-overarching-view">Reducing the number of degrees of freedom, overarching view </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
|
|
<p>
|
|
Many Machine Learning problems involve thousands or even millions of
|
|
features for each training instance. Not only does this make training
|
|
extremely slow, it can also make it much harder to find a good
|
|
solution, as we will see. This problem is often referred to as the
|
|
curse of dimensionality. Fortunately, in real-world problems, it is
|
|
often possible to reduce the number of features considerably, turning
|
|
an intractable problem into a tractable one.
|
|
|
|
<p>
|
|
Later we will discuss some of the most popular dimensionality reduction
|
|
techniques: the principal component analysis (PCA), Kernel PCA, and
|
|
Locally Linear Embedding (LLE).
|
|
|
|
<p>
|
|
Principal component analysis and its various variants deal with the
|
|
problem of fitting a low-dimensional <a href="https://en.wikipedia.org/wiki/Affine_space" target="_blank">affine
|
|
subspace</a> to a set of of
|
|
data points in a high-dimensional space. With its family of methods it
|
|
is one of the most used tools in data modeling, compression and
|
|
visualization.
|
|
|
|
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="preprocessing-our-data">Preprocessing our data </h2>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
|
|
<p>
|
|
Before we proceed however, we will discuss how to preprocess our
|
|
data. Till now and in connection with our previous examples we have
|
|
not met so many cases where we are too sensitive to the scaling of our
|
|
data. Normally the data may need a rescaling and/or may be sensitive
|
|
to extreme values. Scaling the data renders our inputs much more
|
|
suitable for the algorithms we want to employ.
|
|
|
|
<p>
|
|
For data sets gathered for real world applications, it is rather normal that
|
|
different features have very different units and
|
|
numerical scales. For example, a data set detailing health habits may include
|
|
features such as <b>age</b> in the range \( 0-80 \), and <b>caloric intake</b> of order \( 2000 \).
|
|
Many machine learning methods sensitive to the scales of the features and may perform poorly if they
|
|
are very different scales. Therefore, it is typical to scale
|
|
the features in a way to avoid such outlier values.
|
|
|
|
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="functionality-in-scikit-learn">Functionality in Scikit-Learn </h2>
|
|
|
|
<p>
|
|
<b>Scikit-Learn</b> has several functions which allow us to rescale the
|
|
data, normally resulting in much better results in terms of various
|
|
accuracy scores. The <b>StandardScaler</b> function in <b>Scikit-Learn</b>
|
|
ensures that for each feature/predictor we study the mean value is
|
|
zero and the variance is one (every column in the design/feature
|
|
matrix). This scaling has the drawback that it does not ensure that
|
|
we have a particular maximum or minimum in our data set. Another
|
|
function included in <b>Scikit-Learn</b> is the <b>MinMaxScaler</b> which
|
|
ensures that all features are exactly between \( 0 \) and \( 1 \). The
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="more-preprocessing">More preprocessing </h2>
|
|
|
|
<p>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
The <b>Normalizer</b> scales each data
|
|
point such that the feature vector has a euclidean length of one. In other words, it
|
|
projects a data point on the circle (or sphere in the case of higher dimensions) with a
|
|
radius of 1. This means every data point is scaled by a different number (by the
|
|
inverse of it’s length).
|
|
This normalization is often used when only the direction (or angle) of the data matters,
|
|
not the length of the feature vector.
|
|
|
|
<p>
|
|
The <b>RobustScaler</b> works similarly to the StandardScaler in that it
|
|
ensures statistical properties for each feature that guarantee that
|
|
they are on the same scale. However, the RobustScaler uses the median
|
|
and quartiles, instead of mean and variance. This makes the
|
|
RobustScaler ignore data points that are very different from the rest
|
|
(like measurement errors). These odd data points are also called
|
|
outliers, and might often lead to trouble for other scaling
|
|
techniques.
|
|
|
|
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="frequently-used-scaling-functions">Frequently used scaling functions </h2>
|
|
|
|
<p>
|
|
Many features are often scaled using standardization to improve performance. In <b>Scikit-Learn</b> this is given by the <b>StandardScaler</b> function as discussed above. It is easy however to write your own.
|
|
Mathematically, this involves subtracting the mean and divide by the standard deviation over the data set, for each feature:
|
|
|
|
$$
|
|
x_j^{(i)} \rightarrow \frac{x_j^{(i)} - \overline{x}_j}{\sigma(x_j)},
|
|
$$
|
|
|
|
where \( \overline{x}_j \) and \( \sigma(x_j) \) are the mean and standard deviation, respectively, of the feature \( x_j \).
|
|
This ensures that each feature has zero mean and unit standard deviation. For data sets where we do not have the standard deviation or don't wish to calculate it, it is then common to simply set it to one.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="example-of-own-standard-scaling">Example of own Standard scaling </h2>
|
|
|
|
<p>
|
|
Let us consider the following vanilla example where we use both
|
|
<b>Scikit-Learn</b> and write our own function as well. We produce a
|
|
simple test design matrix with random numbers. Each column could then
|
|
represent a specific feature whose mean value is subracted.
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">skl</span>
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> mean_squared_error
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> MinMaxScaler, StandardScaler, Normalizer
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">IPython.display</span> <span style="color: #8B008B; font-weight: bold">import</span> display
|
|
np.random.seed(<span style="color: #B452CD">100</span>)
|
|
<span style="color: #228B22"># setting up a 10 x 5 matrix</span>
|
|
rows = <span style="color: #B452CD">10</span>
|
|
cols = <span style="color: #B452CD">5</span>
|
|
X = np.random.randn(rows,cols)
|
|
XPandas = pd.DataFrame(X)
|
|
display(XPandas)
|
|
<span style="color: #658b00">print</span>(XPandas.mean())
|
|
<span style="color: #658b00">print</span>(XPandas.std())
|
|
XPandas = (XPandas -XPandas.mean())
|
|
display(XPandas)
|
|
scaler = StandardScaler(with_std=<span style="color: #8B008B; font-weight: bold">False</span>)
|
|
scaler.fit(X)
|
|
Xscaled = scaler.transform(X)
|
|
display(XPandas-Xscaled)
|
|
</pre></div>
|
|
<p>
|
|
Small exercise: perform the standard scaling by including the standard deviation.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="min-max-scaling">Min-Max Scaling </h2>
|
|
|
|
<p>
|
|
Another commonly used scaling method is min-max scaling. This is very
|
|
useful for when we want the features to lie in a certain interval. To
|
|
scale the feature \( x_j \) to the interval \( [a, b] \), we can apply the
|
|
transformation
|
|
|
|
$$
|
|
x_j^{(i)} \rightarrow (b-a)\frac{x_j^{(i)} - \min(x_j)}{\max(x_j) - \min(x_j)} - a
|
|
$$
|
|
|
|
where \( \min(x_j) \) and \( \max(x_j) \) return the minimum and maximum value of \( x_j \) over the data set, respectively.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="testing-the-means-squared-error-as-function-of-complexity">Testing the Means Squared Error as function of Complexity </h2>
|
|
One of
|
|
the aims is to reproduce Figure 2.11 of <a href="https://github.com/CompPhysics/MLErasmus/blob/master/doc/Textbooks/elementsstat.pdf" target="_blank">Hastie et al</a>.
|
|
We will also use Ridge and Lasso regression.
|
|
|
|
<p>
|
|
Our data is defined by \( x\in [-3,3] \) with a total of for example \( 100 \) data points.
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>np.random.seed()
|
|
n = <span style="color: #B452CD">100</span>
|
|
maxdegree = <span style="color: #B452CD">14</span>
|
|
<span style="color: #228B22"># Make data set.</span>
|
|
x = np.linspace(-<span style="color: #B452CD">3</span>, <span style="color: #B452CD">3</span>, n).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
|
|
y = np.exp(-x**<span style="color: #B452CD">2</span>) + <span style="color: #B452CD">1.5</span> * np.exp(-(x-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>)+ np.random.normal(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0.1</span>, x.shape)
|
|
</pre></div>
|
|
<p>
|
|
where \( y \) is the function we want to fit with a given polynomial.
|
|
|
|
<p>
|
|
Write a first code which sets up a design matrix \( X \) defined by a fifth-order polynomial. Scale your data and split it in training and test data.
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">import</span> LinearRegression, Ridge, Lasso
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> PolynomialFeatures
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.pipeline</span> <span style="color: #8B008B; font-weight: bold">import</span> make_pipeline
|
|
|
|
|
|
np.random.seed(<span style="color: #B452CD">2018</span>)
|
|
n = <span style="color: #B452CD">50</span>
|
|
maxdegree = <span style="color: #B452CD">5</span>
|
|
<span style="color: #228B22"># Make data set.</span>
|
|
x = np.linspace(-<span style="color: #B452CD">3</span>, <span style="color: #B452CD">3</span>, n).reshape(-<span style="color: #B452CD">1</span>, <span style="color: #B452CD">1</span>)
|
|
y = np.exp(-x**<span style="color: #B452CD">2</span>) + <span style="color: #B452CD">1.5</span> * np.exp(-(x-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>)+ np.random.normal(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">0.1</span>, x.shape)
|
|
TestError = np.zeros(maxdegree)
|
|
TrainError = np.zeros(maxdegree)
|
|
polydegree = np.zeros(maxdegree)
|
|
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=<span style="color: #B452CD">0.2</span>)
|
|
scaler = StandardScaler()
|
|
scaler.fit(x_train)
|
|
x_train_scaled = scaler.transform(x_train)
|
|
x_test_scaled = scaler.transform(x_test)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">for</span> degree <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(maxdegree):
|
|
model = make_pipeline(PolynomialFeatures(degree=degree), LinearRegression(fit_intercept=<span style="color: #8B008B; font-weight: bold">False</span>))
|
|
clf = model.fit(x_train_scaled,y_train)
|
|
y_fit = clf.predict(x_train_scaled)
|
|
y_pred = clf.predict(x_test_scaled)
|
|
polydegree[degree] = degree
|
|
TestError[degree] = np.mean( np.mean((y_test - y_pred)**<span style="color: #B452CD">2</span>) )
|
|
TrainError[degree] = np.mean( np.mean((y_train - y_fit)**<span style="color: #B452CD">2</span>) )
|
|
|
|
plt.plot(polydegree, TestError, label=<span style="color: #CD5555">'Test Error'</span>)
|
|
plt.plot(polydegree, TrainError, label=<span style="color: #CD5555">'Train Error'</span>)
|
|
plt.legend()
|
|
plt.show()
|
|
</pre></div>
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="more-preprocessing-examples-franke-function-and-regression">More preprocessing examples, Franke function and regression </h2>
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># Common imports</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">sklearn.linear_model</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">skl</span>
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> mean_squared_error
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> MinMaxScaler, StandardScaler, Normalizer
|
|
|
|
<span style="color: #228B22"># Where to save the figures and data files</span>
|
|
PROJECT_ROOT_DIR = <span style="color: #CD5555">"Results"</span>
|
|
FIGURE_ID = <span style="color: #CD5555">"Results/FigureFiles"</span>
|
|
DATA_ID = <span style="color: #CD5555">"DataFiles/"</span>
|
|
|
|
<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(PROJECT_ROOT_DIR):
|
|
os.mkdir(PROJECT_ROOT_DIR)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(FIGURE_ID):
|
|
os.makedirs(FIGURE_ID)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #8B008B">not</span> os.path.exists(DATA_ID):
|
|
os.makedirs(DATA_ID)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">image_path</span>(fig_id):
|
|
<span style="color: #8B008B; font-weight: bold">return</span> os.path.join(FIGURE_ID, fig_id)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">data_path</span>(dat_id):
|
|
<span style="color: #8B008B; font-weight: bold">return</span> os.path.join(DATA_ID, dat_id)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">save_fig</span>(fig_id):
|
|
plt.savefig(image_path(fig_id) + <span style="color: #CD5555">".png"</span>, <span style="color: #658b00">format</span>=<span style="color: #CD5555">'png'</span>)
|
|
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">FrankeFunction</span>(x,y):
|
|
term1 = <span style="color: #B452CD">0.75</span>*np.exp(-(<span style="color: #B452CD">0.25</span>*(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>) - <span style="color: #B452CD">0.25</span>*((<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>))
|
|
term2 = <span style="color: #B452CD">0.75</span>*np.exp(-((<span style="color: #B452CD">9</span>*x+<span style="color: #B452CD">1</span>)**<span style="color: #B452CD">2</span>)/<span style="color: #B452CD">49.0</span> - <span style="color: #B452CD">0.1</span>*(<span style="color: #B452CD">9</span>*y+<span style="color: #B452CD">1</span>))
|
|
term3 = <span style="color: #B452CD">0.5</span>*np.exp(-(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">7</span>)**<span style="color: #B452CD">2</span>/<span style="color: #B452CD">4.0</span> - <span style="color: #B452CD">0.25</span>*((<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">3</span>)**<span style="color: #B452CD">2</span>))
|
|
term4 = -<span style="color: #B452CD">0.2</span>*np.exp(-(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">4</span>)**<span style="color: #B452CD">2</span> - (<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">7</span>)**<span style="color: #B452CD">2</span>)
|
|
<span style="color: #8B008B; font-weight: bold">return</span> term1 + term2 + term3 + term4
|
|
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">create_X</span>(x, y, n ):
|
|
<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #658b00">len</span>(x.shape) > <span style="color: #B452CD">1</span>:
|
|
x = np.ravel(x)
|
|
y = np.ravel(y)
|
|
|
|
N = <span style="color: #658b00">len</span>(x)
|
|
l = <span style="color: #658b00">int</span>((n+<span style="color: #B452CD">1</span>)*(n+<span style="color: #B452CD">2</span>)/<span style="color: #B452CD">2</span>) <span style="color: #228B22"># Number of elements in beta</span>
|
|
X = np.ones((N,l))
|
|
|
|
<span style="color: #8B008B; font-weight: bold">for</span> i <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(<span style="color: #B452CD">1</span>,n+<span style="color: #B452CD">1</span>):
|
|
q = <span style="color: #658b00">int</span>((i)*(i+<span style="color: #B452CD">1</span>)/<span style="color: #B452CD">2</span>)
|
|
<span style="color: #8B008B; font-weight: bold">for</span> k <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(i+<span style="color: #B452CD">1</span>):
|
|
X[:,q+k] = (x**(i-k))*(y**k)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">return</span> X
|
|
|
|
|
|
<span style="color: #228B22"># Making meshgrid of datapoints and compute Franke's function</span>
|
|
n = <span style="color: #B452CD">5</span>
|
|
N = <span style="color: #B452CD">1000</span>
|
|
x = np.sort(np.random.uniform(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, N))
|
|
y = np.sort(np.random.uniform(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, N))
|
|
z = FrankeFunction(x, y)
|
|
X = create_X(x, y, n=n)
|
|
<span style="color: #228B22"># split in training and test data</span>
|
|
X_train, X_test, y_train, y_test = train_test_split(X,z,test_size=<span style="color: #B452CD">0.2</span>)
|
|
|
|
|
|
clf = skl.LinearRegression().fit(X_train, y_train)
|
|
|
|
<span style="color: #228B22"># The mean squared error and R2 score</span>
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"MSE before scaling: {:.2f}"</span>.format(mean_squared_error(clf.predict(X_test), y_test)))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"R2 score before scaling {:.2f}"</span>.format(clf.score(X_test,y_test)))
|
|
|
|
scaler = StandardScaler()
|
|
scaler.fit(X_train)
|
|
X_train_scaled = scaler.transform(X_train)
|
|
X_test_scaled = scaler.transform(X_test)
|
|
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Feature min values before scaling:\n {}"</span>.format(X_train.min(axis=<span style="color: #B452CD">0</span>)))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Feature max values before scaling:\n {}"</span>.format(X_train.max(axis=<span style="color: #B452CD">0</span>)))
|
|
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Feature min values after scaling:\n {}"</span>.format(X_train_scaled.min(axis=<span style="color: #B452CD">0</span>)))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Feature max values after scaling:\n {}"</span>.format(X_train_scaled.max(axis=<span style="color: #B452CD">0</span>)))
|
|
|
|
clf = skl.LinearRegression().fit(X_train_scaled, y_train)
|
|
|
|
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"MSE after scaling: {:.2f}"</span>.format(mean_squared_error(clf.predict(X_test_scaled), y_test)))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"R2 score for scaled data: {:.2f}"</span>.format(clf.score(X_test_scaled,y_test)))
|
|
</pre></div>
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="mathematical-interpretation-of-ordinary-least-squares">Mathematical Interpretation of Ordinary Least Squares </h2>
|
|
|
|
<p>
|
|
What is presented here is a mathematical analysis of various regression algorithms (ordinary least squares, Ridge and Lasso Regression). The analysis is based on an important algorithm in linear algebra, the so-called Singular Value Decomposition (SVD).
|
|
|
|
<p>
|
|
We have shown that in ordinary least squares that the optimal parameter \( \beta \) are given by
|
|
|
|
$$
|
|
\hat{\boldsymbol{\beta}} = \left(\boldsymbol{X}^T\boldsymbol{X}\right^{-1}\boldsymbol{X}^T\boldsymbol{y}.
|
|
$$
|
|
|
|
<p>
|
|
This means that our best model is defined as
|
|
|
|
$$
|
|
\tilde{\boldsymbol{y}}=\boldsymbol{X}\hat{\boldsymbol{\beta}} = \boldsymbol{X}\left(\boldsymbol{X}^T\boldsymbol{X}\right^{-1}\boldsymbol{X}^T\boldsymbol{y}.
|
|
$$
|
|
|
|
<p>
|
|
We define now a matrix
|
|
$$
|
|
\boldsymbol{A}=\boldsymbol{X}\left(\boldsymbol{X}^T\boldsymbol{X}\right^{-1}\boldsymbol{X}^T.
|
|
$$
|
|
|
|
<p>
|
|
It means that we can rewrite
|
|
$$
|
|
\tilde{\boldsymbol{y}}=\boldsymbol{X}\hat{\boldsymbol{\beta}} = \boldsymbol{A}\boldsymbol{y}.
|
|
$$
|
|
|
|
<p>
|
|
The matrix \( \boldsymbol{A} \) has the important property that \( \boldsymbol{A}^2=\boldsymbol{A} \). This is the definition of a projection matrix.
|
|
We can then interpret that our optimal model \( \tilde{\boldsymbol{y}} \) ir represented by an orthogonal (it has to be a square matrix like ours, if not we have an oblique projection matrix ) projection of \( \boldsymbol{y} \) onto a space defined by the column vectors of \( \boldsymbol{X} \).
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="residual-error">Residual Error </h2>
|
|
|
|
<p>
|
|
We have defined the residual error as
|
|
$$
|
|
\boldsymbol{\epsilon}=\boldsymbol{y}-\tilde{\boldsymbol{y}}=(\boldsymbol{I}-\boldsymbol{X}\left(\boldsymbol{X}^T\boldsymbol{X}\right^{-1}\boldsymbol{X}^T)\boldsymbol{y}.
|
|
$$
|
|
|
|
<p>
|
|
The residual errors are then the projections of \( \boldsymbol{y} \) onto the orthogonal components of the space defined by the column vectors of \( \boldsymbol{X} \).
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="simple-case">Simple case </h2>
|
|
|
|
<p>
|
|
If the matrix \( \boldsymbol{X} \) is an orthogonal (or unitary in case of complex values) matrix, we have
|
|
|
|
$$
|
|
\boldsymbol{X}^T\boldsymbol{X}=\boldsymbol{X}\boldsymbol{X}^T = \boldsymbol{I}.
|
|
$$
|
|
|
|
<p>
|
|
In this case the matrix \( \boldsymbol{A} \) becomes
|
|
$$
|
|
\boldsymbol{A}=\boldsymbol{X}\left(\boldsymbol{X}^T\boldsymbol{X}\right^{-1}\boldsymbol{X}^T)=\boldsymbol{I},
|
|
$$
|
|
|
|
and we have the obvious case
|
|
$$
|
|
\boldsymbol{\epsilon}=\boldsymbol{y}-\tilde{\boldsymbol{y}}=0.
|
|
$$
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="the-singular-value-decomposition">The singular value decomposition </h2>
|
|
|
|
<p>
|
|
<div class="alert alert-block alert-block alert-text-normal">
|
|
<b></b>
|
|
<p>
|
|
|
|
<p>
|
|
The examples we have looked at so far are cases where we normally can
|
|
invert the matrix \( \boldsymbol{X}^T\boldsymbol{X} \). Using a polynomial expansion as we
|
|
did both for the masses and the fitting of various functions leads to
|
|
row vectors of the design matrix which are essentially orthogonal due
|
|
to the polynomial character of our model. Obtaining the inverse of the
|
|
design matrix is then often done via a so-called LU, QR or Cholesky
|
|
decomposition.
|
|
|
|
<p>
|
|
As we will also see in the first project,
|
|
this may
|
|
however not the be case in general and a standard matrix inversion
|
|
algorithm based on say LU, QR or Cholesky decomposition may lead to singularities. We will see examples of this below.
|
|
|
|
<p>
|
|
There is however a way to partially circumvent this problem and also gain some insights about the ordinary least squares approach, and later shrinkage methods like Ridge and Lasso regressions.
|
|
|
|
<p>
|
|
This is given by the <b>Singular Value Decomposition</b> algorithm, perhaps
|
|
the most powerful linear algebra algorithm. Let us look at a
|
|
different example where we may have problems with the standard matrix
|
|
inversion algorithm. Thereafter we dive into the math of the SVD.
|
|
|
|
|
|
</div>
|
|
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="linear-regression-problems">Linear Regression Problems </h2>
|
|
|
|
<p>
|
|
One of the typical problems we encounter with linear regression, in particular
|
|
when the matrix \( \boldsymbol{X} \) (our so-called design matrix) is high-dimensional,
|
|
are problems with near singular or singular matrices. The column vectors of \( \boldsymbol{X} \)
|
|
may be linearly dependent, normally referred to as super-collinearity.
|
|
This means that the matrix may be rank deficient and it is basically impossible to
|
|
to model the data using linear regression. As an example, consider the matrix
|
|
$$
|
|
\begin{align*}
|
|
\mathbf{X} & = \left[
|
|
\begin{array}{rrr}
|
|
1 & -1 & 2
|
|
\\
|
|
1 & 0 & 1
|
|
\\
|
|
1 & 2 & -1
|
|
\\
|
|
1 & 1 & 0
|
|
\end{array} \right]
|
|
\end{align*}
|
|
$$
|
|
|
|
<p>
|
|
The columns of \( \boldsymbol{X} \) are linearly dependent. We see this easily since the
|
|
the first column is the row-wise sum of the other two columns. The rank (more correct,
|
|
the column rank) of a matrix is the dimension of the space spanned by the
|
|
column vectors. Hence, the rank of \( \mathbf{X} \) is equal to the number
|
|
of linearly independent columns. In this particular case the matrix has rank 2.
|
|
|
|
<p>
|
|
Super-collinearity of an \( (n \times p) \)-dimensional design matrix \( \mathbf{X} \) implies
|
|
that the inverse of the matrix \( \boldsymbol{X}^T\boldsymbol{X} \) (the matrix we need to invert to solve the linear regression equations) is non-invertible. If we have a square matrix that does not have an inverse, we say this matrix singular. The example here demonstrates this
|
|
$$
|
|
\begin{align*}
|
|
\boldsymbol{X} & = \left[
|
|
\begin{array}{rr}
|
|
1 & -1
|
|
\\
|
|
1 & -1
|
|
\end{array} \right].
|
|
\end{align*}
|
|
$$
|
|
|
|
We see easily that \( \mbox{det}(\boldsymbol{X}) = x_{11} x_{22} - x_{12} x_{21} = 1 \times (-1) - 1 \times (-1) = 0 \). Hence, \( \mathbf{X} \) is singular and its inverse is undefined.
|
|
This is equivalent to saying that the matrix \( \boldsymbol{X} \) has at least an eigenvalue which is zero.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="fixing-the-singularity">Fixing the singularity </h2>
|
|
|
|
<p>
|
|
If our design matrix \( \boldsymbol{X} \) which enters the linear regression problem
|
|
$$
|
|
\begin{align}
|
|
\boldsymbol{\beta} & = (\boldsymbol{X}^{T} \boldsymbol{X})^{-1} \boldsymbol{X}^{T} \boldsymbol{y},
|
|
\label{_auto1}
|
|
\end{align}
|
|
$$
|
|
|
|
has linearly dependent column vectors, we will not be able to compute the inverse
|
|
of \( \boldsymbol{X}^T\boldsymbol{X} \) and we cannot find the parameters (estimators) \( \beta_i \).
|
|
The estimators are only well-defined if \( (\boldsymbol{X}^{T}\boldsymbol{X})^{-1} \) exits.
|
|
This is more likely to happen when the matrix \( \boldsymbol{X} \) is high-dimensional. In this case it is likely to encounter a situation where
|
|
the regression parameters \( \beta_i \) cannot be estimated.
|
|
|
|
<p>
|
|
A cheap <em>ad hoc</em> approach is simply to add a small diagonal component to the matrix to invert, that is we change
|
|
$$
|
|
\boldsymbol{X}^{T} \boldsymbol{X} \rightarrow \boldsymbol{X}^{T} \boldsymbol{X}+\lambda \boldsymbol{I},
|
|
$$
|
|
|
|
where \( \boldsymbol{I} \) is the identity matrix. When we discuss <b>Ridge</b> regression this is actually what we end up evaluating. The parameter \( \lambda \) is called a hyperparameter. More about this later.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="basic-math-of-the-svd">Basic math of the SVD </h2>
|
|
|
|
<p>
|
|
From standard linear algebra we know that a square matrix \( \boldsymbol{X} \) can be diagonalized if and only it is
|
|
a so-called <a href="https://en.wikipedia.org/wiki/Normal_matrix" target="_blank">normal matrix</a>, that is if \( \boldsymbol{X}\in {\mathbb{R}}^{n\times n} \)
|
|
we have \( \boldsymbol{X}\boldsymbol{X}^T=\boldsymbol{X}^T\boldsymbol{X} \) or if \( \boldsymbol{X}\in {\mathbb{C}}^{n\times n} \) we have \( \boldsymbol{X}\boldsymbol{X}^{\dagger}=\boldsymbol{X}^{\dagger}\boldsymbol{X} \).
|
|
The matrix has then a set of eigenpairs
|
|
|
|
$$
|
|
(\lambda_1,\boldsymbol{u}_1),\dots, (\lambda_n,\boldsymbol{u}_n),
|
|
$$
|
|
|
|
and the eigenvalues are given by the diagonal matrix
|
|
$$
|
|
\boldsymbol{\Sigma}=\mathrm{Diag}(\lambda_1, \dots,\lambda_n).
|
|
$$
|
|
|
|
The matrix \( \boldsymbol{X} \) can be written in terms of an orthogonal/unitary transformation \( \boldsymbol{U} \)
|
|
$$
|
|
\boldsymbol{X} = \boldsymbol{U}\boldsymbol{\Sigma}\boldsymbol{V}^T,
|
|
$$
|
|
|
|
with \( \boldsymbol{U}\boldsymbol{U}^T=\boldsymbol{I} \) or \( \boldsymbol{U}\boldsymbol{U}^{\dagger}=\boldsymbol{I} \).
|
|
|
|
<p>
|
|
Not all square matrices are diagonalizable. A matrix like the one discussed above
|
|
$$
|
|
\boldsymbol{X} = \begin{bmatrix}
|
|
1& -1 \\
|
|
1& -1\\
|
|
\end{bmatrix}
|
|
$$
|
|
|
|
is not diagonalizable, it is a so-called <a href="https://en.wikipedia.org/wiki/Defective_matrix" target="_blank">defective matrix</a>. It is easy to see that the condition
|
|
\( \boldsymbol{X}\boldsymbol{X}^T=\boldsymbol{X}^T\boldsymbol{X} \) is not fulfilled.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="the-svd-a-fantastic-algorithm">The SVD, a Fantastic Algorithm </h2>
|
|
|
|
<p>
|
|
However, and this is the strength of the SVD algorithm, any general
|
|
matrix \( \boldsymbol{X} \) can be decomposed in terms of a diagonal matrix and
|
|
two orthogonal/unitary matrices. The <a href="https://en.wikipedia.org/wiki/Singular_value_decomposition" target="_blank">Singular Value Decompostion
|
|
(SVD) theorem</a>
|
|
states that a general \( m\times n \) matrix \( \boldsymbol{X} \) can be written in
|
|
terms of a diagonal matrix \( \boldsymbol{\Sigma} \) of dimensionality \( m\times n \)
|
|
and two orthognal matrices \( \boldsymbol{U} \) and \( \boldsymbol{V} \), where the first has
|
|
dimensionality \( m \times m \) and the last dimensionality \( n\times n \).
|
|
We have then
|
|
|
|
$$
|
|
\boldsymbol{X} = \boldsymbol{U}\boldsymbol{\Sigma}\boldsymbol{V}^T
|
|
$$
|
|
|
|
<p>
|
|
As an example, the above defective matrix can be decomposed as
|
|
|
|
$$
|
|
\boldsymbol{X} = \frac{1}{\sqrt{2}}\begin{bmatrix} 1& 1 \\ 1& -1\\ \end{bmatrix} \begin{bmatrix} 2& 0 \\ 0& 0\\ \end{bmatrix} \frac{1}{\sqrt{2}}\begin{bmatrix} 1& -1 \\ 1& 1\\ \end{bmatrix}=\boldsymbol{U}\boldsymbol{\Sigma}\boldsymbol{V}^T,
|
|
$$
|
|
|
|
<p>
|
|
with eigenvalues \( \sigma_1=2 \) and \( \sigma_2=0 \).
|
|
The SVD exits always!
|
|
|
|
<p>
|
|
The SVD
|
|
decomposition (singular values) gives eigenvalues
|
|
\( \sigma_i\geq\sigma_{i+1} \) for all \( i \) and for dimensions larger than \( i=p \), the
|
|
eigenvalues (singular values) are zero.
|
|
|
|
<p>
|
|
In the general case, where our design matrix \( \boldsymbol{X} \) has dimension
|
|
\( n\times p \), the matrix is thus decomposed into an \( n\times n \)
|
|
orthogonal matrix \( \boldsymbol{U} \), a \( p\times p \) orthogonal matrix \( \boldsymbol{V} \)
|
|
and a diagonal matrix \( \boldsymbol{\Sigma} \) with \( r=\mathrm{min}(n,p) \)
|
|
singular values \( \sigma_i\geq 0 \) on the main diagonal and zeros filling
|
|
the rest of the matrix. There are at most \( p \) singular values
|
|
assuming that \( n > p \). In our regression examples for the nuclear
|
|
masses and the equation of state this is indeed the case, while for
|
|
the Ising model we have \( p > n \). These are often cases that lead to
|
|
near singular or singular matrices.
|
|
|
|
<p>
|
|
The columns of \( \boldsymbol{U} \) are called the left singular vectors while the columns of \( \boldsymbol{V} \) are the right singular vectors.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="economy-size-svd">Economy-size SVD </h2>
|
|
|
|
<p>
|
|
If we assume that \( n > p \), then our matrix \( \boldsymbol{U} \) has dimension \( n
|
|
\times n \). The last \( n-p \) columns of \( \boldsymbol{U} \) become however
|
|
irrelevant in our calculations since they are multiplied with the
|
|
zeros in \( \boldsymbol{\Sigma} \).
|
|
|
|
<p>
|
|
The economy-size decomposition removes extra rows or columns of zeros
|
|
from the diagonal matrix of singular values, \( \boldsymbol{\Sigma} \), along with the columns
|
|
in either \( \boldsymbol{U} \) or \( \boldsymbol{V} \) that multiply those zeros in the expression.
|
|
Removing these zeros and columns can improve execution time
|
|
and reduce storage requirements without compromising the accuracy of
|
|
the decomposition.
|
|
|
|
<p>
|
|
If \( n > p \), we keep only the first \( p \) columns of \( \boldsymbol{U} \) and \( \boldsymbol{\Sigma} \) has dimension \( p\times p \).
|
|
If \( p > n \), then only the first \( n \) columns of \( \boldsymbol{V} \) are computed and \( \boldsymbol{\Sigma} \) has dimension \( n\times n \).
|
|
The \( n=p \) case is obvious, we retain the full SVD.
|
|
In general the economy-size SVD leads to less FLOPS and still conserving the desired accuracy.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="codes-for-the-svd">Codes for the SVD </h2>
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
<span style="color: #228B22"># SVD inversion</span>
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">SVDinv</span>(A):
|
|
<span style="color: #CD5555">''' Takes as input a numpy matrix A and returns inv(A) based on singular value decomposition (SVD).</span>
|
|
<span style="color: #CD5555"> SVD is numerically more stable than the inversion algorithms provided by</span>
|
|
<span style="color: #CD5555"> numpy and scipy.linalg at the cost of being slower.</span>
|
|
<span style="color: #CD5555"> '''</span>
|
|
U, S, VT = np.linalg.svd(A,full_matrices=<span style="color: #8B008B; font-weight: bold">True</span>)
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">'test U'</span>)
|
|
<span style="color: #658b00">print</span>( (np.transpose(U) @ U - U <span style="color: #707a7c">@np</span>.transpose(U)))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">'test VT'</span>)
|
|
<span style="color: #658b00">print</span>( (np.transpose(VT) @ VT - VT <span style="color: #707a7c">@np</span>.transpose(VT)))
|
|
<span style="color: #658b00">print</span>(U)
|
|
<span style="color: #658b00">print</span>(S)
|
|
<span style="color: #658b00">print</span>(VT)
|
|
|
|
D = np.zeros((<span style="color: #658b00">len</span>(U),<span style="color: #658b00">len</span>(VT)))
|
|
<span style="color: #8B008B; font-weight: bold">for</span> i <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(<span style="color: #B452CD">0</span>,<span style="color: #658b00">len</span>(VT)):
|
|
D[i,i]=S[i]
|
|
<span style="color: #8B008B; font-weight: bold">return</span> U @ D @ VT
|
|
|
|
|
|
X = np.array([ [<span style="color: #B452CD">1.0</span>,-<span style="color: #B452CD">1.0</span>], [<span style="color: #B452CD">1.0</span>,-<span style="color: #B452CD">1.0</span>]])
|
|
<span style="color: #228B22">#X = np.array([[1, 2], [3, 4], [5, 6]])</span>
|
|
|
|
<span style="color: #658b00">print</span>(X)
|
|
C = SVDinv(X)
|
|
<span style="color: #228B22"># Print the difference between the original matrix and the SVD one</span>
|
|
<span style="color: #658b00">print</span>(C-X)
|
|
</pre></div>
|
|
<p>
|
|
The matrix \( \boldsymbol{X} \) has columns that are linearly dependent. The first
|
|
column is the row-wise sum of the other two columns. The rank of a
|
|
matrix (the column rank) is the dimension of space spanned by the
|
|
column vectors. The rank of the matrix is the number of linearly
|
|
independent columns, in this case just \( 2 \). We see this from the
|
|
singular values when running the above code. Running the standard
|
|
inversion algorithm for matrix inversion with \( \boldsymbol{X}^T\boldsymbol{X} \) results
|
|
in the program terminating due to a singular matrix.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="note-about-svd-calculations">Note about SVD Calculations </h2>
|
|
|
|
<p>
|
|
The \( U \), \( S \), and \( V \) matrices returned from the <b>svd()</b> function
|
|
cannot be multiplied directly.
|
|
|
|
<p>
|
|
As you can see from the code, the \( S \)
|
|
vector must be converted into a diagonal matrix. This may cause a
|
|
as
|
|
the size of the matrices do not fit the rules of matrix
|
|
multiplication, where the number of columns in a matrix must match the
|
|
number of rows in the subsequent matrix.
|
|
|
|
<p>
|
|
If you wish to include the zero singular values, you will need to resize the matrices. More about this later.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="mathematical-properties">Mathematical Properties </h2>
|
|
|
|
<p>
|
|
There are several interesting mathematical properties which will be
|
|
relevant when we are going to discuss the differences between say
|
|
ordinary least squares (OLS) and <b>Ridge</b> regression.
|
|
|
|
<p>
|
|
We have from OLS that the parameters of the linear approximation are given by
|
|
$$
|
|
\boldsymbol{\tilde{y}} = \boldsymbol{X}\boldsymbol{\beta} = \boldsymbol{X}\left(\boldsymbol{X}^T\boldsymbol{X}\right)^{-1}\boldsymbol{X}^T\boldsymbol{y}.
|
|
$$
|
|
|
|
<p>
|
|
The matrix to invert can be rewritten in terms of our SVD decomposition as
|
|
|
|
$$
|
|
\boldsymbol{X}^T\boldsymbol{X} = \boldsymbol{V}\boldsymbol{\Sigma}^T\boldsymbol{U}^T\boldsymbol{U}\boldsymbol{\Sigma}\boldsymbol{V}^T.
|
|
$$
|
|
|
|
Using the orthogonality properties of \( \boldsymbol{U} \) we have
|
|
|
|
$$
|
|
\boldsymbol{X}^T\boldsymbol{X} = \boldsymbol{V}\boldsymbol{\Sigma}^T\boldsymbol{\Sigma}\boldsymbol{V}^T = \boldsymbol{V}\boldsymbol{D}\boldsymbol{V}^T,
|
|
$$
|
|
|
|
with \( \boldsymbol{D} \) being a diagonal matrix with values along the diagonal given by the singular values squared.
|
|
|
|
<p>
|
|
This means that
|
|
$$
|
|
(\boldsymbol{X}^T\boldsymbol{X})\boldsymbol{V} = \boldsymbol{V}\boldsymbol{D},
|
|
$$
|
|
|
|
that is the eigenvectors of \( (\boldsymbol{X}^T\boldsymbol{X}) \) are given by the columns of the right singular matrix of \( \boldsymbol{X} \) and the eigenvalues are the squared singular values. It is easy to show (show this) that
|
|
$$
|
|
(\boldsymbol{X}\boldsymbol{X}^T)\boldsymbol{U} = \boldsymbol{U}\boldsymbol{D},
|
|
$$
|
|
|
|
that is, the eigenvectors of \( (\boldsymbol{X}\boldsymbol{X})^T \) are the columns of the left singular matrix and the eigenvalues are the same.
|
|
|
|
<p>
|
|
Going back to our OLS equation we have
|
|
$$
|
|
\boldsymbol{X}\boldsymbol{\beta} = \boldsymbol{X}\left(\boldsymbol{V}\boldsymbol{D}\boldsymbol{V}^T \right)^{-1}\boldsymbol{X}^T\boldsymbol{y}=\boldsymbol{U\Sigma V^T}\left(\boldsymbol{V}\boldsymbol{D}\boldsymbol{V}^T \right)^{-1}(\boldsymbol{U\Sigma V^T})^T\boldsymbol{y}=\boldsymbol{U}\boldsymbol{U}^T\boldsymbol{y}.
|
|
$$
|
|
|
|
We will come back to this expression when we discuss Ridge regression.
|
|
|
|
$$ \tilde{y}^{OLS}={\bf X}\hat{\beta}^{OLS}=\sum_{j=1}^p {\bf u}_j{\bf u}_j^T{\bf y}$$ and for Ridge we have 
|
|
|
|
$$ \tilde{y}^{Ridge}={\bf X}\hat{\beta}^{Ridge}=\sum_{j=1}^p {\bf u}_j\frac{\sigma_j^2}{\sigma_j^2+\lambda}{\bf u}_j^T{\bf y}$$ . 
|
|
|
|
<p>
|
|
It is indeed the economy-sized SVD, note the summation runs up tp $$p$$ only and not $$n$$. 
|
|
|
|
<p>
|
|
Here we have that $${\bf X} = {\bf U}{\bf \Sigma}{\bf V}^T$$, with $$\Sigma$$ being an $$ n\times p$$ matrix and $${\bf V}$$ being a $$ p\times p$$ matrix. We also have assumed here that $$ n > p$$. 
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="friday-september-3">Friday September 3 </h2>
|
|
|
|
<p>
|
|
<a href="https://www.uio.no/studier/emner/matnat/fys/FYS-STK4155/h20/forelesningsvideoer/LectureSeptember11.mp4?vrtx=view-as-webpage" target="_blank">Video of Lecture from 2020</a> and <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/HandWrittenNotes/NotesSeptember11.pdf" target="_blank">handwritten notes</a>
|
|
|
|
<p>
|
|
More material will be added here, see handwritten notes also.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="ridge-and-lasso-regression">Ridge and LASSO Regression </h2>
|
|
|
|
<p>
|
|
Let us remind ourselves about the expression for the standard Mean Squared Error (MSE) which we used to define our cost function and the equations for the ordinary least squares (OLS) method, that is
|
|
our optimization problem is
|
|
$$
|
|
{\displaystyle \min_{\boldsymbol{\beta}\in {\mathbb{R}}^{p}}}\frac{1}{n}\left\{\left(\boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right)^T\left(\boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\right)\right\}.
|
|
$$
|
|
|
|
or we can state it as
|
|
$$
|
|
{\displaystyle \min_{\boldsymbol{\beta}\in
|
|
{\mathbb{R}}^{p}}}\frac{1}{n}\sum_{i=0}^{n-1}\left(y_i-\tilde{y}_i\right)^2=\frac{1}{n}\vert\vert \boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\vert\vert_2^2,
|
|
$$
|
|
|
|
where we have used the definition of a norm-2 vector, that is
|
|
$$
|
|
\vert\vert \boldsymbol{x}\vert\vert_2 = \sqrt{\sum_i x_i^2}.
|
|
$$
|
|
|
|
<p>
|
|
By minimizing the above equation with respect to the parameters
|
|
\( \boldsymbol{\beta} \) we could then obtain an analytical expression for the
|
|
parameters \( \boldsymbol{\beta} \). We can add a regularization parameter \( \lambda \) by
|
|
defining a new cost function to be optimized, that is
|
|
|
|
$$
|
|
{\displaystyle \min_{\boldsymbol{\beta}\in
|
|
{\mathbb{R}}^{p}}}\frac{1}{n}\vert\vert \boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\vert\vert_2^2+\lambda\vert\vert \boldsymbol{\beta}\vert\vert_2^2
|
|
$$
|
|
|
|
<p>
|
|
which leads to the Ridge regression minimization problem where we
|
|
require that \( \vert\vert \boldsymbol{\beta}\vert\vert_2^2\le t \), where \( t \) is
|
|
a finite number larger than zero. By defining
|
|
|
|
$$
|
|
C(\boldsymbol{X},\boldsymbol{\beta})=\frac{1}{n}\vert\vert \boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\vert\vert_2^2+\lambda\vert\vert \boldsymbol{\beta}\vert\vert_1,
|
|
$$
|
|
|
|
<p>
|
|
we have a new optimization equation
|
|
$$
|
|
{\displaystyle \min_{\boldsymbol{\beta}\in
|
|
{\mathbb{R}}^{p}}}\frac{1}{n}\vert\vert \boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta}\vert\vert_2^2+\lambda\vert\vert \boldsymbol{\beta}\vert\vert_1
|
|
$$
|
|
|
|
which leads to Lasso regression. Lasso stands for least absolute shrinkage and selection operator.
|
|
|
|
<p>
|
|
Here we have defined the norm-1 as
|
|
$$
|
|
\vert\vert \boldsymbol{x}\vert\vert_1 = \sum_i \vert x_i\vert.
|
|
$$
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="more-on-ridge-regression">More on Ridge Regression </h2>
|
|
|
|
<p>
|
|
Using the matrix-vector expression for Ridge regression,
|
|
|
|
$$
|
|
C(\boldsymbol{X},\boldsymbol{\beta})=\frac{1}{n}\left\{(\boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta})^T(\boldsymbol{y}-\boldsymbol{X}\boldsymbol{\beta})\right\}+\lambda\boldsymbol{\beta}^T\boldsymbol{\beta},
|
|
$$
|
|
|
|
<p>
|
|
by taking the derivatives with respect to \( \boldsymbol{\beta} \) we obtain then
|
|
a slightly modified matrix inversion problem which for finite values
|
|
of \( \lambda \) does not suffer from singularity problems. We obtain
|
|
|
|
$$
|
|
\boldsymbol{\beta}^{\mathrm{Ridge}} = \left(\boldsymbol{X}^T\boldsymbol{X}+\lambda\boldsymbol{I}\right)^{-1}\boldsymbol{X}^T\boldsymbol{y},
|
|
$$
|
|
|
|
<p>
|
|
with \( \boldsymbol{I} \) being a \( p\times p \) identity matrix with the constraint that
|
|
|
|
$$
|
|
\sum_{i=0}^{p-1} \beta_i^2 \leq t,
|
|
$$
|
|
|
|
<p>
|
|
with \( t \) a finite positive number.
|
|
|
|
<p>
|
|
We see that Ridge regression is nothing but the standard
|
|
OLS with a modified diagonal term added to \( \boldsymbol{X}^T\boldsymbol{X} \). The
|
|
consequences, in particular for our discussion of the bias-variance tradeoff
|
|
are rather interesting.
|
|
|
|
<p>
|
|
Furthermore, if we use the result above in terms of the SVD decomposition (our analysis was done for the OLS method), we had
|
|
$$
|
|
(\boldsymbol{X}\boldsymbol{X}^T)\boldsymbol{U} = \boldsymbol{U}\boldsymbol{D}.
|
|
$$
|
|
|
|
<p>
|
|
We can analyse the OLS solutions in terms of the eigenvectors (the columns) of the right singular value matrix \( \boldsymbol{U} \) as
|
|
$$
|
|
\boldsymbol{X}\boldsymbol{\beta} = \boldsymbol{X}\left(\boldsymbol{V}\boldsymbol{D}\boldsymbol{V}^T \right)^{-1}\boldsymbol{X}^T\boldsymbol{y}=\boldsymbol{U\Sigma V^T}\left(\boldsymbol{V}\boldsymbol{D}\boldsymbol{V}^T \right)^{-1}(\boldsymbol{U\Sigma V^T})^T\boldsymbol{y}=\boldsymbol{U}\boldsymbol{U}^T\boldsymbol{y}
|
|
$$
|
|
|
|
<p>
|
|
For Ridge regression this becomes
|
|
|
|
$$
|
|
\boldsymbol{X}\boldsymbol{\beta}^{\mathrm{Ridge}} = \boldsymbol{U\Sigma V^T}\left(\boldsymbol{V}\boldsymbol{D}\boldsymbol{V}^T+\lambda\boldsymbol{I} \right)^{-1}(\boldsymbol{U\Sigma V^T})^T\boldsymbol{y}=\sum_{j=0}^{p-1}\boldsymbol{u}_j\boldsymbol{u}_j^T\frac{\sigma_j^2}{\sigma_j^2+\lambda}\boldsymbol{y},
|
|
$$
|
|
|
|
<p>
|
|
with the vectors \( \boldsymbol{u}_j \) being the columns of \( \boldsymbol{U} \).
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="interpreting-the-ridge-results">Interpreting the Ridge results </h2>
|
|
|
|
<p>
|
|
Since \( \lambda \geq 0 \), it means that compared to OLS, we have
|
|
|
|
$$
|
|
\frac{\sigma_j^2}{\sigma_j^2+\lambda} \leq 1.
|
|
$$
|
|
|
|
<p>
|
|
Ridge regression finds the coordinates of \( \boldsymbol{y} \) with respect to the
|
|
orthonormal basis \( \boldsymbol{U} \), it then shrinks the coordinates by
|
|
\( \frac{\sigma_j^2}{\sigma_j^2+\lambda} \). Recall that the SVD has
|
|
eigenvalues ordered in a descending way, that is \( \sigma_i \geq
|
|
\sigma_{i+1} \).
|
|
|
|
<p>
|
|
For small eigenvalues \( \sigma_i \) it means that their contributions become less important, a fact which can be used to reduce the number of degrees of freedom.
|
|
Actually, calculating the variance of \( \boldsymbol{X}\boldsymbol{v}_j \) shows that this quantity is equal to \( \sigma_j^2/n \).
|
|
With a parameter \( \lambda \) we can thus shrink the role of specific parameters.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="more-interpretations">More interpretations </h2>
|
|
|
|
<p>
|
|
For the sake of simplicity, let us assume that the design matrix is orthonormal, that is
|
|
|
|
$$
|
|
\boldsymbol{X}^T\boldsymbol{X}=(\boldsymbol{X}^T\boldsymbol{X})^{-1} =\boldsymbol{I}.
|
|
$$
|
|
|
|
<p>
|
|
In this case the standard OLS results in
|
|
$$
|
|
\boldsymbol{\beta}^{\mathrm{OLS}} = \boldsymbol{X}^T\boldsymbol{y}=\sum_{i=0}^{p-1}\boldsymbol{u}_j\boldsymbol{u}_j^T\boldsymbol{y},
|
|
$$
|
|
|
|
<p>
|
|
and
|
|
|
|
$$
|
|
\boldsymbol{\beta}^{\mathrm{Ridge}} = \left(\boldsymbol{I}+\lambda\boldsymbol{I}\right)^{-1}\boldsymbol{X}^T\boldsymbol{y}=\left(1+\lambda\right)^{-1}\boldsymbol{\beta}^{\mathrm{OLS}},
|
|
$$
|
|
|
|
<p>
|
|
that is the Ridge estimator scales the OLS estimator by the inverse of a factor \( 1+\lambda \), and
|
|
the Ridge estimator converges to zero when the hyperparameter goes to
|
|
infinity.
|
|
|
|
<p>
|
|
We will come back to more interpreations after we have gone through some of the statistical analysis part.
|
|
|
|
<p>
|
|
For more discussions of Ridge and Lasso regression, <a href="https://arxiv.org/abs/1509.09169" target="_blank">Wessel van Wieringen's</a> article is highly recommended.
|
|
Similarly, <a href="https://arxiv.org/abs/1803.08823" target="_blank">Mehta et al's article</a> is also recommended.
|
|
|
|
<p>
|
|
<!-- !split -->
|
|
|
|
<h2 id="a-better-understanding-of-regularization">A better understanding of regularization </h2>
|
|
|
|
<p>
|
|
The parameter \( \lambda \) that we have introduced in the Ridge (and
|
|
Lasso as well) regression is often called a regularization parameter
|
|
or shrinkage parameter. It is common to call it a hyperparameter. What does it mean mathemtically?
|
|
|
|
<p>
|
|
Here we will first look at how to analyze the difference between the
|
|
standard OLS equations and the Ridge expressions in terms of a linear
|
|
algebra analysis using the SVD algorithm. Thereafter, we will link
|
|
(see the material on the bias-variance tradeoff below) these
|
|
observation to the statisical analysis of the results. In particular
|
|
we consider how the variance of the parameters \( \boldsymbol{\beta} \) is
|
|
affected by changing the parameter \( \lambda \).
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="decomposing-the-ols-and-ridge-expressions">Decomposing the OLS and Ridge expressions </h2>
|
|
|
|
<p>
|
|
We have our design matrix
|
|
\( \boldsymbol{X}\in {\mathbb{R}}^{n\times p} \). With the SVD we decompose it as
|
|
|
|
$$
|
|
\boldsymbol{X} = \boldsymbol{U\Sigma V^T},
|
|
$$
|
|
|
|
<p>
|
|
with \( \boldsymbol{U}\in {\mathbb{R}}^{n\times n} \), \( \boldsymbol{\Sigma}\in {\mathbb{R}}^{n\times p} \)
|
|
and \( \boldsymbol{V}\in {\mathbb{R}}^{p\times p} \).
|
|
|
|
<p>
|
|
The matrices \( \boldsymbol{U} \) and \( \boldsymbol{V} \) are unitary/orthonormal matrices, that is in case the matrices are real we have \( \boldsymbol{U}^T\boldsymbol{U}=\boldsymbol{U}\boldsymbol{U}^T=\boldsymbol{I} \) and \( \boldsymbol{V}^T\boldsymbol{V}=\boldsymbol{V}\boldsymbol{V}^T=\boldsymbol{I} \).
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="introducing-the-covariance-and-correlation-functions">Introducing the Covariance and Correlation functions </h2>
|
|
|
|
<p>
|
|
Before we discuss the link between for example Ridge regression and the singular value decomposition, we need to remind ourselves about
|
|
the definition of the covariance and the correlation function. These are quantities
|
|
|
|
<p>
|
|
Suppose we have defined two vectors
|
|
\( \hat{x} \) and \( \hat{y} \) with \( n \) elements each. The covariance matrix \( \boldsymbol{C} \) is defined as
|
|
$$
|
|
\boldsymbol{C}[\boldsymbol{x},\boldsymbol{y}] = \begin{bmatrix} \mathrm{cov}[\boldsymbol{x},\boldsymbol{x}] & \mathrm{cov}[\boldsymbol{x},\boldsymbol{y}] \\
|
|
\mathrm{cov}[\boldsymbol{y},\boldsymbol{x}] & \mathrm{cov}[\boldsymbol{y},\boldsymbol{y}] \\
|
|
\end{bmatrix},
|
|
$$
|
|
|
|
where for example
|
|
$$
|
|
\mathrm{cov}[\boldsymbol{x},\boldsymbol{y}] =\frac{1}{n} \sum_{i=0}^{n-1}(x_i- \overline{x})(y_i- \overline{y}).
|
|
$$
|
|
|
|
With this definition and recalling that the variance is defined as
|
|
$$
|
|
\mathrm{var}[\boldsymbol{x}]=\frac{1}{n} \sum_{i=0}^{n-1}(x_i- \overline{x})^2,
|
|
$$
|
|
|
|
we can rewrite the covariance matrix as
|
|
$$
|
|
\boldsymbol{C}[\boldsymbol{x},\boldsymbol{y}] = \begin{bmatrix} \mathrm{var}[\boldsymbol{x}] & \mathrm{cov}[\boldsymbol{x},\boldsymbol{y}] \\
|
|
\mathrm{cov}[\boldsymbol{x},\boldsymbol{y}] & \mathrm{var}[\boldsymbol{y}] \\
|
|
\end{bmatrix}.
|
|
$$
|
|
|
|
<p>
|
|
The covariance takes values between zero and infinity and may thus
|
|
lead to problems with loss of numerical precision for particularly
|
|
large values. It is common to scale the covariance matrix by
|
|
introducing instead the correlation matrix defined via the so-called
|
|
correlation function
|
|
|
|
$$
|
|
\mathrm{corr}[\boldsymbol{x},\boldsymbol{y}]=\frac{\mathrm{cov}[\boldsymbol{x},\boldsymbol{y}]}{\sqrt{\mathrm{var}[\boldsymbol{x}] \mathrm{var}[\boldsymbol{y}]}}.
|
|
$$
|
|
|
|
<p>
|
|
The correlation function is then given by values \( \mathrm{corr}[\boldsymbol{x},\boldsymbol{y}]
|
|
\in [-1,1] \). This avoids eventual problems with too large values. We
|
|
can then define the correlation matrix for the two vectors \( \boldsymbol{x} \)
|
|
and \( \boldsymbol{y} \) as
|
|
|
|
$$
|
|
\boldsymbol{K}[\boldsymbol{x},\boldsymbol{y}] = \begin{bmatrix} 1 & \mathrm{corr}[\boldsymbol{x},\boldsymbol{y}] \\
|
|
\mathrm{corr}[\boldsymbol{y},\boldsymbol{x}] & 1 \\
|
|
\end{bmatrix},
|
|
$$
|
|
|
|
<p>
|
|
In the above example this is the function we constructed using <b>pandas</b>.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="correlation-function-and-design-feature-matrix">Correlation Function and Design/Feature Matrix </h2>
|
|
|
|
<p>
|
|
In our derivation of the various regression algorithms like <b>Ordinary Least Squares</b> or <b>Ridge regression</b>
|
|
we defined the design/feature matrix \( \boldsymbol{X} \) as
|
|
|
|
$$
|
|
\boldsymbol{X}=\begin{bmatrix}
|
|
x_{0,0} & x_{0,1} & x_{0,2}& \dots & \dots x_{0,p-1}\\
|
|
x_{1,0} & x_{1,1} & x_{1,2}& \dots & \dots x_{1,p-1}\\
|
|
x_{2,0} & x_{2,1} & x_{2,2}& \dots & \dots x_{2,p-1}\\
|
|
\dots & \dots & \dots & \dots \dots & \dots \\
|
|
x_{n-2,0} & x_{n-2,1} & x_{n-2,2}& \dots & \dots x_{n-2,p-1}\\
|
|
x_{n-1,0} & x_{n-1,1} & x_{n-1,2}& \dots & \dots x_{n-1,p-1}\\
|
|
\end{bmatrix},
|
|
$$
|
|
|
|
with \( \boldsymbol{X}\in {\mathbb{R}}^{n\times p} \), with the predictors/features \( p \) refering to the column numbers and the
|
|
entries \( n \) being the row elements.
|
|
We can rewrite the design/feature matrix in terms of its column vectors as
|
|
$$
|
|
\boldsymbol{X}=\begin{bmatrix} \boldsymbol{x}_0 & \boldsymbol{x}_1 & \boldsymbol{x}_2 & \dots & \dots & \boldsymbol{x}_{p-1}\end{bmatrix},
|
|
$$
|
|
|
|
with a given vector
|
|
$$
|
|
\boldsymbol{x}_i^T = \begin{bmatrix}x_{0,i} & x_{1,i} & x_{2,i}& \dots & \dots x_{n-1,i}\end{bmatrix}.
|
|
$$
|
|
|
|
<p>
|
|
With these definitions, we can now rewrite our \( 2\times 2 \)
|
|
correaltion/covariance matrix in terms of a moe general design/feature
|
|
matrix \( \boldsymbol{X}\in {\mathbb{R}}^{n\times p} \). This leads to a \( p\times p \)
|
|
covariance matrix for the vectors \( \boldsymbol{x}_i \) with \( i=0,1,\dots,p-1 \)
|
|
|
|
$$
|
|
\boldsymbol{C}[\boldsymbol{x}] = \begin{bmatrix}
|
|
\mathrm{var}[\boldsymbol{x}_0] & \mathrm{cov}[\boldsymbol{x}_0,\boldsymbol{x}_1] & \mathrm{cov}[\boldsymbol{x}_0,\boldsymbol{x}_2] & \dots & \dots & \mathrm{cov}[\boldsymbol{x}_0,\boldsymbol{x}_{p-1}]\\
|
|
\mathrm{cov}[\boldsymbol{x}_1,\boldsymbol{x}_0] & \mathrm{var}[\boldsymbol{x}_1] & \mathrm{cov}[\boldsymbol{x}_1,\boldsymbol{x}_2] & \dots & \dots & \mathrm{cov}[\boldsymbol{x}_1,\boldsymbol{x}_{p-1}]\\
|
|
\mathrm{cov}[\boldsymbol{x}_2,\boldsymbol{x}_0] & \mathrm{cov}[\boldsymbol{x}_2,\boldsymbol{x}_1] & \mathrm{var}[\boldsymbol{x}_2] & \dots & \dots & \mathrm{cov}[\boldsymbol{x}_2,\boldsymbol{x}_{p-1}]\\
|
|
\dots & \dots & \dots & \dots & \dots & \dots \\
|
|
\dots & \dots & \dots & \dots & \dots & \dots \\
|
|
\mathrm{cov}[\boldsymbol{x}_{p-1},\boldsymbol{x}_0] & \mathrm{cov}[\boldsymbol{x}_{p-1},\boldsymbol{x}_1] & \mathrm{cov}[\boldsymbol{x}_{p-1},\boldsymbol{x}_{2}] & \dots & \dots & \mathrm{var}[\boldsymbol{x}_{p-1}]\\
|
|
\end{bmatrix},
|
|
$$
|
|
|
|
and the correlation matrix
|
|
$$
|
|
\boldsymbol{K}[\boldsymbol{x}] = \begin{bmatrix}
|
|
1 & \mathrm{corr}[\boldsymbol{x}_0,\boldsymbol{x}_1] & \mathrm{corr}[\boldsymbol{x}_0,\boldsymbol{x}_2] & \dots & \dots & \mathrm{corr}[\boldsymbol{x}_0,\boldsymbol{x}_{p-1}]\\
|
|
\mathrm{corr}[\boldsymbol{x}_1,\boldsymbol{x}_0] & 1 & \mathrm{corr}[\boldsymbol{x}_1,\boldsymbol{x}_2] & \dots & \dots & \mathrm{corr}[\boldsymbol{x}_1,\boldsymbol{x}_{p-1}]\\
|
|
\mathrm{corr}[\boldsymbol{x}_2,\boldsymbol{x}_0] & \mathrm{corr}[\boldsymbol{x}_2,\boldsymbol{x}_1] & 1 & \dots & \dots & \mathrm{corr}[\boldsymbol{x}_2,\boldsymbol{x}_{p-1}]\\
|
|
\dots & \dots & \dots & \dots & \dots & \dots \\
|
|
\dots & \dots & \dots & \dots & \dots & \dots \\
|
|
\mathrm{corr}[\boldsymbol{x}_{p-1},\boldsymbol{x}_0] & \mathrm{corr}[\boldsymbol{x}_{p-1},\boldsymbol{x}_1] & \mathrm{corr}[\boldsymbol{x}_{p-1},\boldsymbol{x}_{2}] & \dots & \dots & 1\\
|
|
\end{bmatrix},
|
|
$$
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="covariance-matrix-examples">Covariance Matrix Examples </h2>
|
|
|
|
<p>
|
|
The Numpy function <b>np.cov</b> calculates the covariance elements using
|
|
the factor \( 1/(n-1) \) instead of \( 1/n \) since it assumes we do not have
|
|
the exact mean values. The following simple function uses the
|
|
<b>np.vstack</b> function which takes each vector of dimension \( 1\times n \)
|
|
and produces a \( 2\times n \) matrix \( \boldsymbol{W} \)
|
|
|
|
$$
|
|
\boldsymbol{W} = \begin{bmatrix} x_0 & y_0 \\
|
|
x_1 & y_1 \\
|
|
x_2 & y_2\\
|
|
\dots & \dots \\
|
|
x_{n-2} & y_{n-2}\\
|
|
x_{n-1} & y_{n-1} &
|
|
\end{bmatrix},
|
|
$$
|
|
|
|
<p>
|
|
which in turn is converted into into the \( 2\times 2 \) covariance matrix
|
|
\( \boldsymbol{C} \) via the Numpy function <b>np.cov()</b>. We note that we can also calculate
|
|
the mean value of each set of samples \( \boldsymbol{x} \) etc using the Numpy
|
|
function <b>np.mean(x)</b>. We can also extract the eigenvalues of the
|
|
covariance matrix through the <b>np.linalg.eig()</b> function.
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># Importing various packages</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
n = <span style="color: #B452CD">100</span>
|
|
x = np.random.normal(size=n)
|
|
<span style="color: #658b00">print</span>(np.mean(x))
|
|
y = <span style="color: #B452CD">4</span>+<span style="color: #B452CD">3</span>*x+np.random.normal(size=n)
|
|
<span style="color: #658b00">print</span>(np.mean(y))
|
|
W = np.vstack((x, y))
|
|
C = np.cov(W)
|
|
<span style="color: #658b00">print</span>(C)
|
|
</pre></div>
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="correlation-matrix">Correlation Matrix </h2>
|
|
|
|
<p>
|
|
The previous example can be converted into the correlation matrix by
|
|
simply scaling the matrix elements with the variances. We should also
|
|
subtract the mean values for each column. This leads to the following
|
|
code which sets up the correlations matrix for the previous example in
|
|
a more brute force way. Here we scale the mean values for each column of the design matrix, calculate the relevant mean values and variances and then finally set up the \( 2\times 2 \) correlation matrix (since we have only two vectors).
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
n = <span style="color: #B452CD">100</span>
|
|
<span style="color: #228B22"># define two vectors </span>
|
|
x = np.random.random(size=n)
|
|
y = <span style="color: #B452CD">4</span>+<span style="color: #B452CD">3</span>*x+np.random.normal(size=n)
|
|
<span style="color: #228B22">#scaling the x and y vectors </span>
|
|
x = x - np.mean(x)
|
|
y = y - np.mean(y)
|
|
variance_x = np.sum(x<span style="color: #707a7c">@x</span>)/n
|
|
variance_y = np.sum(y<span style="color: #707a7c">@y</span>)/n
|
|
<span style="color: #658b00">print</span>(variance_x)
|
|
<span style="color: #658b00">print</span>(variance_y)
|
|
cov_xy = np.sum(x<span style="color: #707a7c">@y</span>)/n
|
|
cov_xx = np.sum(x<span style="color: #707a7c">@x</span>)/n
|
|
cov_yy = np.sum(y<span style="color: #707a7c">@y</span>)/n
|
|
C = np.zeros((<span style="color: #B452CD">2</span>,<span style="color: #B452CD">2</span>))
|
|
C[<span style="color: #B452CD">0</span>,<span style="color: #B452CD">0</span>]= cov_xx/variance_x
|
|
C[<span style="color: #B452CD">1</span>,<span style="color: #B452CD">1</span>]= cov_yy/variance_y
|
|
C[<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>]= cov_xy/np.sqrt(variance_y*variance_x)
|
|
C[<span style="color: #B452CD">1</span>,<span style="color: #B452CD">0</span>]= C[<span style="color: #B452CD">0</span>,<span style="color: #B452CD">1</span>]
|
|
<span style="color: #658b00">print</span>(C)
|
|
</pre></div>
|
|
<p>
|
|
We see that the matrix elements along the diagonal are one as they
|
|
should be and that the matrix is symmetric. Furthermore, diagonalizing
|
|
this matrix we easily see that it is a positive definite matrix.
|
|
|
|
<p>
|
|
The above procedure with <b>numpy</b> can be made more compact if we use <b>pandas</b>.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="correlation-matrix-with-pandas">Correlation Matrix with Pandas </h2>
|
|
|
|
<p>
|
|
We whow here how we can set up the correlation matrix using <b>pandas</b>, as done in this simple code
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
|
|
n = <span style="color: #B452CD">10</span>
|
|
x = np.random.normal(size=n)
|
|
x = x - np.mean(x)
|
|
y = <span style="color: #B452CD">4</span>+<span style="color: #B452CD">3</span>*x+np.random.normal(size=n)
|
|
y = y - np.mean(y)
|
|
X = (np.vstack((x, y))).T
|
|
<span style="color: #658b00">print</span>(X)
|
|
Xpd = pd.DataFrame(X)
|
|
<span style="color: #658b00">print</span>(Xpd)
|
|
correlation_matrix = Xpd.corr()
|
|
<span style="color: #658b00">print</span>(correlation_matrix)
|
|
</pre></div>
|
|
<p>
|
|
We expand this model to the Franke function discussed above.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="correlation-matrix-with-pandas-and-the-franke-function">Correlation Matrix with Pandas and the Franke function </h2>
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># Common imports</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
|
|
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">FrankeFunction</span>(x,y):
|
|
term1 = <span style="color: #B452CD">0.75</span>*np.exp(-(<span style="color: #B452CD">0.25</span>*(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>) - <span style="color: #B452CD">0.25</span>*((<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>))
|
|
term2 = <span style="color: #B452CD">0.75</span>*np.exp(-((<span style="color: #B452CD">9</span>*x+<span style="color: #B452CD">1</span>)**<span style="color: #B452CD">2</span>)/<span style="color: #B452CD">49.0</span> - <span style="color: #B452CD">0.1</span>*(<span style="color: #B452CD">9</span>*y+<span style="color: #B452CD">1</span>))
|
|
term3 = <span style="color: #B452CD">0.5</span>*np.exp(-(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">7</span>)**<span style="color: #B452CD">2</span>/<span style="color: #B452CD">4.0</span> - <span style="color: #B452CD">0.25</span>*((<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">3</span>)**<span style="color: #B452CD">2</span>))
|
|
term4 = -<span style="color: #B452CD">0.2</span>*np.exp(-(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">4</span>)**<span style="color: #B452CD">2</span> - (<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">7</span>)**<span style="color: #B452CD">2</span>)
|
|
<span style="color: #8B008B; font-weight: bold">return</span> term1 + term2 + term3 + term4
|
|
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">create_X</span>(x, y, n ):
|
|
<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #658b00">len</span>(x.shape) > <span style="color: #B452CD">1</span>:
|
|
x = np.ravel(x)
|
|
y = np.ravel(y)
|
|
|
|
N = <span style="color: #658b00">len</span>(x)
|
|
l = <span style="color: #658b00">int</span>((n+<span style="color: #B452CD">1</span>)*(n+<span style="color: #B452CD">2</span>)/<span style="color: #B452CD">2</span>) <span style="color: #228B22"># Number of elements in beta</span>
|
|
X = np.ones((N,l))
|
|
|
|
<span style="color: #8B008B; font-weight: bold">for</span> i <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(<span style="color: #B452CD">1</span>,n+<span style="color: #B452CD">1</span>):
|
|
q = <span style="color: #658b00">int</span>((i)*(i+<span style="color: #B452CD">1</span>)/<span style="color: #B452CD">2</span>)
|
|
<span style="color: #8B008B; font-weight: bold">for</span> k <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(i+<span style="color: #B452CD">1</span>):
|
|
X[:,q+k] = (x**(i-k))*(y**k)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">return</span> X
|
|
|
|
|
|
<span style="color: #228B22"># Making meshgrid of datapoints and compute Franke's function</span>
|
|
n = <span style="color: #B452CD">4</span>
|
|
N = <span style="color: #B452CD">100</span>
|
|
x = np.sort(np.random.uniform(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, N))
|
|
y = np.sort(np.random.uniform(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, N))
|
|
z = FrankeFunction(x, y)
|
|
X = create_X(x, y, n=n)
|
|
|
|
Xpd = pd.DataFrame(X)
|
|
<span style="color: #228B22"># subtract the mean values and set up the covariance matrix</span>
|
|
Xpd = Xpd - Xpd.mean()
|
|
covariance_matrix = Xpd.cov()
|
|
<span style="color: #658b00">print</span>(covariance_matrix)
|
|
</pre></div>
|
|
<p>
|
|
We note here that the covariance is zero for the first rows and
|
|
columns since all matrix elements in the design matrix were set to one
|
|
(we are fitting the function in terms of a polynomial of degree \( n \)).
|
|
|
|
<p>
|
|
This means that the variance for these elements will be zero and will
|
|
cause problems when we set up the correlation matrix. We can simply
|
|
drop these elements and construct a correlation
|
|
matrix without these elements.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="rewriting-the-covariance-and-or-correlation-matrix">Rewriting the Covariance and/or Correlation Matrix </h2>
|
|
|
|
<p>
|
|
We can rewrite the covariance matrix in a more compact form in terms of the design/feature matrix \( \boldsymbol{X} \) as
|
|
$$
|
|
\boldsymbol{C}[\boldsymbol{x}] = \frac{1}{n}\boldsymbol{X}^T\boldsymbol{X}= \mathbb{E}[\boldsymbol{X}^T\boldsymbol{X}].
|
|
$$
|
|
|
|
<p>
|
|
To see this let us simply look at a design matrix \( \boldsymbol{X}\in {\mathbb{R}}^{2\times 2} \)
|
|
$$
|
|
\boldsymbol{X}=\begin{bmatrix}
|
|
x_{00} & x_{01}\\
|
|
x_{10} & x_{11}\\
|
|
\end{bmatrix}=\begin{bmatrix}
|
|
\boldsymbol{x}_{0} & \boldsymbol{x}_{1}\\
|
|
\end{bmatrix}.
|
|
$$
|
|
|
|
<p>
|
|
If we then compute the expectation value
|
|
$$
|
|
\mathbb{E}[\boldsymbol{X}^T\boldsymbol{X}] = \frac{1}{n}\boldsymbol{X}^T\boldsymbol{X}=\begin{bmatrix}
|
|
x_{00}^2+x_{01}^2 & x_{00}x_{10}+x_{01}x_{11}\\
|
|
x_{10}x_{00}+x_{11}x_{01} & x_{10}^2+x_{11}^2\\
|
|
\end{bmatrix},
|
|
$$
|
|
|
|
which is just
|
|
$$
|
|
\boldsymbol{C}[\boldsymbol{x}_0,\boldsymbol{x}_1] = \boldsymbol{C}[\boldsymbol{x}]=\begin{bmatrix} \mathrm{var}[\boldsymbol{x}_0] & \mathrm{cov}[\boldsymbol{x}_0,\boldsymbol{x}_1] \\
|
|
\mathrm{cov}[\boldsymbol{x}_1,\boldsymbol{x}_0] & \mathrm{var}[\boldsymbol{x}_1] \\
|
|
\end{bmatrix},
|
|
$$
|
|
|
|
where we wrote $$\boldsymbol{C}[\boldsymbol{x}_0,\boldsymbol{x}_1] = \boldsymbol{C}[\boldsymbol{x}]$$ to indicate that this the covariance of the vectors \( \boldsymbol{x} \) of the design/feature matrix \( \boldsymbol{X} \).
|
|
|
|
<p>
|
|
It is easy to generalize this to a matrix \( \boldsymbol{X}\in {\mathbb{R}}^{n\times p} \).
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="linking-with-svd">Linking with SVD </h2>
|
|
|
|
<p>
|
|
More material will be added here.
|
|
|
|
<p>
|
|
<!-- !split --><br><br><br><br><br><br><br><br><br><br>
|
|
|
|
<h2 id="exercises-for-week-37-september-6-10">Exercises for week 37, September 6-10 </h2>
|
|
|
|
<p>
|
|
The exercises here are meant to prepare you for work with project 1. The first exercise is a follow-up of exercise 2 from week 36 August 30-September 3).
|
|
|
|
<p>
|
|
<!-- --- begin exercise --- -->
|
|
|
|
<h2 id="exercise-1-adding-ridge-and-lasso-regression">Exercise 1: Adding Ridge and Lasso Regression </h2>
|
|
|
|
<p>
|
|
This exercise is a continuation of exercise 2 from exercise set 1 (week 36, August 30-September 3)). We will
|
|
use the same function to generate our data set, still staying with a
|
|
simple function \( y(x) \) which we want to fit using linear regression,
|
|
but now extending the analysis to include the Ridge and the Lasso
|
|
regression methods.
|
|
|
|
<p>
|
|
We will thus again generate our own dataset for a function \( y(x) \) where
|
|
\( x \in [0,1] \) and defined by random numbers computed with the uniform
|
|
distribution. The function \( y \) is a quadratic polynomial in \( x \) with
|
|
added stochastic noise according to the normal distribution \( \cal{N}(0,1) \).
|
|
|
|
<p>
|
|
The following simple Python instructions define our \( x \) and \( y \) values (with 100 data points).
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>x = np.random.rand(<span style="color: #B452CD">100</span>)
|
|
y = <span style="color: #B452CD">2.0</span>+<span style="color: #B452CD">5</span>*x*x+<span style="color: #B452CD">0.1</span>*np.random.randn(<span style="color: #B452CD">100</span>)
|
|
</pre></div>
|
|
<p>
|
|
Write your own code for the Ridge method (see chapter 3.4 of Hastie <em>et al.</em>, equations (3.43) and (3.44)) and compute the parametrization for different values of \( \lambda \). Compare and analyze your results with those from exercise 3. Study the dependence on \( \lambda \) while also varying the strength of the noise in your expression for \( y(x) \).
|
|
|
|
<p>
|
|
The code here allows you to perform your own Ridge calculation and perform calculations for various values of the regularization parameter \( \lambda \). This program can easily be extended upon.
|
|
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">os</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">pandas</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">pd</span>
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.preprocessing</span> <span style="color: #8B008B; font-weight: bold">import</span> StandardScaler
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">R2</span>(y_data, y_model):
|
|
<span style="color: #8B008B; font-weight: bold">return</span> <span style="color: #B452CD">1</span> - np.sum((y_data - y_model) ** <span style="color: #B452CD">2</span>) / np.sum((y_data - np.mean(y_data)) ** <span style="color: #B452CD">2</span>)
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">MSE</span>(y_data,y_model):
|
|
n = np.size(y_model)
|
|
<span style="color: #8B008B; font-weight: bold">return</span> np.sum((y_data-y_model)**<span style="color: #B452CD">2</span>)/n
|
|
|
|
|
|
<span style="color: #228B22"># A seed just to ensure that the random numbers are the same for every run.</span>
|
|
<span style="color: #228B22"># Useful for eventual debugging.</span>
|
|
np.random.seed(<span style="color: #B452CD">3155</span>)
|
|
|
|
x = np.random.rand(<span style="color: #B452CD">100</span>)
|
|
y = <span style="color: #B452CD">2.0</span>+<span style="color: #B452CD">5</span>*x*x+<span style="color: #B452CD">0.1</span>*np.random.randn(<span style="color: #B452CD">100</span>)
|
|
|
|
<span style="color: #228B22"># number of features p (here degree of polynomial</span>
|
|
p = <span style="color: #B452CD">3</span>
|
|
<span style="color: #228B22"># The design matrix now as function of a given polynomial</span>
|
|
X = np.zeros((<span style="color: #658b00">len</span>(x),p))
|
|
X[:,<span style="color: #B452CD">0</span>] = <span style="color: #B452CD">1.0</span>
|
|
X[:,<span style="color: #B452CD">1</span>] = x
|
|
X[:,<span style="color: #B452CD">2</span>] = x*x
|
|
<span style="color: #228B22"># We split the data in test and training data</span>
|
|
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=<span style="color: #B452CD">0.2</span>)
|
|
scaler = StandardScaler()
|
|
scaler.fit(X_train)
|
|
X_train_scaled = scaler.transform(X_train)
|
|
X_test_scaled = scaler.transform(X_test)
|
|
|
|
<span style="color: #228B22"># matrix inversion to find beta</span>
|
|
OLSbeta = np.linalg.inv(X_train.T @ X_train) @ X_train.T @ y_train
|
|
<span style="color: #658b00">print</span>(OLSbeta)
|
|
<span style="color: #228B22"># and then make the prediction</span>
|
|
ytildeOLS = X_train @ OLSbeta
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Training R2 for OLS"</span>)
|
|
<span style="color: #658b00">print</span>(R2(y_train,ytildeOLS))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Training MSE for OLS"</span>)
|
|
<span style="color: #658b00">print</span>(MSE(y_train,ytildeOLS))
|
|
ypredictOLS = X_test @ OLSbeta
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Test R2 for OLS"</span>)
|
|
<span style="color: #658b00">print</span>(R2(y_test,ypredictOLS))
|
|
<span style="color: #658b00">print</span>(<span style="color: #CD5555">"Test MSE OLS"</span>)
|
|
<span style="color: #658b00">print</span>(MSE(y_test,ypredictOLS))
|
|
|
|
<span style="color: #228B22"># Repeat now for Ridge regression and various values of the regularization parameter</span>
|
|
I = np.eye(p,p)
|
|
<span style="color: #228B22"># Decide which values of lambda to use</span>
|
|
nlambdas = <span style="color: #B452CD">20</span>
|
|
MSEPredict = np.zeros(nlambdas)
|
|
MSETrain = np.zeros(nlambdas)
|
|
lambdas = np.logspace(-<span style="color: #B452CD">4</span>, <span style="color: #B452CD">1</span>, nlambdas)
|
|
<span style="color: #8B008B; font-weight: bold">for</span> i <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(nlambdas):
|
|
lmb = lambdas[i]
|
|
Ridgebeta = np.linalg.inv(X_train.T @ X_train+lmb*I) @ X_train.T @ y_train
|
|
<span style="color: #228B22"># and then make the prediction</span>
|
|
ytildeRidge = X_train @ Ridgebeta
|
|
ypredictRidge = X_test @ Ridgebeta
|
|
MSEPredict[i] = MSE(y_test,ypredictRidge)
|
|
MSETrain[i] = MSE(y_train,ytildeRidge)
|
|
<span style="color: #228B22"># Now plot the results</span>
|
|
plt.figure()
|
|
plt.plot(np.log10(lambdas), MSETrain, label = <span style="color: #CD5555">'MSE Ridge train'</span>)
|
|
plt.plot(np.log10(lambdas), MSEPredict, <span style="color: #CD5555">'r--'</span>, label = <span style="color: #CD5555">'MSE Ridge Test'</span>)
|
|
plt.xlabel(<span style="color: #CD5555">'log10(lambda)'</span>)
|
|
plt.ylabel(<span style="color: #CD5555">'MSE'</span>)
|
|
plt.legend()
|
|
plt.show()
|
|
</pre></div>
|
|
<p>
|
|
Repeat the above but using the functionality of <b>Scikit-Learn</b>. Compare your code with the results from <b>Scikit-Learn</b>. Remember to run with the same random numbers for generating \( x \) and \( y \).
|
|
|
|
<p>
|
|
Finally, using <b>Scikit-Learn</b> or your own code, compute also the mean square error, a risk metric corresponding to the expected value of the squared (quadratic) error defined as
|
|
$$ MSE(\hat{y},\hat{\tilde{y}}) = \frac{1}{n}
|
|
\sum_{i=0}^{n-1}(y_i-\tilde{y}_i)^2,
|
|
$$
|
|
|
|
and the \( R^2 \) score function.
|
|
If \( \tilde{\hat{y}}_i \) is the predicted value of the \( i-th \) sample and \( y_i \) is the corresponding true value, then the score \( R^2 \) is defined as
|
|
$$
|
|
R^2(\hat{y}, \tilde{\hat{y}}) = 1 - \frac{\sum_{i=0}^{n - 1} (y_i - \tilde{y}_i)^2}{\sum_{i=0}^{n - 1} (y_i - \bar{y})^2},
|
|
$$
|
|
|
|
where we have defined the mean value of \( \hat{y} \) as
|
|
$$
|
|
\bar{y} = \frac{1}{n} \sum_{i=0}^{n - 1} y_i.
|
|
$$
|
|
|
|
Discuss these quantities as functions of the variable \( \lambda \) in the Ridge and Lasso regression methods.
|
|
|
|
<h3 id="exercise-linear-regression-for-a-two-dimensional-function">Exercise: Linear Regression for a two-dimensional function </h3>
|
|
|
|
<p>
|
|
This is a longer exercise and the aim is to study in more detail various
|
|
regression methods, including the Ordinary Least Squares (OLS) method,
|
|
Ridge regression and finally Lasso regression.
|
|
This exercise forms a part of project 1.
|
|
|
|
<p>
|
|
We will study how to fit polynomials to a specific
|
|
two-dimensional function called <a href="http://www.dtic.mil/dtic/tr/fulltext/u2/a081688.pdf" target="_blank">Franke's
|
|
function</a>. This
|
|
is a function which has been widely used when testing various
|
|
interpolation and fitting algorithms.
|
|
|
|
<p>
|
|
The Franke function, which is a weighted sum of four exponentials reads as follows
|
|
$$
|
|
\begin{align*}
|
|
f(x,y) &= \frac{3}{4}\exp{\left(-\frac{(9x-2)^2}{4} - \frac{(9y-2)^2}{4}\right)}+\frac{3}{4}\exp{\left(-\frac{(9x+1)^2}{49}- \frac{(9y+1)}{10}\right)} \\
|
|
&+\frac{1}{2}\exp{\left(-\frac{(9x-7)^2}{4} - \frac{(9y-3)^2}{4}\right)} -\frac{1}{5}\exp{\left(-(9x-4)^2 - (9y-7)^2\right) }.
|
|
\end{align*}
|
|
$$
|
|
|
|
<p>
|
|
The function will be defined for \( x,y\in [0,1] \). Our first step will
|
|
be to perform an OLS regression analysis of this function, trying out
|
|
a polynomial fit with an \( x \) and \( y \) dependence of the form \( [x, y,
|
|
x^2, y^2, xy, \dots] \). We will fit a
|
|
function (for example a polynomial) of \( x \) and \( y \). Thereafter we
|
|
will repeat much of the same procedure using the Ridge and Lasso
|
|
regression methods, introducing thus a dependence on the bias
|
|
(penalty) \( \lambda \).
|
|
|
|
<p>
|
|
The Python fucntion for the Franke function is included here (it performs also a three-dimensional plot of it)
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">mpl_toolkits.mplot3d</span> <span style="color: #8B008B; font-weight: bold">import</span> Axes3D
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">matplotlib</span> <span style="color: #8B008B; font-weight: bold">import</span> cm
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">matplotlib.ticker</span> <span style="color: #8B008B; font-weight: bold">import</span> LinearLocator, FormatStrFormatter
|
|
<span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
|
|
<span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">random</span> <span style="color: #8B008B; font-weight: bold">import</span> random, seed
|
|
|
|
fig = plt.figure()
|
|
ax = fig.gca(projection=<span style="color: #CD5555">'3d'</span>)
|
|
|
|
<span style="color: #228B22"># Make data.</span>
|
|
x = np.arange(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">0.05</span>)
|
|
y = np.arange(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">0.05</span>)
|
|
x, y = np.meshgrid(x,y)
|
|
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">FrankeFunction</span>(x,y):
|
|
term1 = <span style="color: #B452CD">0.75</span>*np.exp(-(<span style="color: #B452CD">0.25</span>*(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>) - <span style="color: #B452CD">0.25</span>*((<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>))
|
|
term2 = <span style="color: #B452CD">0.75</span>*np.exp(-((<span style="color: #B452CD">9</span>*x+<span style="color: #B452CD">1</span>)**<span style="color: #B452CD">2</span>)/<span style="color: #B452CD">49.0</span> - <span style="color: #B452CD">0.1</span>*(<span style="color: #B452CD">9</span>*y+<span style="color: #B452CD">1</span>))
|
|
term3 = <span style="color: #B452CD">0.5</span>*np.exp(-(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">7</span>)**<span style="color: #B452CD">2</span>/<span style="color: #B452CD">4.0</span> - <span style="color: #B452CD">0.25</span>*((<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">3</span>)**<span style="color: #B452CD">2</span>))
|
|
term4 = -<span style="color: #B452CD">0.2</span>*np.exp(-(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">4</span>)**<span style="color: #B452CD">2</span> - (<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">7</span>)**<span style="color: #B452CD">2</span>)
|
|
<span style="color: #8B008B; font-weight: bold">return</span> term1 + term2 + term3 + term4
|
|
|
|
|
|
z = FrankeFunction(x, y)
|
|
|
|
<span style="color: #228B22"># Plot the surface.</span>
|
|
surf = ax.plot_surface(x, y, z, cmap=cm.coolwarm,
|
|
linewidth=<span style="color: #B452CD">0</span>, antialiased=<span style="color: #8B008B; font-weight: bold">False</span>)
|
|
|
|
<span style="color: #228B22"># Customize the z axis.</span>
|
|
ax.set_zlim(-<span style="color: #B452CD">0.10</span>, <span style="color: #B452CD">1.40</span>)
|
|
ax.zaxis.set_major_locator(LinearLocator(<span style="color: #B452CD">10</span>))
|
|
ax.zaxis.set_major_formatter(FormatStrFormatter(<span style="color: #CD5555">'%.02f'</span>))
|
|
|
|
<span style="color: #228B22"># Add a color bar which maps values to colors.</span>
|
|
fig.colorbar(surf, shrink=<span style="color: #B452CD">0.5</span>, aspect=<span style="color: #B452CD">5</span>)
|
|
|
|
plt.show()
|
|
</pre></div>
|
|
<p>
|
|
We will generate our own dataset for a function
|
|
\( \mathrm{FrankeFunction}(x,y) \) with \( x,y \in [0,1] \). The function
|
|
\( f(x,y) \) is the Franke function. You should explore also the addition
|
|
an added stochastic noise to this function using the normal
|
|
distribution \( \cal{N}(0,1) \).
|
|
|
|
<p>
|
|
Write your own code (using either a matrix inversion or a singular
|
|
value decomposition from e.g., <b>numpy</b> ) or use your code and perform a standard least square regression
|
|
analysis using polynomials in \( x \) and \( y \) up to fifth order. You can use <b>scikit-learn</b> as well.
|
|
|
|
<p>
|
|
Evaluate the Mean Squared error (MSE)
|
|
|
|
$$ MSE(\hat{y},\hat{\tilde{y}}) = \frac{1}{n}
|
|
\sum_{i=0}^{n-1}(y_i-\tilde{y}_i)^2,
|
|
$$
|
|
|
|
<p>
|
|
and the \( R^2 \) score function. If \( \tilde{\hat{y}}_i \) is the predicted
|
|
value of the \( i-th \) sample and \( y_i \) is the corresponding true value,
|
|
then the score \( R^2 \) is defined as
|
|
|
|
$$
|
|
R^2(\hat{y}, \tilde{\hat{y}}) = 1 - \frac{\sum_{i=0}^{n - 1} (y_i - \tilde{y}_i)^2}{\sum_{i=0}^{n - 1} (y_i - \bar{y})^2},
|
|
$$
|
|
|
|
<p>
|
|
where we have defined the mean value of \( \hat{y} \) as
|
|
|
|
$$
|
|
\bar{y} = \frac{1}{n} \sum_{i=0}^{n - 1} y_i.
|
|
$$
|
|
|
|
<p>
|
|
You should split your data in train and test and also consider scaling the data.
|
|
|
|
<p>
|
|
To set up the design matrix, the following code can be used
|
|
<p>
|
|
|
|
<!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
|
|
<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">FrankeFunction</span>(x,y):
|
|
term1 = <span style="color: #B452CD">0.75</span>*np.exp(-(<span style="color: #B452CD">0.25</span>*(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>) - <span style="color: #B452CD">0.25</span>*((<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">2</span>)**<span style="color: #B452CD">2</span>))
|
|
term2 = <span style="color: #B452CD">0.75</span>*np.exp(-((<span style="color: #B452CD">9</span>*x+<span style="color: #B452CD">1</span>)**<span style="color: #B452CD">2</span>)/<span style="color: #B452CD">49.0</span> - <span style="color: #B452CD">0.1</span>*(<span style="color: #B452CD">9</span>*y+<span style="color: #B452CD">1</span>))
|
|
term3 = <span style="color: #B452CD">0.5</span>*np.exp(-(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">7</span>)**<span style="color: #B452CD">2</span>/<span style="color: #B452CD">4.0</span> - <span style="color: #B452CD">0.25</span>*((<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">3</span>)**<span style="color: #B452CD">2</span>))
|
|
term4 = -<span style="color: #B452CD">0.2</span>*np.exp(-(<span style="color: #B452CD">9</span>*x-<span style="color: #B452CD">4</span>)**<span style="color: #B452CD">2</span> - (<span style="color: #B452CD">9</span>*y-<span style="color: #B452CD">7</span>)**<span style="color: #B452CD">2</span>)
|
|
<span style="color: #8B008B; font-weight: bold">return</span> term1 + term2 + term3 + term4
|
|
|
|
|
|
<span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">create_X</span>(x, y, n ):
|
|
<span style="color: #8B008B; font-weight: bold">if</span> <span style="color: #658b00">len</span>(x.shape) > <span style="color: #B452CD">1</span>:
|
|
x = np.ravel(x)
|
|
y = np.ravel(y)
|
|
|
|
N = <span style="color: #658b00">len</span>(x)
|
|
l = <span style="color: #658b00">int</span>((n+<span style="color: #B452CD">1</span>)*(n+<span style="color: #B452CD">2</span>)/<span style="color: #B452CD">2</span>) <span style="color: #228B22"># Number of elements in beta</span>
|
|
X = np.ones((N,l))
|
|
|
|
<span style="color: #8B008B; font-weight: bold">for</span> i <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(<span style="color: #B452CD">1</span>,n+<span style="color: #B452CD">1</span>):
|
|
q = <span style="color: #658b00">int</span>((i)*(i+<span style="color: #B452CD">1</span>)/<span style="color: #B452CD">2</span>)
|
|
<span style="color: #8B008B; font-weight: bold">for</span> k <span style="color: #8B008B">in</span> <span style="color: #658b00">range</span>(i+<span style="color: #B452CD">1</span>):
|
|
X[:,q+k] = (x**(i-k))*(y**k)
|
|
|
|
<span style="color: #8B008B; font-weight: bold">return</span> X
|
|
|
|
|
|
<span style="color: #228B22"># Making meshgrid of datapoints and compute Franke's function</span>
|
|
n = <span style="color: #B452CD">5</span>
|
|
N = <span style="color: #B452CD">1000</span>
|
|
x = np.sort(np.random.uniform(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, N))
|
|
y = np.sort(np.random.uniform(<span style="color: #B452CD">0</span>, <span style="color: #B452CD">1</span>, N))
|
|
z = FrankeFunction(x, y)
|
|
X = create_X(x, y, n=n)
|
|
</pre></div>
|
|
<p>
|
|
Write then your own code for the Ridge method or use <b>Scikit-Learn</b>.
|
|
Perform the same analysis as you did for ordinary Least Squares (for the same polynomials) but now for different values of \( \lambda \). Compare and
|
|
analyze your results with those obtained with ordinary Least Squares. Study the
|
|
dependence on \( \lambda \).
|
|
|
|
<p>
|
|
This part is essentially a repeat of the previous ones, but now
|
|
with Lasso regression. Write either your own code or
|
|
use the functionalities of <b>Scikit-Learn</b> (recommended).
|
|
Give a
|
|
critical discussion of the three methods and a judgement of which
|
|
model fits the data best.
|
|
|
|
<p>
|
|
<!-- --- end exercise --- -->
|
|
|
|
<!-- ------------------- end of main content --------------- -->
|
|
|
|
|
|
<center style="font-size:80%">
|
|
<!-- copyright --> © 1999-2021, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
|
|
</center>
|
|
|
|
|
|
</body>
|
|
</html>
|
|
|
|
|