CS545 fall 2016

This is an old revision of the document!

Warning: Declaration of syntax_plugin_fontsize2::handle($match, $state, $pos, &$handler) should be compatible with DokuWiki_Syntax_Plugin::handle($match, $state, $pos, Doku_Handler $handler) in /s/bach/b/class/cs545/public_html/fall16/lib/plugins/fontsize2/syntax.php on line 19

Warning: Declaration of syntax_plugin_fontsize2::render($mode, &$renderer, $data) should be compatible with DokuWiki_Syntax_Plugin::render($format, Doku_Renderer $renderer, $data) in /s/bach/b/class/cs545/public_html/fall16/lib/plugins/fontsize2/syntax.php on line 19

Warning: Declaration of syntax_plugin_comment::handle($match, $state, $pos, &$handler) should be compatible with DokuWiki_Syntax_Plugin::handle($match, $state, $pos, Doku_Handler $handler) in /s/bach/b/class/cs545/public_html/fall16/lib/plugins/comment/syntax.php on line 30

Warning: Declaration of syntax_plugin_comment::render($mode, &$renderer, $data) should be compatible with DokuWiki_Syntax_Plugin::render($format, Doku_Renderer $renderer, $data) in /s/bach/b/class/cs545/public_html/fall16/lib/plugins/comment/syntax.php on line 30

Warning: Declaration of syntax_plugin_tablewidth::handle($match, $state, $pos, &$handler) should be compatible with DokuWiki_Syntax_Plugin::handle($match, $state, $pos, Doku_Handler $handler) in /s/bach/b/class/cs545/public_html/fall16/lib/plugins/tablewidth/syntax.php on line 57

Warning: Declaration of syntax_plugin_tablewidth::render($mode, &$renderer, $data) should be compatible with DokuWiki_Syntax_Plugin::render($format, Doku_Renderer $renderer, $data) in /s/bach/b/class/cs545/public_html/fall16/lib/plugins/tablewidth/syntax.php on line 57

Warning: Declaration of syntax_plugin_mathjax_protecttex::render($mode, &$renderer, $data) should be compatible with DokuWiki_Syntax_Plugin::render($format, Doku_Renderer $renderer, $data) in /s/bach/b/class/cs545/public_html/fall16/lib/plugins/mathjax/syntax/protecttex.php on line 157

model selection and cross validation in scikit-learn

First let's import some modules and read in some data:

In [1]: import numpy as np
 
In [2]: from sklearn import cross_validation
 
In [3]: from sklearn import svm
 
In [4]: from sklearn import metrics
 
In [5]: data=np.genfromtxt("../data/heart_scale.data", delimiter=",")
 
In [6]: X=data[:,1:]
 
In [7]: y=data[:,0]

The simplest form of model evaluation uses a validation/test set:

In [9]: X_train, X_test, y_train, y_test = cross_validation.train_test_split(X, y, test_size=0.4, random_state=0)
 
In [10]: classifier = svm.SVC(kernel='linear', C=1).fit(X_train, y_train)
 
In [11]: classifier.score(X_test, y_test)
Out[11]: 0.7592592592592593

Next, let'd perform cross-validation:

In [12]: scores = cross_validation.cross_val_score(classifier, X, y, cv=5, scoring='accuracy')
 
In [13]: 
 
In [13]: scores = cross_validation.cross_val_score(classifier, X, y, cv=5, scoring='roc_auc')
 
In [14]: # you can also obtain the predictions by cross-validation and then compute the accuracy:
 
In [15]: y_predict = cross_validation.cross_val_predict(classifier, X, y, cv=5)
 
In [16]: metrics.accuracy_score(y, y_predict)
Out[16]: 0.83703703703703702

CS545 fall 2016

User Tools

Site Tools

Sidebar

model selection and cross validation in scikit-learn

Page Tools