CS545 fall 2016

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.

=== Feature selection in scikit-learn ===

This demo uses the a file of yeast gene expression data that is available {{wiki:yeast2.data | here}}.

<file python feature_selection.py>

"""
=================================================
SVM with RFE feature selection
=================================================

How to perform feature selection in scikit-learn
"""

import numpy as np
from sklearn import svm, feature_selection, cross_validation
from sklearn.pipeline import Pipeline
from sklearn.pipeline import make_pipeline
from sklearn.feature_selection import RFE,RFECV
from sklearn.svm import SVC,LinearSVC

# let's read in the yeast data that you used in an earlier assignment:
data = np.genfromtxt("../data/yeast2.csv", delimiter = ",")
X = data[:,1:]
y = data[:,0]
# add some extra features as noise
X = np.hstack((X, np.random.randn(len(y), 250)))

# create an instance of RFE that uses an SVM to define weights
# for the features (any linear classifier will work):
classifier = LinearSVC()
selector = RFE(classifier, step=0.1, n_features_to_select=25)
# run feature selection:
selector = selector.fit(X, y)

# check which features got chosen:
print selector.support_
print selector.ranking_

# to actually perform feature selection:
Xt=selector.fit_transform(X,y)

# the wrong way to perform cross-validation:
cv = cross_validation.StratifiedKFold(y, 5, shuffle=True, random_state=0)
print (np.mean(cross_validation.cross_val_score(classifier, Xt, y, cv=cv)))

# now let's perform nested cross-validation:
classifier = LinearSVC()
selector = RFE(classifier, step=0.1,n_features_to_select=25)
rfe_svm = make_pipeline(selector, classifier)

print (np.mean(cross_validation.cross_val_score(rfe_svm, X, y, cv=cv)))

# feature selection using a univariate filter method:
from sklearn.feature_selection import SelectKBest, f_regression
filter_selector = SelectKBest(f_regression, k=25)
classifier = LinearSVC()
filter_svm = make_pipeline(filter_selector, classifier)

print (np.mean(cross_validation.cross_val_score(filter_svm, X, y, cv=cv)))

</file>
CS545 fall 2016

User Tools

Site Tools

Page Tools
