FeatureSelection.py 文件源码

def test():
    #??????
    np.random.seed(13)
    X=pd.DataFrame(np.random.randn(20,10))
    X.columns=['x%d'%i for i in range(10)]
    y=pd.Series(np.random.choice([0,1],20))

    #??sklearn?????????????
    clf_sklearn=feature_selection.SelectKBest(feature_selection.f_classif,k=4)
    clf=SklearnSelector(estimator=clf_sklearn)
    clf.fit(X,y)
    clf.transform(X)
    print(clf.feature_selected)

    clf_sklearn=SelectFromModel(LogisticRegression())
    clf=SklearnSelector(estimator=clf_sklearn)
    clf.fit(X,y)
    clf.transform(X)
    print(clf.feature_selected)

    #?????
    clf_selectkbest=feature_selection.SelectKBest(feature_selection.f_classif,k=4)
    clf_selectfrommodel=SelectFromModel(LogisticRegression())
    clf_baseselector=SklearnSelector(clf_selectkbest)
    clf=VotingSelector(selectors=[('clf_selectkbest',clf_selectkbest),
                                  ('clf_selectfrommodel',clf_selectfrommodel),
                                  ('clf_baseselector',clf_baseselector)],threshold=0.5)
    clf.fit(X,y)
    clf.transform(X)
    print(clf.feature_selected)
    print(clf.df_voting)
    print(clf.score)