CODEWITHSUNDEEP
pythonpandasscikit-learn
Fluxy
Fluxy

Reputation: 2978

How to find out the feature names selected by RFECV?

How can I know which particular features are selected with RFECV? If I do X_rfecv_train=selector.transform(X_train), I get a numpy array, but I don't know feature names...

rf_rec = RandomForestClassifier(n_jobs=-1, max_depth = 20, max_features = 0.9, 
                                min_samples_leaf = 2, min_samples_split = 0.1, 
                                n_estimators = 100, oob_score=True, class_weight="balanced", 
                                random_state=0)
selector = RFECV(rf_rec, step=1, cv=5)
selector = selector.fit(X_train, y_train)

X_rfecv_train=selector.transform(X_train)
X_rfecv_test=selector.transform(X_test)

Upvotes: 1

Views: 1526

Answers (1)

Vivek Kumar
Vivek Kumar

Reputation: 36599

You can use the get_support(). All selectors in scikit-learn should have this method.

print(selector.get_support(indices=True))

This will print the index of your original features which are selected and present in X_rfecv_train and X_rfecv_test.

Upvotes: 3

Related Questions