My data df_X has 11 features and y is multi-class label. I used multi-class SVM to select importance of features. estimator_.coef_ should return the score of each feature. But why here it returns a list of scores? The same case occured for multi-class LogisticRegression().

By the way, what's the difference between SelectKBest and SelectFromModel for feature selection in sklearn.

enter image description here

  • $\begingroup$ Do you have 11 classes? $\endgroup$
    – Sycorax
    Commented Jan 5, 2023 at 15:32
  • $\begingroup$ @Sycorax no, the classes in samples are 3,4,5,6,7,8. $\endgroup$ Commented Jan 6, 2023 at 1:52


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.