Random Forest

Question

I am currently working on Career Guidance prediction using Machine Learning.

The dataset has 38 features. For feature selection I tried using mutual_info_classif for getting the mutual information of my features and got the list of important features.

The Second approach I followed is using SelectKBest with mutual_info_classif as my score_func. On this approach I got some other list of features.

Is it normal to get different results ?
Can anyone please help me out?

Answer 1

Hello @Priyesh Dave

Thanks for reaching out to us here. Based on my understanding, if you keep X, Y the same, and you are using sklearn.feature_selection.SelectKBest(score_func= mutual_info_classif , *, k= n) the result should be the same.

Are you manually rank the feature with mutual_info_classif score? The SelectKBest function is basically return the K highest feature based on the X and Y.

If this is not your case, I would highly recommend you to check with Scikit-learn team by email : https://mail.python.org/mailman/listinfo/scikit-learn

Or forum: https://stackoverflow.com/questions/tagged/scikit-learn

Hope this will help. Please let us know if any further queries.

------------------------------

• Please don't forget to click on or upvote button whenever the information provided helps you. Original posters help the community find answers faster by identifying the correct answer. Here is how
• Want a reminder to come back and check responses? Here is how to subscribe to a notification
• If you are interested in joining the VM program and help shape the future of Q&A: Here is how you can be part of Q&A Volunteer Moderators