The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here

features selected in training phase

mar_granimar_grani Member Posts: 1 Learner I
edited December 2018 in Help

Hi!

I have a question: I tried to train a Logistic model with AutoModel, but I don't understand how to see the features used in the classification (I suppose that not all the features were selected as usefull and some features were deleted in the training phase).

Thanks a lot!

Answers

  • MartinLiebigMartinLiebig Administrator, Moderator, Employee-RapidMiner, RapidMiner Certified Analyst, RapidMiner Certified Expert, University Professor Posts: 3,533 RM Data Scientist

    Hi @mar_grani,

     

    i think the easiest way is to just add a break point before your learner.

     

    Cheers,

    Martin

    - Sr. Director Data Solutions, Altair RapidMiner -
    Dortmund, Germany
  • IngoRMIngoRM Employee-RapidMiner, RapidMiner Certified Analyst, RapidMiner Certified Expert, Community Manager, RMResearcher, Member, University Professor Posts: 1,751 RM Founder

    Martin is right, you could indeed open the process generated by Auto Model and place a breakpoint before the learner and execute the process.  You will then see exactly what the learner has used.  But this would actually be the same data which is shown in Auto Model under "General" - "Data".  The global importance of the features (independent of the model) is shown in "General" - "Weights" by the way.  Most models perform some form of feature selection in addition to how the data was prepared.  You can see the used features in each model's visualization (if there was any selection, otherwise all features have been used).

     

    The next version of Auto Model will come by the way with more advanced ways of automatic feature selection and feature generation to further improve this.

     

    Hope this helps,

    Ingo

Sign In or Register to comment.