Improvement Description
add optimize_feature_selection option to classify-samples-ncv by adding recursive feature elimination to the inner loop of the nested CV training
Current Behavior
no correct way of performing feature elimination in classify-samples-ncv (see Q2 Forum post)
Proposed Behavior
allow for feature elimination procedure as in classify-samples with parameter optimize_feature_selection
Comments
- Suggestion to enhance outputted average feature importance table by adding an additional column named "number of times feature was selected" to account for features being dropped/used in different folds.
References
- answer on StackExchange
- answer on StackOverflow
Improvement Description
add
optimize_feature_selectionoption toclassify-samples-ncvby adding recursive feature elimination to the inner loop of the nested CV trainingCurrent Behavior
no correct way of performing feature elimination in classify-samples-ncv (see Q2 Forum post)
Proposed Behavior
allow for feature elimination procedure as in classify-samples with parameter
optimize_feature_selectionComments
References