python - Regression function by weighted least squares using scikit-learn -
i want use scikit-learn
's decisiontreeregressor, need different criterion. there way make own decision function? rather mean squared error, need use regression tree based on weighted least squares.
what best modification be? i'm thinking need modify tree change way feature_importances_
calculated.
however, checked code of tree.py
, while mentions tree_.computer_feature_importances()
, can't sem find make edits.
Comments
Post a Comment