Closed Issam28 closed 6 years ago
Fortunately, these are not dramatic issues.
The first one is due to a mistake in the authors' example code. 'auto' is not a valid parameter for class_weight. You have to replace it by a valid one, that is balanced or balanced_subsample. (the error message you get is explicit on this)
E.g:
rf = RandomForestClassifier(n_jobs=-1, class_weight='auto', max_depth=5)
see my pull request.
As per your second point. It is a slight coding error in the definition of the _get_imp() function. See my other pull request.
Once you are done with these changes, it shall work ok.
sorry, scikit learn changed the RF api.. this used to be 'auto' for sure. Thanks a lot for the fix @tagomatech
i can't reproduce your example with my dataset . here's the error that i'm getting
any idea how to solve this ?