Problems of measure_importance() from "randomForestExplainer" Package

I want to get the mean minimal depths for all my variables and I used function measure_importance(rfm). When I used randomforest() to get my model, everything is good, but the output dataframe has a problem when I used train(). It calculated the mean_min_depth of each category for a categorical variable. I am curious if there is any difference of final model between train() and randomforest(). Because I want to use train() rather than randomforest(). How can I get the mean_min_depth for each variable not for each category?

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.