Why only GBT?

[sodaliAyran]

I understand why you use tree based algorithms to find Zero Importance Features. But why Gradient Boosted Trees is the only option? I think RandomForest and maybe DecisionTree algorithms should be one of the options. Or are they already?

[WillKoehrsen]

Random forests would also work for model-based feature importances. This is an option I'm looking to add when I have some time. Thanks for the suggestion.

[sodaliAyran]

@WillKoehrsen thank you.