search

interpretml / interpret

Fit interpretable models. Explain blackbox machine learning.
https://interpret.ml/docs
MIT License
6.04k stars 715 forks source link

Method for extracting where the steps are in the variable plots #506

Open AntonyJScott opened 3 months ago

AntonyJScott commented 3 months ago

I am looking to extract the data underlying feature importance plots and the stepped line plots for individual features in order to make custom piots.

Is there a way for determining at what variable values the steps in the stepped line plots occur please?

paulbkoch commented 3 months ago

Hi @AntonyJScott --

The feature importances can get obtained with the term_importances function on the EBM objects, which is documented here.

For drawing or otherwise extracting the partial response plots, here are a two issues that might be of use to you: https://github.com/interpretml/interpret/issues/325 and https://github.com/interpretml/interpret/issues/475

And you might also find our documentation on EBM internals useful: https://interpret.ml/docs/ebm-internals-regression.html