search

konstantint / SKompiler

A tool for compiling trained SKLearn models into other representations (such as SQL, Sympy or Excel formulas)
MIT License
171 stars 10 forks source link

Support for Explainable Boosting Machine #13

Open KalinNonchev opened 1 year ago

KalinNonchev commented 1 year ago

Hello,

thank you for the package. Would you also offer support for Explainable Boosting Machine from InterpretML? It is similar to Gradient Boosting Machine of sklearn.

Thank you in advance!

Best,

konstantint commented 1 year ago

I don't believe I would personally find time to implement that very soon. I am not familiar with the EBM model currently but looking into implementation I can't immediately tell how complex it is to translate it into a SK expression (from a quick glance it seems the logic is more complicated than what one would expect from a straightforward boosting model).

paulbkoch commented 1 year ago

Our docs have a simplified replica of the predict function that might be useful.

https://interpret.ml/docs/ebm-internals-regression.html