InterpretML

Hi @hexgnu, Good idea! We reset the index right now for engineering simplicity, but we can look into preserving this in the future. Nice to hear that this would be...

Hi @kjk111, Thanks for reporting this. Do you mind sharing your Python and interpret version numbers with us? I'm having trouble reproducing this on my side. You can get your...

Hi @aman63, Good question! You can access these through the `.data()` method on any global explanation. For example: ```python3 ebm_global = ebm.explain_global() graph_data = ebm_global.data(0) # Gets graph related info...

Hi @1230pitchanqw , Thanks for bringing this up. We currently log the validation performance of every 10th epoch in our logging system. You can enable this by running the following...

Hi @andrewjylee, Thanks for bringing this up and looking into it so carefully! The main issue is that we estimate and include the interaction terms for EBMs during the fit...

Hi @ZhangTP1996, It's possible to extend EBMs for ranking use cases, and building this feature is currently on our backlog. However, we don't expect to get to it in the...

Seems like it could be a useful feature. We'll look into showing some real time metrics with a verbose flag, maybe as an experimental feature to see how it works....

Hi @ruedika, Thanks for the insightful question, and for taking the time to provide code! Unfortunately, no GAM fitting method can universally approximate all possible generator functions, and EBMs are...

Hi @brianfhead -- We're happy to hear that EBMs are performing comparably on your datasets. You are correct that the current R implementation only supports global explanations. The R package...

Hi @bverhoeff, Thanks for your numerous thoughtful questions, and for being a fan of the package! Funnily enough, we used to scale each plot independently in the earliest releases of...