Quantile regression with Explainable Boosting Machine

[gmartinonQM]

Hi all !

I am really interested in using Explainable Boosting Machine for quantile regression. Is the pinball loss implemented ? Do you have a running example available ?

Thanks in advance for your time.

[paulbkoch]

Hi @gmartinonQM -- Custom losses are not implemented currently. It's something that we've put a lot of work into, but the feature is not complete and we're currently working on other issues for a bit before returning to it, so it'll be a few months at least before these become available. We're tracking custom losses in #281

-InterpretML team

[paulbkoch]

I'll close this issue for now since we're tracking custom losses in issue #281