Add absorbing regression for GLM
aeturrell opened this issue · comments
Apologies if this is totally out of scope.
As far as I can tell, there's no mature, cross-OS, maintained Python package for absorbing regression for generalized linear models. There's a prototype version using jax + GPUs in fastreg but I haven't seen anything else, and it's not available in statsmodels. Definitely of secondary importance, but it would be great to have it available via a formula API too.
It would be fab to have this feature!
Hi @aeturrell, GLM is out of scope for their project. I think the natural home would be statsmodels, given the extensive GLM capabilities there.