Generalized additive models in julia?


#1

Hi,

Is anyone using generalized additive models in julia?

I know I can link to pyGAM with PyCall or gam with RCall, but I find both packages don’t scale well. My current dataset is of the order 100million rows.

I’m willing to collaborate if somewhen has some code lying around or if someone could point me to some - preferably permissive - code to port to julia (pyGAM and R’s gam are both GPL).

Thanks!
Cheers
Ken


#2

Did you find any help for this project? It would be indeed very useful to have a top GAM fitting package in Julia…


#3

I talked to someone at JuliaCon who does this.


#4

Well, David, that’s a bit vague… How did he or she look like? :smile:

Best,
Ken


#5

You could take one of the splines packages to create basis functions, add them to a design matrix, and fit with MixedModels, with random effects acting as a smoothing penalty, or even using MCMC. Look at Tutorial in biostatistics: spline smoothing with linear mixed models..