[Feature Proposal]: Generalized Additive Models

[osheari1]

I'm working on a project that requires GAMs. I'm more than willing to implement this myself, as long as it fits in with the Smile road map.

One other caveat is, I would only be able to implement it in Scala (rather than Java). Would this be an issue?

[haifengl]

Scala is fine. Please implements it in the scala module and corresponding package, e.g. smile.regression.

[osheari1]

👍

[osheari1]

Would it be OK to add an additional dependency for vectorization vs iterating over double arrays? If not I can just deal with the looping.

[haifengl]

You would like to add a library dependency for vector processing? In general, we hide this details in math module (mostly in Matrix interface). With different plugin, we can have pure java implementation or native one.

[ValdarT]

Hi. Any updates on this?

[osheari1]

Hey,

Unfortunately I haven't had any time to work on this in the past couple months due to moving / new job ect. After starting the project, it turned out it would've required way more work than I initially expected for 1 person. If we could get more people to contribute, finishing it would become a lot more viable.

[ValdarT]

Why close the issue? GAMs seem like a great addition so the feature request makes sense (whether there is ongoing work on it or not)?