R package and FM extensions

[adrtod]

Hello!
Thanks for this great open source code!

Is there any R interface available for your library?
If not I might be interested in developing it.

Have you considered extending to Field-aware FM and Higher-order FM?
If not, how difficult would it be to adapt the current implementation?

[ibayer]

@adrtod
I'm currently completely rewriting and extending the solvers in c++ (currently C)#94. The new version will
be much easier to extend. The rewrite will include 3'rd FMs with the original formulation (not HOFMs). Its currently to early to say if I will also add Field-aware and HOFMs, partly because I haven't see strong evidence that they actually perform better. However, I intent to include an CD solver for implicit feedback[0].

We don't have a R interface yet but I'm very motivated to change that. One of the goals for the c++ implementation is to simply interfacing from R and Python. I would be very interested in a collaboration!

[0] A Generic Coordinate Descent Framework for Learning from Implicit Feedback

[adrtod]

Sounds good!
I started a skeleton of the package here:
https://github.com/adrtod/rfastFM