Add Spectral Mixture Kernel

It would be a good idea, I think, to add in the spectral mixture (SM) kernel introduced by Andrew Wilson and Ryan Adams in 2014. There’s been a decent literature based on this work (over 60 citations in 2 years) including some work I’m starting on and it’d be helpful to not have to hand roll GP optimization every time I use this kernel.

I looked around the list of currently implemented kernels and couldn’t find it but if I just missed it let me know.

I guess the broader question here is how close you want this package to follow the literature. sklearn, for example, has a policy of not closely following the literature, so that’s not necessarily a bad thing, but it’d be helpful to have a guideline on what the package’s goals are in this domain.

Let me know what you guys think. Thanks 😀 !

(I’d be willing to help out in adding this though I’m not as of yet familiar with the internals of the package)

[imsrgadich]

for people interested I have SM kernel in GPflow.

https://github.com/imsrgadich/gprsm

[imsrgadich]

@cdipaolo were you able to find the issues with the initializations?