I am looking for non-stationary kernels for a GP for spatial modelling that can be fitted with Stan. So far, I have a couple of really interesting papers like this, but the maths there are above my paygrade and I really cannot implement it myself.
I did find this nsCorr function from the package BayesNSGP, which I think is what I want, but I am struggling with implementing that in Stan.
I’d appreciate any help or pointers.