Problem for accounting for spatial correlation with NNGP model

[avehtari]

@BenCretois

As I said I am trying to limit my use of packages type brms or INLA to get a better grasp of what I am doing though

@BenCretois

I fitted the model in brms with the gp() and although this is very slow compared to the NNGP this seem to work.

brms has the Practical Hilbert space approximate Bayesian Gaussian processes for probabilistic programming implemented. Just add option k to gp(). From gp: Set up Gaussian process terms in 'brms' in brms: Bayesian Regression Models using 'Stan'

• k Optional number of basis functions for computing approximate GPs. If NA (the default), exact GPs are computed.