Mosaic plot for bivariate gaussian process

Dear all,

I am using the package “brms” for fitting models with STAN. I noticed that the package also supports Gaussian Processes with the function “gp”, which I use a lot to model interactions between continuous predictors.

However, the built-in function for plotting mosaic plots of bivariate smooths (conditional_smooth) does not seem to work for Gaussian Processes, but only for penalized splines.

Do you have any idea on how I can produce mosaic plots (or, even better, surface plots) from outputs of a multivariate Gaussian process? Unfortunately, no mainstream package for mosaic/surface plots supports “brms” and I have no idea on how to proceed. I guess that shifting to rstan might be an option, but again I do not know how to work through outputs for plotting them.

To better understand, this is the example for bivariate GP in brms

dat <- mgcv::gamSim(1, n = 30, scale = 2)
fit3 <- brm(y ~ gp(x1, x2), dat, chains = 2)
summary(fit3)

Best,

Jacopo Cerri

1 Like