Question on PMCAR with multiple likelihoods

[BasMatsuura]

Dear Professor Gómez-Rubio,

I am trying to implement your multivariate effects, particularly the PMCAR, for multiple likelihoods using INLA. Your book helped me figure out how to structure univariate data for multiple likelihoods, but I fail to understand how to do so for multivariate data. My aim is to then fit the PMCAR effect to it, using the copy() function to retain the hyperparameters across the likelihoods.

Have you any idea whether this would be possible? And if so, any suggestions on how to structure the data when passing it to INLA and the inla.PMCAR effect? Right now I am stuck performing trial and error to see if it runs with different configurations. Was hoping you had already given this a try!

Yours sincerely,

Bas