Joint posterior inference for latent Gaussian models with R-INLA

Efficient Bayesian inference remains a computational challenge in hierarchical models. Simulation-based approaches such as Markov Chain Monte Carlo methods are still popular but have large cost. When dealing with the class of Latent Gaussian Models, INLA methodology embedded R-INLA software provides accurate by computing deterministic mixture representation to approximate joint posterior, from which marginals computed. The approach has beginning been targeting univariate posteriors. In this paper, we lay out development foundation tools for also providing approximations subsets latent field. These inherit copula structure and additionally provide corrections skewness. same idea is carried forward sampling representation, now can adjust