Estimating Bayesian Hierarchical Models using bayesGDS

Braun and Damien (2015), henceforth known as BD, introduce an alternative to MCMC for sampling from posterior distributions. The main advantages of BD over MCMC are that samples can be collected in parallel, and that the algorithm is scalable (linear complexity) for hierarchical models with conditionally independent heterogeneous units. These features make BD an attractive estimation method for Bayesian hierarchical models of large datasets.