wwrechard/weierstrass: Combining posterior samples from multiple subsets.

The Weierstrass sampler is a "divide-conquer-combine" type parallel MCMC sampler. The implementation in this package is the particular rejection sampler which directly works on the subset posterior samples as a post-processing sampler. The method make use of rejection sampling or importance sampling to combine the subset posterior samples for approximating the posterior obtained on the full data set.

install.packages("remotes")
remotes::install_github("wwrechard/weierstrass")