资 源 简 介
PIMSA is an implementation of a reversible jump MCMC algorithm that samples across the space of all possible models in the context of variable selection. Output from the sampler can be used for inference of a posterior distribution. Key features of this program are the ability to sample higher-order interactions (the current version is hard-coded at a limit of 5), and incorporate a variety of prior sources of knowledge to inform the variable selection kernel.
