BAS: Bayesian Model Averaging using Bayesian Adaptive Sampling
Package for Bayesian Model Averaging in linear models and generalized linear models using stochastic or deterministic sampling without replacement from posterior distributions. Prior distributions on coefficients are from Zellner's g-prior or mixtures of g-priors corresponding to the Zellner-Siow Cauchy Priors or the mixture of g-priors from Liang et al (2008) <http://dx.doi.org/10.1198/016214507000001337> for linear models or mixtures of g-priors in GLMs of Li and Clyde (2015) <http://arxiv.org/abs/1503.06913>. Other model selection criteria include AIC, BIC and Empirical Bayes estimates of g. Sampling probabilities may be updated based on the sampled models using Sampling w/out Replacement or an efficient MCMC algorithm samples models using the BAS tree structure as an efficient hash table. Uniform priors over all models or beta-binomial prior distributions on model size are allowed, and for large p truncated priors on the model space may be used. The user may force variables to always be included. Details behind the sampling algorithm are provided in Clyde, Ghosh and Littman (2010) <http://dx.doi.org/10.1198/jcgs.2010.09049>.