Adaptive Monte Carlo for Bayesian variable selection in regression models
Griffin, Jim E.
Steel, Mark F.J.
MetadataShow full item record
This article describesmethods for efficient posterior simulation for Bayesian variable selection in generalized linear models with many regressors but few observations. The algorithms use a proposal on model space that contains a tuneable parameter. An adaptive approach to choosing this tuning parameter is described that allows automatic, efficient computation in these models. The method is applied to examples from normal linear and probit regression. Relevant code and datasets are posted online as supplementary materials.