BayesVarSel (version 1.3)
Bayesian Variable selection in Linear Models
Description
Within the context of the linear regression model, this package provides tools for the analysis of the variable selection problem from a Bayesian perspective. The default implementation takes advantage of a closed-form expression for the posterior probabilities that the prior proposed in Bayarri, Berger, Forte and Garcia-Donato (2012) produces. Alternatively, other priors, like Zellner (1986) g-prior, Zellner-Siow (1980,1984) or Liang, Paulo, Molina, Clyde and Berger (2008) can be used. BayesVarSel allows the calculations to be performed either exactly (sequential or parallel computation) or heuristically, using a Gibbs sampling algorithm studied in Garcia-Donato and Martinez-Beneito (2013).