Fast Monte Carlo Markov chains for Bayesian shrinkage models with random effects (Q1755112)

The paper under review deals with performing Bayesian data analysis using a general linear mixed model. Since the (conditionally conjugate) multivariate Gaussian prior on \(\beta\) coefficients does not perform well in the high-dimensional setting, the authors consider a model in which this prior is replaced by the normal-gamma shrinkage prior developed by \textit{J. E. Griffin} and \textit{P. J. Brown} [Bayesian Anal. 5, No. 1, 171--188 (2010; Zbl 1330.62128)]. The resulting posterior density is complex, and the authors develop a MCMC algorithm for exploring it. This algorithm is easier to analyze than the more obvious three-step Gibbs sampler. And it is proved that the algorithm is geometrically ergodic in many practical settings.