Convergence analysis of a collapsed Gibbs sampler for Bayesian vector autoregressions (Q2044318)

This paper presents remarkable results about convergence complexity analysis (study of how convergence properties of MCMC algorithms are affected by changes in the data), an emerging research area with practical and theoretical importance. The practical importance stems from the fact that MCMC algorithms frequently are used in the Bayesian inference context. The theoretical importance stems from the wide variety of situations which the paper results apply, serving as a foundation for further research in the convergence complexity analysis area.