This paper improves the existing literature on the shrinkage of high dimensional model and parameter spaces through Bayesian priors and Markov Chains algorithms. A hierarchical semiparametric Bayes approach is developed to overtake limits and misspecificity involved in compressed regression...