Bayesian adaptive Lasso Tobit regression

Abstract

In this paper, we introduce a new procedure for model selection in Tobit regression, we suggest the Bayesian adaptive Lasso Tobit regression (BALTR) for variable selection (VS) and coefficient estimation. We submitted a Bayesian hierarchical model and Gibbs sampler (GS) for our procedure. Our proposed procedure is clarified by means of simulations and a real data analysis. Results demonstrate our procedure performs well in comparison to further procedures.