A Bayesian Semiparametric Random Effect Model for Meta-Regression

In this paper, we will introduce a Bayesian semiparametric model concerned with both constant and coefficients. In Meta-Analysis or Meta-Regression, we usually use a parametric family. However, lately the increasing tendency to use Bayesian nonparametric and semiparametric models, entered this area too. On the other hand, although we have some works on Bayesian nonparametric or semiparametric models, they just focus on intercept and do not pay much attention to regressor coefficient(s). We also would check the efficiency of the proposed model via simulation and give an illustrating example.

A Bayesian Semiparametric Random Effect Model for Meta-Regression Keywords:

Meta-analysis , Meta-regression , Dirichlet process , Bayesian Model Selection , Gibbs Sampling