This paper proposes the use of Bayesian model averaging (BMA) as a tool to select the predictors' set for bridge models. BMA is a computationally feasible method that allows us to explore the model space even in the presence of a large set of candidate predictors. We test the performance of BMA...