Abstract
Abstract Bayesian elastic net and classical elastic net are regularization methods that provide variable selection procedure. We discuss the Bayesian elastic net by setting the scale mixture of normal distribution mixing with Rayleigh distribution as double exponential (Laplace) prior distribution of regression coefficient. The new proposed scale mixture produced normal distribution mixing with truncated gamma distribution. The hierarchical prior distributions and new Gibbs sample algorithm have developed. Therefore, variable selection have discussed through some simulation examples. The simulation results show the outperformance of the proposed model.
