1: \begin{abstract}

2: 

3: Empirical likelihood is a popular nonparametric statistical tool that does not

4: require any distributional assumptions. 

5: In this paper, we explore the possibility of conducting variable selection via Bayesian empirical likelihood. 

6: We show theoretically that when the prior distribution

7: satisfies certain mild conditions, the corresponding Bayesian empirical

8: likelihood estimators are posteriorly consistent and variable selection consistent. 

9: As special cases, we show the prior of Bayesian empirical likelihood LASSO and SCAD satisfies such conditions 

10: and thus can identify the non-zero elements of the parameters with probability tending to $1$. 

11: In addition, it is easy to verify that those conditions are met for other widely used priors such as ridge, elastic net and adaptive LASSO.

12: Empirical likelihood depends on a parameter that needs to be obtained by numerically solving a non-linear equation. 

13: Thus, there exists no conjugate prior for the posterior distribution, 

14: which causes the slow convergence of the MCMC sampling algorithm in some cases. 

15: To solve this problem, we propose a novel approach, which uses an approximation distribution as the proposal. 

16: The computational results demonstrate quick convergence for the examples used in the paper. 

17: We use both simulation and real data analyses to illustrate the advantages of the proposed methods.

18: \end{abstract}

19: