Empirical likelihood (EL) is a nonparametric analog of standard likelihood, and inherits many of its properties.  Accordingly, Bayesian EL, which replaces the ordinary likelihood function with an empirical likelihood function, has been developed.  In this talk, I will briefly introduce empirical likelihood in both its standard and Bayesian versions.  A recent development is Bayesian EL for regularized methods such as ridge regression.  The theoretical background for Bayesian ridge and Bayesian lasso empirical likelihoods will be presented, along with discussion of computational issues.   

This is joint work with Adel Bedoui.