An Improved Approach for Estimating Empirical likelihood Based on Random Walk Metropolis Algorithm

摘要

It has been proved that empirical likelihood can be a likelihood function for Bayesian analysis.Based on this discovery,we propose a random walk metropolis algorithm to estimate the maximum empirical likelihood.First,a simulation dataset is generated which take a uniform distribution as prior,and then obtain a posterior.Finally,we prove this distribution is the normal one and the max value is similar to empirical likelihood estimator by sequential quadratic programming.