Analysis of Empirical Likelihood Confidence Regions in Data Loss Situation of Linear E V Random Model

摘要

By using the complementary approach of linear regression, this paper has complement the missing data of linear EV model in the case of missing completely at random to response variable, has constructed empirical likelihood ratio statistics of unknown parameters, and has proved the empirical likelihood ratio statistics which had constructed above approximate chi square distribution, finally, according to the results, this article also has constructed confidence regions of unknown parameters.