Empirical Likelihood Based Bias-Correction for Linear EV Model with Missing Data

摘要

The empirical likelihood inference based bias correction in linear EV model with missing responses is studied. In the literature, the usual method to deal with missing data is imputation by which the missing value of response is to be filled in with appropriate values to create a completed data set. An empirical likelihood based bias-correction method is developed. It can be shown that the directly biascorrected empirical likelihood ratio is asymptotically standard chi-square. A simulation study indicates that the proposed method performs competitively in terms of the average lengths and coverage probabilities of confidence intervals.