Generalized Empirical Likelihood Inference in Semiparametric Regression Model for Longitudinal Data  

作　　者：Gao Rong LI Ping TIAN Liu Gen XUE 

机构地区：[1]College of Applied Sciences, Beijing University of Technology, Beijing 100022, P. R. China [2]School of Finance and Statistics, East China Normal University, Shanghai 200241, P. R. China [3]Department of Mathematics, Xuchang University, Xuchang 461000, P. R. China

出　　处：《Acta Mathematica Sinica,English Series》2008年第12期2029-2040,共12页数学学报（英文版）

基　　金：China Postdoctoral Science Foundation Funded Project (20080430633);Shanghai Postdoctoral Scientific Program (08R214121);the National Natural Science Foundation of China (10871013);the Research Fund for the Doctoral Program of Higher Education (20070005003);the Natural Science Foundation of Beijing (1072004);the Basic Research and Frontier Technology Foundation of He'nan (072300410090)

摘　　要：In this paper, we consider the semiparametric regression model for longitudinal data. Due to the correlation within groups, a generalized empirical log-likelihood ratio statistic for the unknown parameters in the model is suggested by introducing the working covariance matrix. It is proved that the proposed statistic is asymptotically standard chi-squared under some suitable conditions, and hence it can be used to construct the confidence regions of the parameters. A simulation study is conducted to compare the proposed method with the generalized least squares method in terms of coverage accuracy and average lengths of the confidence intervals.In this paper, we consider the semiparametric regression model for longitudinal data. Due to the correlation within groups, a generalized empirical log-likelihood ratio statistic for the unknown parameters in the model is suggested by introducing the working covariance matrix. It is proved that the proposed statistic is asymptotically standard chi-squared under some suitable conditions, and hence it can be used to construct the confidence regions of the parameters. A simulation study is conducted to compare the proposed method with the generalized least squares method in terms of coverage accuracy and average lengths of the confidence intervals.

关 键 词：longitudinal data semiparametric regression model empirical likelihood confidence region 

分 类 号：O212[理学—概率论与数理统计]