Optimal Model Averaging Estimation for Generalized Linear Models and Generalized Linear Mixed-Effects Models

被引：143

作者：

Zhang, Xinyu ^{[1
,2
]}

Yu, Dalei ^{[3
]}

Zou, Guohua ^{[1
,2
]}

Liang, Hua ^{[4
]}

机构：

[1] Chinese Acad Sci, Acad Math & Syst Sci, Beijing, Peoples R China

[2] Sch Math Sci, Beijing, Peoples R China

[3] Yunnan Univ Finance & Econ, Stat & Math Coll, Kunming, Peoples R China

[4] Geoge Washington Univ, Dept Stat, Washington, DC USA

来源：

JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION | 2016年 / 111卷 / 516期

基金：

中国国家自然科学基金;

关键词：

Asymptotic optimality; Kullback-Leibler loss; Misspecification; Penalized generalized weighted least squares (PGWLS); Prediction accuracy; FOCUSED INFORMATION CRITERION; CONDITIONAL INFERENCE; REGRESSION; SELECTION;

D O I：

10.1080/01621459.2015.1115762

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 ; 070103 ; 0714 ;

摘要：

Considering model averaging estimation in generalized linear models, we propose a weight choice criterion based on the Kullback-Leibler (KL) loss with a penalty term. This criterion is different from that for continuous observations in principle, but reduces to the Mallows criterion in the situation. We prove that the corresponding model averaging estimator is asymptotically optimal under certain assumptions. We further extend our concern to the generalized linear mixed-effects model framework and establish associated theory. Numerical experiments illustrate that the proposed method is promising.

引用

页码：1775 / 1790

页数：16