Variable selection via the composite likelihood method for multilevel longitudinal data with missing responses and covariates

被引：1

作者：

Li, Haocheng ^{[1
]}

Shu, Di ^{[2
]}

He, Wenqing ^{[3
]}

Yi, Grace Y. ^{[2
]}

机构：

[1] Univ Calgary, Dept Math & Stat, Calgary, AB, Canada

[2] Univ Waterloo, Dept Stat & Actuarial Sci, Waterloo, ON, Canada

[3] Western Univ, Dept Stat & Actuarial Sci, London, ON, Canada

来源：

COMPUTATIONAL STATISTICS & DATA ANALYSIS | 2019年 / 135卷

基金：

加拿大自然科学与工程研究理事会;

关键词：

Composite likelihood; Longitudinal data; Missing data; Missing not at random; Multilevel structure; Variable selection; MODEL SELECTION; BINARY DATA; DROP-OUT;

D O I：

10.1016/j.csda.2019.01.011

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

Longitudinal data with multilevel structures are commonly collected when following up subjects in clusters over a period of time. Missing values and variable selection issues are common for such data. Biased results may be produced if incompleteness of data is ignored in the analysis. On the other hand, incorporating a large number of irrelevant covariates into inferential procedures may lead to difficulty in computation and interpretation. A unified penalized composite likelihood framework is developed to handle data with missingness and variable selection issues. It is flexible to handle the situation where responses and covariates are missing not simultaneously under the assumption of missing not at random. The method is justified both rigorously with theoretical results and numerically with simulation studies. The method is also applied to the Waterloo Smoking Prevention Project data. (C) 2019 Elsevier B.V. All rights reserved.

引用

页码：25 / 34

页数：10

共 30 条

[1] Joint Variable Selection for Fixed and Random Effects in Linear Mixed-Effects Models [J].