具有时变协变量的纵向资料的线性混合效应模型及算法  

The Linear Mixed Effects Models for Time-Varying Covariates in Longitudinal Data

作　　者：王晓东[1] 田俊[2] 

机构地区：[1]泉州师范学院数学与计算机科学学院,福建泉州362000 [2]福建医科大学公共卫生学院,福建福州350004

出　　处：《数学的实践与认识》2012年第12期55-63,共9页Mathematics in Practice and Theory

基　　金：国家自然科学基金(81041066)

摘　　要：许多研究领域中都会涉及到纵向资料的分析处理.在纵向资料的分析中,常常遇到带有时变协变量的情况.传统的方差分析难以处理具有时变协变量的纵向资料.对线性混合效应模型应用于分析这类资料进行了方法学的探讨,并编制了模型拟合的SAS程序.通过实例应用,对线性混合效应模型分析纵向资料的方法和过程给出详细的介绍.The treatment and analysis of longitudinal data may be encountered in many research areas. In longitudinal data analysis, often encountered with a situation of time- varying covariates. It is difficult to deal with the longitudinal data with time-varying covariates by using a traditional variance analysis method. In this paper, we explore methodologically that how the linear mixed effects model can be applied to analyze such data. We have also designed the SAS programs for the linear mixed effects model fitting. A detailed description and process of the linear mixed model and its application to the analysis of longitudinal data are given by example applications.

关 键 词：重复观测 协变量 随机效应 线性模型 

分 类 号：O212.1[理学—概率论与数理统计]