摘要
刻画纵向数据的协方差结构有三个可能因素:随机效应、序列相关和随机误差.在纵向数据分析中,模型方差的齐性是一个基本假定.但是,该假设未必正确. Zhang和Weiss研究了具有随机效应的线性模型的异方差检验.林金官和韦博成将Zhang和Weiss的结果推广到非线性情形.本文对具有自相关误差的非线性纵向数据模型,研究了方差齐性和相关系数的齐性检验,得到了检验的score统计量并应用于血浆渗透数据(见Davidian和Giltian.最后,本文还给出了模拟结果.
There are three possible factors to model covariance structure of longitudinal data: randon. effects, serial correlation and measurement errors. In longitudinal data analysis, homogeneity of variance is a basic assumption. However, this assumption is not necessarily appropriate. Zhang, Weiss, Lin, Wei respectively tested for heteroscedasticity in linear and nonlinear models with random effects based on longitudinal data. This paper discusses the tests for homogeneity of within-individual variances and between-individual autocorrelation coefficients in nonlinear models with AR(1) errors based on longitudinal data and obtains several score test statistics. The plasma concentrations data (Davidian, Giltian) is used to illustrate our results.
出处
《应用数学学报》
CSCD
北大核心
2004年第3期466-480,共15页
Acta Mathematicae Applicatae Sinica
基金
国家自然科学基金(10371016号)
国家社会科学基金(04BTJ002号)
东南大学博士后基金资助项目.
关键词
纵向数据
非线性模型
异方差
自相关系数
齐性检验
统计学
Longitudinal data, nonlinear regression, heteroscedasticity, autocorrelation coefficient, AR(1) error
