文档介绍：删失情形下纵向数据变系数分位回归模型的
统计推断
杜江,谢田法
北京工业大学应用数理学院,北京 100124
摘要:本文讨论纵向数据变系数分位回归模型的响应变量固定删失时的统计推断。其中参数估
计时考虑了对非参变系数部分用样条逼近。给出了线性协变量某些系数是否为零的检验方法。
所提方法具有容易实现且对模型误差分布不加限定等优点。为了说明推断方法的有限样本性
质,给出了大量的模拟结果。同时,把所提方法用于艾滋病临床数据的实例分析。
关键词:纵向数据;分位回归;固定删失;样条函数;秩得分检验
中图分类号: O212
Statistical inference for censored partially linear
varying coefficient quantile models with
longitudinal data
DU Jiang, XIE Tian-Fa
College of Applied Sciences, Beijing University of Technology, Beijing 100124
Abstract: In this paper, an inference procedure is proposed for partially linear varying
coefficient quantile models with longitudinal data where some measurement of the response
are censored by fixed constants. The nonparametric functional coefficients are approximated
by splines. A hypothesis testing procedure is proposed to test whether some linear covariates
are noninformative. The proposed procedures can be easily implemented, and no specification
of the error distribution is needed. Finite sample performance of the proposed procedures is
assessed by Monte Carlo simulation studies. The proposed methodology is illustrated by an
analysis of AIDS clinical trial dataset.
Key words: Longitudinal data, quantile regression, fixed censoring, spline function, rank
score test.
基金项目: XIE’s research is supported by the National Natural Science Foundation of China(11101015) and specialized
research fund for the doctoral program of higher education (20091103120012)
作者简介: DU Jiang(1984-),male,PhD Student,major research direction:quantile regression. Correspondence
author:XIE Tian-Fa(1976-),male,associate professor,major research direction:Data depth, Statistical applications.
-1-
0 Introduction
In many longitudinal studies, repeated measurements of the response variable are collected
at irregular and possibly subject-specific time points. Partially linear varying coefficient is one of
the tools to deal with the longitudinal data gained, which was studied by Huang, Wu and Zhou
(2002), Qu and Li