多类型复发事件下变系数加性乘积比率回归模型  

A Flexible Additive multiplicative Rates Model for Multiple Type Recurrent Event Data

作　　者：戴家佳[1,2] 刘焕彬[3] 杨振海[2] 

机构地区：[1]贵州大学理学院,贵阳550025 [2]北京工业大学应用数理学院,北京100124 [3]黄冈师范学院数学与信息科学学院,黄冈438000

出　　处：《应用数学学报》2009年第5期938-950,共13页Acta Mathematicae Applicatae Sinica

基　　金：贵州省科学技术基金(J-2009-2063)资助项目

摘　　要：本文基于多类型复发事件数据,讨论了一个新的加性乘积比率回归模型,该模型包括两部分,其中第一部分为可加Aalen模型,其中协变量影响为加性的且与时间有关.第二部分为Cox回归模型,其中协变量有乘性影响.利用估计方程的方法,给出了该模型中未知参数和非参数函数的一种估计方法,并利用现代经验过程理论证明了所得估计的相合性和渐近正态性.In this article, we propose a new additive multiplicative rates models for multiple type recurrent event data. The proposed model includes two parts. The first part is an Aalen＇s additive model which allows for time-varying covariate effects. The second part is a Cox regression model which contains multiplicative covariate effects. Estimating equation approaches are developed to estimate the nonparametric and parametric components of the proposed model. Based on modern empirical process theory, the consistency and asymptotic normality properties of the proposed estimators are established.

关 键 词：多类型复发事件 加性乘积比率模型 加性Aalen模型 估计方程 

分 类 号：O212.7[理学—概率论与数理统计]