多类型复发事件数据半参数模型的统计推断

发布时间：2018-03-15 07:19

本文选题：多类型复发事件　切入点：估计方程　出处：《中央民族大学》2015年硕士论文　论文类型：学位论文

【摘要】：在生物医学、机械工业、人口学和生态学等领域中,研究的个体对象在一定的时间内可能经历多次我们感兴趣的事件,例如机器发生故障,患者疾病的发作等,这种事件在生存分析中被称复发事件,相应的复发时间组成的数据称为复发事件数据。当只对一种事件类型的结果感兴趣时,得到的就是单类型复发事件数据,如果研究中感兴趣的事件类型结果不止一种时,得到的就是多类型复发事件数据。由于复发事件数据具有次序性和相依性,并且由于删失时间的存在,导致复发事件数据往往是很复杂的。自上世纪80年代起,越来越多的国内外学者致力于研究解决这方面的问题,也得到了许多具有实际意义的结果。关于复发事件数据,我们感兴趣的往往是协变量对于复发事件的影响程度,至今为止学者们提出了各种各样的回归模型来研究协变量对于复发事件的影响,学者们主要从复发事件的间隔时间和复发事件本身这两方面入手来研究复发事件数据。本文第一章首先介绍了复发事件的一些背景知识和常用概念以及生存分析在复发事件研究用到一些定义和基础知识和本文用到的基本模型和基本估计方法,第二章综述了学者们在不同的角度下对复发事件数据进行研究的一些成果,本文的第三章在Lin等人于2000年提出的模型的基础上,提出了一个在一定条件下可以作为不同事件类型在时刻t与参考事件类型的平均复发次数的比较的模型,并给出该模型的参数估计值,以及估计的渐进性质。
[Abstract]:In biomedical, mechanical, demographic, and ecological fields, individual subjects may experience many events of interest to us over a certain period of time, such as machine failure, patient illness, etc. This kind of event is called recurrence event in survival analysis, and the corresponding recurrence time component data is called recurrence event data. When only interested in the result of one type of event, we get the single type recurrence event data. If there is more than one result of the event type of interest in the study, what is obtained is the multi-type recurrence event data, because of the order and dependence of the recurrence event data, and because of the existence of the censored time, Since -20s, more and more scholars at home and abroad have devoted themselves to the research and solution of this problem, and obtained a lot of practical results. We are often interested in the degree of influence of covariables on recurrent events. So far, scholars have proposed various regression models to study the effects of covariables on recurrent events. The first chapter of this paper introduces some background knowledge, common concepts and survival analysis of recurrent events. In the study of recurrence events, some definitions and basic knowledge, basic models and basic estimation methods used in this paper are used. In the second chapter, we summarize some achievements of scholars' research on recurrence event data from different angles. The third chapter is based on the model proposed by Lin et al in 2000. In this paper, we present a model which can be used to compare the average recurrence times of different event types at time t with that of reference event types under certain conditions, and give the parameter estimates of the model and the asymptotic properties of the estimations.
【学位授予单位】：中央民族大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：O212

【相似文献】