加法风险率模型下聚类的当前状态数据的回归分析

发布时间：2018-05-15 08:54

本文选题：加法风险率模型 + 当前状态数据　；参考：《武汉大学》2017年硕士论文

【摘要】：生存分析最初起源于现代医学,工程等科学研究中的实际问题,是数理统计研究中的一个重要分支。自二十世纪七十年代中期以来,生存分析迅速发展,它着重对删失数据进行研究。生存分析理论结合概率统计理论,不仅能有效的处理生活中的常见删失数据问题,而且促进了数理统计的发展。在生存分析研究中,当前状态数据常常出现。当前状态数据出现时失效时间是不能直接观察到的,而只是知道失效时间在删失时间的前面还是后面。我们都知道,对于失效时间的研究方法很多。本文研究了在加法风险率模型下聚类的当前状态数据(I型区间删失数据)的回归分析的问题.该文着重考虑了当相关的失效时间数据与簇类的规模相关时的情形.为了研究这一问题,我们介绍了一个簇内再抽样方法并且给出了相应估计量的极限分布理论,最后,我们还通过大量的模拟研究验证了该种方法的有效性.
[Abstract]:Survival analysis originated from the practical problems in modern medicine, engineering and other scientific research, and it is an important branch of mathematical statistics research. Since the mid-seventies of the 20th century, survival analysis has developed rapidly, and it focuses on the research of censored data. The theory of survival analysis combined with the theory of probability and statistics can not only effectively deal with the common problem of deleted data in daily life, but also promote the development of mathematical statistics. In the research of survival analysis, current state data often appear. The failure time when the current state data appears can not be observed directly, but only knows whether the failure time is before or behind the censored time. As we all know, there are many methods for studying failure time. In this paper, we study the problem of regression analysis of the current state data (type I interval censored data) of clustering under the additive risk rate model. This paper focuses on the case where the relevant failure time data are related to the size of the cluster class. In order to study this problem, we introduce a method of intra-cluster re-sampling and give the limit distribution theory of the corresponding estimator. Finally, we verify the validity of this method by a large number of simulation studies.
【学位授予单位】：武汉大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O212.3

【相似文献】