广义指数比例风险模型下区间删失数据的贝叶斯估计  

Bayesian Estimation of Interval Censored Data with Proportional Hazards Model under Generalized Exponential Distribution

作　　者：刘斌霞 王纯杰 LIU Bin-xia;WANG Chun-jie(College of Statistics and Mathematics,Zhongnan University of Economics and Law,Wuhan 430000,China;College of Mathematics and Statistics,Changchun University of Technology,Changchun 130012,China)

机构地区：[1]中南财经政法大学统计与数学学院,湖北武汉430000 [2]长春工业大学数学与统计学院,吉林长春130012

出　　处：《数理统计与管理》2023年第2期293-301,共9页Journal of Applied Statistics and Management

基　　金：国家自然科学基金项目(11671054);中南财经政法大学研究生教育创新计划项目(201931311)。

摘　　要：本文研究了失效时间服从广义指数分布,且风险率函数为比例风险模型时,Ⅱ型区间删失数据的贝叶斯估计.假定参数的先验分布为无信息先验,建立贝叶斯层次模型从而得到后验密度函数.通过MH算法得到参数估计值,数值模拟结果验证了所提方法的有效性.最后将所提方法应用到乳腺癌患者和血友病患者这两个实际数据中进行分析.In this paper,we investigate the Bayesian estimation of intervalⅡcensored data when failure time obeys the generalized exponential distribution and the hazard function satisfies the proportional hazards model.Assuming that the prior distribution of parameter is non-informative prior,then establish the Bayesian hierarchal model to obtain the posterior density function.Using MH algorithm to obtain the parameter estimation,and the results of numerical simulation verify the effectiveness of the proposed method.Finally,the proposed method is applied to the real data of breast cancer and hemophilia.

关 键 词：广义指数分布 Ⅱ型区间删失数据 比例风险模型 贝叶斯估计 

分 类 号：O212[理学—概率论与数理统计]