左截断右删失数据下伽玛分布参数多变点的贝叶斯估计  

Bayesian Estimation of Parameter of Gamma Distribution with Multiple Change Points for Randomly Truncated and Censored Data

作　　者：何朝兵[1] 

机构地区：[1]安阳师范学院数学与统计学院,河南安阳455000

出　　处：《四川师范大学学报（自然科学版）》2015年第3期398-403,共6页Journal of Sichuan Normal University（Natural Science）

基　　金：国家自然科学基金(61174099);河南省教育厅自然科学基金(2011B110001)资助项目

摘　　要：首先通过添加数据得到了左截断右删失数据下伽玛分布的完全数据似然函数,然后研究了变点位置和其它参数的满条件分布,接着利用Gibbs抽样与Metropolis-Hastings算法相结合的MCMC(Markov Chain Monte Carlo)方法得到了参数的Gibbs样本,把Gibbs样本的均值作为各参数的贝叶斯估计,随机模拟试验的结果表明各参数贝叶斯估计的精度都较高.In this paper,the complete-data likelihood function of g（em）ma distribution for randomly truncated and censored data was obtained after adding data.The full conditional distributions of change-point positions and other parameters were investigated to get Gibbs samples of the parameters by MCMC method of Gibbs sampling together with Metropolis-Hastings algorithm.Taking the means of Gibbs samples as the Bayesian estimations of the parameters,random simulation test results show that the Bayesian estimations of the parameters are fairly accurate.

关 键 词：完全数据似然函数 满条件分布 MCMC方法 GIBBS抽样 Metropolis-Hastings算法 

分 类 号：O213.2[理学—概率论与数理统计] O212.8[理学—数学]