IIRCT下泊松分布参数多变点模型的贝叶斯估计  

Bayesian Estimation of Parameter of Poisson Distribution with Multiple Change Points for Random Censoring Test Model with Incomplete Information

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：《数学的实践与认识》2014年第11期176-184,共9页Mathematics in Practice and Theory

基　　金：河南省教育厅自然科学基金(2011B110001)

摘　　要：通过添加缺损的寿命变量数据得到了带有不完全信息随机截尾试验下泊松分布参数多变点模型的完全数据似然函数,研究了变点位置参数和其它参数的满条件分布.利用Gibbs抽样与Metropolis-Hastings算法相结合的MCMC方法对各参数的满条件分布分别进行了抽样,把Gibbs样本的均值作为各参数的贝叶斯估计,并且详细介绍了MCMC方法的实施步骤.最后进行了随机模拟试验,试验结果表明各参数贝叶斯估计的精度都较高.By filling in the missing data of the life variable, the complete data likelihoodfunction of Poisson distribution with multiple change points for IIRCT is obtained. Thefull conditional distributions of change-point positions and other parameters are studied.Every parameter is sampled from its full conditional distribution respectively, using MCMCmethod of Gibbs sampling together with Metropolis-Hastings algorithm, and the means ofGibbs samples are taken as Bayesian estimations of the parameters. The implementationsteps of MCMC method are introduced in detail. The random simulation tests are conducted,and the results show that 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：O212.8[理学—概率论与数理统计]