IIRCT下泊松分布参数单变点的贝叶斯估计

Bayes Estimation of Parameter of Poisson Distribution with Single Change Point for Random Censoring Test Model with Incomplete Information

首先通过添加数据得到了带有不完全信息随机截尾试验下泊松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方法对参数进行了估计,最后进行了随机模拟,试验结果表明参数贝叶斯估计的精度较高.

This paper firstly obtained the complete-data likelihood function of Poisson distribution for IIRCT after adding data,then studied the full conditional distributions of change-point position and other parameters,and estimated the parameters by MCMC method of Gibbs sampling together with Metropolis-Hastings algorithm.Finally random simulation tests were conducted,and the results showed that Bayes estimations of the parameters were fairly accurate.