| IIRCT下负二项分布参数多变点的贝叶斯估计 |
| |
| 引用本文： | 何朝兵,刘华文.IIRCT下负二项分布参数多变点的贝叶斯估计[J].郑州大学学报(自然科学版),2014(2):11-15. |
| |
| 作者姓名： | 何朝兵 刘华文 |
| |
| 作者单位： | [1]安阳师范学院数学与统计学院,河南安阳455000 [2]山东大学数学学院,山东济南250100 |
| |
| 基金项目： | 国家自然科学基金资助项目,编号61174099;河南省教育厅科学技术研究重点项目,编号14B110011. |
| |
| 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负二项分布的完全数据似然函数.给出了变点位置和其他参数的满条件分布.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.详细介绍了MCMC方法的实施步骤,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.
|
| 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 |
Bayesian Estimation of Parameter of Negative Binomial Distribution with Multiple Change Points for IIRCT |
| |
| HE Chao-bing,LIU Hua-wen.Bayesian Estimation of Parameter of Negative Binomial Distribution with Multiple Change Points for IIRCT[J].Journal of Zhengzhou University (Natural Science),2014(2):11-15. |
| |
| Authors: | HE Chao-bing LIU Hua-wen |
| |
| Institution: | 1.School of Mathematics and Statistics,Anyang Normal University,Anyang 455000,China; 2.School of Mathematics,Shandong University,Ji＇nan 250100,China) |
| |
| Abstract: | By filling in the missing data of the life variable,the complete-data likelihood function of negative binomial distribution for IIRCT was obtained.The full conditional distributions of change-point positions and other parameters were given.Every parameter was sampled from the full conditional distributions respectively,using MCMC method of Gibbs sampling together with Metropolis-Hastings algorithm.The implementation steps of MCMC method were introduced in detail.Gibbs samples of the parameters were obtained,and the means of Gibbs samples were taken as Bayesian estimations of the parameters.The random simulation test results showed that Bayesian estimations of the parameters were fairly accurate. |
| |
| Keywords: | complete-data likelihood function full conditional distribution MCMC method Gibbs sampling Metropolis-Hastings algorithm |
| 本文献已被 维普 等数据库收录！ |
