k阶Erlang分布参数单变点的贝叶斯估计  

Bayesian estimation of parameter on k-th Erlang distribution with single change point

作　　者：季海波[1] 王丽 JI Haibo;WANG Li(School of Liberal and Science,Suqian College,Suqian 223800,China)

机构地区：[1]宿迁学院文理学院,江苏宿迁223800

出　　处：《高师理科学刊》2021年第12期10-13,共4页Journal of Science of Teachers＇College and University

基　　金：江苏省宿迁市科技计划项目(Z2019110)。

摘　　要：利用贝叶斯方法研究了k阶Erlang分布参数单变点模型,得到该分布的似然函数.利用Fisher信息矩阵选取无信息先验分布,从而得到每个参数的满条件分布.使用Gibbs抽样与M-H算法相结合的MCMC方法对各参数进行随机模拟得到贝叶斯估计.模拟结果表明,各参数的估计值的精度在较高水平上都是有效的.Bayesian method has been used to study the single change point model of k-th Erlang distribution,and the likelihood function of the distribution was get.Non-information prior distribution was selected by Fisher information matrix,and the full conditional distributions of every parameters were obtained.The MCMC method combining Gibbs sampling and M-H algorithm is used to randomly simulate each parameter to obtain Bayesian estimation.The simulation results showed that Bayesian estimation of parameters were fairly accurate at higher level.

关 键 词：k阶Erlang分布 Fisher信息矩阵 MCMC方法 贝叶斯估计 

分 类 号：O212.8[理学—概率论与数理统计]