艾拉姆咖分布参数变点的Bayes估计  

Bayes Estimation of Parametric Change-point ofЭрлангаDistribution

作　　者：张晗 周菊玲[1] 董翠玲[1] 刘贞 ZHANG Han;ZHOU Ju-ling;DONG Cui-ling;LIU Zhen(School of Mathematics and Science,Xinjiang Normal University,Urumqi 830017,China)

机构地区：[1]新疆师范大学数学科学学院,新疆乌鲁木齐830017

出　　处：《数学的实践与认识》2021年第7期146-151,共6页Mathematics in Practice and Theory

基　　金：国家自然科学基金青年项目(11801488);新疆师范大学校级科研平台招标课题(XJNUSYS2019B05);新疆师范大学重点实验室项目(XJNUSYS082018A01);新疆师范大学教改工程项目(SDJG2018-46)。

摘　　要：研究了艾拉姆咖分布变点估计的非迭代抽样算法(IBF)和MCMC算法.在贝叶斯框架下,选取无信息先验分布,得到关于变点位置的后验分布和各参数的满条件分布,并且详细介绍了IBF算法和MCMC方法的实施步骤.最后进行随机模拟试验,结果表明两种算法都能够有效的估计变点位置,并且IBF算法的计算速度优于MCMC方法.This article studies the non-iterative sampling algorithm(IBF)and MCMC algorithm for the estimation of the change-point of theЭрлангаdistribution.Under the Bayesian framework,select the uninformed prior distribution to obtain the posterior distribution of the position of the change-point and the fully conditional distribution of each parameter and then the implement steps of the IBF algorithm and the MCMC method are introduced in detail.Finally,a random simulation test is carried out.The results show that both algorithms can effectively estimate the position of the change-point and the calculation speed of the IBF algorithm is better than that of the MCMC algorithm.

关 键 词：艾拉姆咖分布 IBF算法 MCMC算法 

分 类 号：O212.8[理学—概率论与数理统计]