基于记录值的几何分布模型的Bayes可靠性分析  

Bayesian reliability analysis of geometric distribution model based on record values

作　　者：黄文宜[1] 

机构地区：[1]宜春学院数学与计算机科学学院,江西宜春336000

出　　处：《安徽大学学报（自然科学版）》2015年第5期19-22,共4页Journal of Anhui University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(11361067)

摘　　要：针对来自几何分布的记录值样本,研究可靠度参数的Bayes估计问题.首先给出可靠度的最大似然估计和最小方差无偏估计;在可靠度参数的先验分布为贝塔分布、损失函数为平方误差损失和一类新的加权平方损失函数下,导出了可靠度的Bayes估计;最后通过Monte Carlo数值模拟例子说明论文给出估计的有效性.For the record sample from geometric distribution,Bayes estimation of reliability was studied under squared error loss and a new weighted square error loss functions.The maximum likelihood estimator and minimum variance unbiased estimator were first derived,then Bayes estimators of reliability were obtained when the prior distribution of reliability was conjugate beta prior distribution,and the loss functions were squared error loss and a new weighted square error loss functions.Finally the effectiveness of the estimators was proposed in this paper by using a Monte Carlo numerical example.

关 键 词：记录值 可靠度 BAYES估计 平方损失函数 几何分布 

分 类 号：O212[理学—概率论与数理统计]