双边定数截尾下Pareto分布的可靠性分析  

Reliability Analysis for Pareto Distribution under Type-Ⅱ Doubly Censored Sample

作　　者：侯华蕾[1] 师义民[1] 李豪亮[1] 

机构地区：[1]西北工业大学应用数学系,陕西西安710072

出　　处：《数理统计与管理》2009年第5期826-830,共5页Journal of Applied Statistics and Management

基　　金：国家自然科学基金资助(No.70471057)

摘　　要：基于双边定数截尾样本,选取未知参数的先验分布为无信息先验和Gamma分布,分别在平方损失和LINEX损失下,研究了Pareto分布的形状参数和可靠性指标(可靠度和失效率)的Bayes估计.为了研究估计的精度,采用Monte-Carlo模拟的方法给出了数值检验的例子.结果表明在LINEX损失下并选用Gamma先验分布时,参数的Bayes估计是最优的.Based on type-Ⅱ doubly censored sample, choosing the Non-informative prior and Gamma prior distributions for the unknown parameter , the estimators of and reliability measures（failure rate r（t）, reliability R（t）） for Pareto distribution are obtained relative to squared error loss and LINEX loss functions. In order to investigate the accuracy of estimations, an illustrative example is examined numerically by means of Monte-Carlo simulation. The conclusion is that the Bayes estimator is the best estimator under LINEX loss function and Gamma prior distribution.

关 键 词：双边定数截尾 PARETO分布 BAYES估计 可靠性分析 

分 类 号：O212[理学—概率论与数理统计] O213.2[理学—数学]