逐次Ⅱ型截尾样本下剩余寿命ZZ分布参数估计  

Estimation of parameters of ZZ distribution for residual lifetime prediction from progressive Type-Ⅱcensoring data

作　　者：袁喜琴 闫在在[1] YUAN Xiqin;YAN Zaizai(School of Science,Inner Mongolia University of Technology,Hohhot 010051,China)

机构地区：[1]内蒙古工业大学理学院,呼和浩特010051

出　　处：《内蒙古工业大学学报（自然科学版）》2022年第6期488-496,共9页Journal of Inner Mongolia University of Technology：Natural Science Edition

基　　金：国家自然科学基金项目(11861049);内蒙古自治区自然科学基金项目(2022MS01006,2020LH01002);内蒙古自治区直属高校基本科研业务费项目(JY2022083)。

摘　　要：在逐次Ⅱ型截尾试验下,提出了一种基于ZZ分布的剩余寿命预测模型,给出其参数和可靠性特征的极大似然估计,及其渐进置信区间,Bootstrap置信区间。并在广义熵损失函数(GELF)下,运用Metropolis-Hastings抽样,讨论其参数及可靠性特征的Bayes估计和最高后验密度(HPD)可信区间。最后,通过真实数据验证,验证了本文方法的可行性和有效性。In this paper,a residual life prediction model based on ZZ distribution is proposed under progressive Type-Ⅱcensoring data.The maximum likelihood estimation of its parameters,reliability characteristics,asymptotic confidence interval and Bootstrap confidence interval are given.Under the generalized entropy loss function(GELF),Metropolis-Hastings sampling is applied.The Bayes estimation and the highest posterior density(HPD)confidence interval of its parameters and reliability characteristics are discussed.Finally,a real example is given to illustrate the feasibility and validity of the methods proposed in this paper.

关 键 词：剩余寿命 ZZ分布 逐次Ⅱ型截尾样本 极大似然估计 BAYES估计 Metropolis-Hastings抽样 

分 类 号：O212.5[理学—概率论与数理统计]