恒加速寿命试验广义半正态分布的统计推断  

Statistical Inference of Generalized Half-Normal Distribution for Constant Accelerated Life Tests

作　　者：韩秀蓉 陈建伟[2] 蔡南莲[1] HAN Xiurong;CHEN Jianwei;CAI Nanlian(School of Science,Jimei University,Xiamen 361021,China;Department of Mathematics and Statistics,San Diego State University,San Diego 92182,USA)

机构地区：[1]集美大学理学院,福建厦门361021 [2]圣地亚哥州立大学数学与统计系,圣地亚哥92182

出　　处：《集美大学学报（自然科学版）》2024年第6期505-512,共8页Journal of Jimei University：Natural Science

基　　金：国家社会科学基金项目“半参数动态空间的PROBIT模型的统计推断与贝叶斯算法及应用”(18BTJ031)。

摘　　要：在恒加速寿命试验下,研究具有广义半正态分布寿命的参数估计问题。证明当恒加速寿命试验的应力和广义半正态分布参数满足一定条件时,寿命形成了一个几何过程;探讨模型参数的最大似然估计、渐近置信区间和bootstrap置信区间的求法。对得到的估计结果进行数值模拟,模拟结果表明,模型参数估计值及置信区间是有效的。In this paper,under constant-stress accelerated life test,the parameter estimations on a population with generalized half-normal distribution were considered.Firstly,it was proved that the product life forms a geometric process when certain conditions of the stress and parameters of generalized half-normal distribution are satisfied.Secondly,the maximum likelihood estimations,asymptotic confidence interval and bootstrap confidence interval estimations of the parameters were proposed.Finally,the obtained estimation results were numerically simulated,which showed that the model parameter estimations and confidence interval estimations were valid.

关 键 词：加速寿命试验 几何过程 最大似然估计 bootstrap置信区间 渐近分布 

分 类 号：O213.2[理学—概率论与数理统计] O212.1[理学—数学]