基于混合Ⅰ型删失数据的威布尔模型精确推断(英文)  

Exact Inference Based on Type-Ⅰ Hybrid Censored Sample from Weibull Distribution

作　　者：程从华[1] 陈进源[2] 

机构地区：[1]湛江师范学院数学与计算科学学院,广东湛江524048 [2]兰州大学数学与统计学院,甘肃兰州730000

出　　处：《数学研究》2012年第3期271-281,共11页Journal of Mathematical Study

基　　金：supported by NSFC(71171103);Research Fund for the Docoral Program of Zhanjiang Normal University(ZL1102)

摘　　要：威布尔分布是可靠性和寿命测试试验中常用的模型.本文中,我们考虑了基于混合Ⅰ型删失数据的威布尔模型精确推断.我们得到了威布尔分布未知参数最大似然估计的精确分布以及基于精确分布的置信区间.由于精确分布函数较为复杂,我们也给出了未知参数的另外几种置信区间,比如,基于近似方法的置信区间,Bootstrap置信区间.为了评价本文的方法,我们给出了一些数值模拟的结果.The Weibull distribution is one of the popular distributions in reliability and life testing experiments. In this paper, we discuss the exact inference of Weibull distribution by considering the Type-I hybrid censoring schemes. We derive the exact conditional distribution of the maximum hkelihood estimator （MLE） as well as the exact confidence intervals for the scale parameter of the Weibull distribution. We also propose some other methods to construct the confidence intervals, for example, asymptotic method, bootstrap method. Some simulation studies are carried out to assess their performances

关 键 词：最大似然估计 混合Ⅰ型删失 置信区间 威布尔分布 

分 类 号：O212.1[理学—概率论与数理统计]