对数稳定分布加速寿命试验的贝叶斯分析  

BAYESIAN STATISTICAL ANALYSIS OF ACCELERATED LIFE TESTS FOR LOG-STABLE DISTRIBUTION

作　　者：李海芬[1] 汤银才[1] 

机构地区：[1]华东师范大学金融与统计学院,上海200241

出　　处：《系统科学与数学》2011年第4期448-457,共10页Journal of Systems Science and Mathematical Sciences

基　　金：国家科技863计划(2007AA01Z189);国家自然科学基金(90718013;10571057);上海市教育发展基金(2007CG58)资助课题

摘　　要：对数稳定分布与对数正态分布相比具有厚尾等良好的特性,由于没有显式的密度函数,其参数估计有一定的困难.运用Gibbs抽样方法给出了对数稳定分布的Bayes参数估计,与对数正态分布参数估计进行了比较,并应用到恒加寿命试验数据的分析中.模拟研究说明了该方法的有效性,对数稳定分布作为新的寿命分布是合适的.Log-stable distribution has better properties than log-normal distribution.It is difficult to estimate the parameters because of the lack of closed formulae for its density and cumulative distribution.Bayesian approach via Gibbs sampling is used to estimate the parameters in the log-stable distribution and compared with log-normal distribution.The approach is further applied to the constant stress accelerated life test.Simulation shows that the approach is efficient and the log-stable distribution is suitable as a new life time distribution.

关 键 词：对数稳定分布 BAYES方法 恒加寿命试验 GIBBS抽样 

分 类 号：O212.8[理学—概率论与数理统计]