Bayes可靠性增长模型先验参数的确定  

Selection of Prior Parameters to Bayesian Reliability Growth Model

作　　者：韩明[1] 

机构地区：[1]宁波大学数学系,315211

出　　处：《系统工程与电子技术》2000年第4期88-90,共3页Systems Engineering and Electronics

基　　金：全国统计科学研究 (计划 )重点项目!(980 0 3);浙江省教委基金项目!(971 0 7) ;宁波大学 2 1世纪人才基金项目!(982 1 0 2G)资

摘　　要：Gamma分布Gamma(a ,b) (其中a >0 ,b>0 ) ,通常用来作指数分布Bayes可靠性增长模型的先验分布 ,参数a和b的不同 ,将使可靠性增长试验的评估结果有很大的差异。根据可靠性增长和“弱”先验分布原理 ,通过对数值例的分析、比较 ,得出了先验参数的三种取法对可靠性评估结果的影响 ,并给出了指数分布Bayes可靠性增长模型的先验参数a和b的一种确定方法。Gamma distribution Gamma(a,b)(in which:a>0,b>0) is always used as a prior distribution of Bayesian exponential distribution reliability growth model. Different value of a and b will make considerable difference in evaluating results. Three kinds of methods to select prior parameters are observed by using according to the reliability growth and “weak”prior distribution theory, Bayesian prior evaluation is calculated and is compared with each other, the effect of different Gamma distribution parameters on reliability growth evaluating results is found out, and the selection method of prior parameters a and b to exponential distribution Bayesian reliability growth model are also given out.

关 键 词：可靠性增长模型 指数分布 贝叶斯定理 

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