治愈率模型在屏蔽数据可靠性分析中的应用研究  

Application of Cure Rate Model in Reliability Analysis for Masked Data

作　　者：林静[1] 韩玉启[1] 朱慧明[2] 

机构地区：[1]南京理工大学经济管理学院,南京210094 [2]湖南大学统计学院,长沙410079

出　　处：《数学的实践与认识》2007年第7期69-75,共7页Mathematics in Practice and Theory

基　　金：湖南省自然科学基金(05JJ0130)

摘　　要：针对传统方法中的不足,在引入标准治愈率模型的基础上,提出在屏蔽数据可靠性分析中应用一种扩展的治愈率模型的建模方法;分析证明了利用该方法进行建模分析时仅需对模型作较少的前提假设,在信息不足的情况下能够识别出伴随变量对系统寿命分布的影响,进而有效提高模型估计的稳健性.通过运用基于Gibbs抽样的MCMC方法动态模拟出相关参数后验分布的马尔可夫链,给出随机截尾条件下模型参数的贝叶斯估计;实例分析的结果,证明了该模型在可靠性应用中的直观性与有效性.Aimed at the fault of the traditional methods, the standard cure rate model was introduced, and application in reliability evaluation for masked data of extensive cure rate model was given out. By that, the method can improve the robustness of the model and recognize the effect of the covarlates was proved, especially when there was not enough information, because when utilized it needs less hypothesis. Whatrs more, the MCMC method based on Gibbs sampling was used to simulate dynamically the Markov chain of the parameters＇ posterior distribution. And then, the parameters＇ Bayesian estimation in the condition of the random truncated test was given out. Finally, the result of the data analysis was used to prove the objectivity and validity of the model.

关 键 词：贝叶斯分析 可靠性 马尔可夫链蒙特卡罗模拟 治愈率模型 屏蔽数据 

分 类 号：O213.2[理学—概率论与数理统计]