一种新的J-M模型参数区间估计工程算法

New confidence interval estimation method for parameters of J-M model

将Bayes统计方法应用于J-M模型参数的区间估计研究.首先给出J-M模型参数的基于杰弗莱原则的先验分布,在此基础上给出模型参数的Bayes后验分布密度,由此得出模型参数的Bayes点估计.然后,在点估计基础上,给出模型参数的Bayes区间估计方法.最后,分别将J-M模型参数的最大似然、最小二乘以及Bayes区间估计法应用于多组公开的失效数据集,求解模型参数的区间估计值.计算结果表明,Bayes区间估计算法的计算精度、适用性以及稳定性是三者当中最好的.

Bayes statistics is applied to confidence interval estimation for p ly, Prior distribution of J-M model＇s parameters based on Jeffrey principl arameters of J-M model. First e is obtained to get the posteri or distribution density which is used to achieve the point estimation of parameters based on Bayes theorem. Bayes interval estimation method is developed consequently. Finally, three interval estimation methods respectively based on maximum likelihood, least square and Bayes theorem are applied to several open failure data sets. The results show that the precision, stability and applicability of interval estimation method based on Bayes theorem is the best in these three methods.