论文部分内容阅读
对于寿命试验普阿松分布的应力参数(失效率),推导了一种平滑的经验贝叶斯估计量。可靠性函数既可利用参数的经验贝叶斯估计量来确定,也可利用求得的可靠性函数期望值来确定。利用Monte Carlo模拟法研究了经验贝叶斯方法的特性。其中,对经验贝叶斯估计量的平均平方误差与传统的估计量例如最小方差无偏估计(MVU)或最大似然估计(ML)进行了比较。结果表明,经验贝叶斯估计量的平均平方误差明显比传统估计量的小。
For the stress parameter (failure rate) of the Pursat distribution over the lifetime test, a smooth empirical Bayesian estimator is derived. The reliability function can be determined either by empirical Bayesian estimators of the parameters or by using the expected reliability function obtained. The Monte Carlo simulation method is used to study the characteristics of empirical Bayesian method. Among them, the average squared error of the empirical Bayesian estimator is compared with the traditional estimators such as the minimum variance unbiased estimation (MVU) or the maximum likelihood estimation (ML). The results show that the average squared error of empirical Bayes estimator is obviously smaller than the traditional estimator.