双边定时截尾下指数威布尔分布的参数估计

Parameter Estimation of Exponentiated Weibull Distribution under Doubly Type-Ⅰ Censored Sample

基于双边定时截尾样本,考察了指数威布尔分布的极大似然估计,由于无法得到似然方程的显式解的表达式,所以证明了解的唯一存在性.用EM算法可以处理不完全数据下参数的估计问题,得到了EM算法估计的迭代式.用R软件进行了随机模拟,结果表明当样本容量n较大时,采用极大似然估计较为准确,当样本容量n较小时,采用EM算法估计较为准确.

Under type-I doubly censored sample,the maximum likelihood method was used to estimate the parameter in exponentiated Weibull distribution. The explicit expression of parameter estimation could not be obtained.But, it was proved that the maximum likelihood estimation had been unique..The EM algorithm can be used to deal with the estimation of parameters under incomplete data, and the iterative formula of EM algorithm is obtained. The random simulation with R software shows that when the sample size n is large, the maximum likelihood estimation is more accurate, and when the sample size n is small, the EM algorithm is more accurate.