在可靠性统计的定时截尾寿命试验中,最后一个失效时间与截尾时刻之间的信息常被忽略.利用EM算法来处理这一情形,得出Weibull分布中尺度参数的迭代解.并将EM算法与传统的极大似然估计进行了比较,可以看出EM算法明显优于传统的极大似然估计.
The paper deals with Type-I Censoring sample by EM algorithm, in which some information is usually neglected. The study has got the iterative solution of the parameter of Weibull distribution. Moreover. a comparison of EM algorithm and maximum likelihood estimation is made. It is a clear result that the EM algorithm is superior to maximum likelihood estimation.