广义指数分布参数基于截尾样本在序约束下的极大似然估计  

Maximum Likelihood Estimation of Parameters of Generalized Exponential Distribution Based on Censored Samples under Order Restrictions

作　　者：何彪[1] 李树有[1] 宓颖[1] 

机构地区：[1]辽宁工业大学理学院,辽宁锦州121001

出　　处：《辽宁工业大学学报（自然科学版）》2013年第5期341-346,共6页Journal of Liaoning University of Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(11271064)

摘　　要：研究了多个广义指数分布参数基于截尾样本在序约束下的极大似然估计问题。应用简单迭代法和EM迭代法给出广义指数分布的两个未知参数的极大似然估计的算法,并讨论了其在简单半序、伞形半序、简单树半序和简单环半序约束下的极大似然估计问题,并给出模拟结果。Maximum likelihood estimations of parameters of k generalized exponential distributions based on censored samples under order restrictions were studied. With the simple iterative methods and the EM iteration, the algorithms about the maximum likelihood estimates of the two unknown parameters of the generalized exponential distribution were given. Maximum likelihood estimations of parameters of k generalized exponential distributions based on censored samples under the order restrictions, for example, simple order, umbrella semi-order, simple tree order and simple ring order were discussed, with simulation results given.

关 键 词：广义指数分布 简单迭代法 EM算法 序约束 

分 类 号：O212[理学—概率论与数理统计]