正态总体均值与标准差比在树序约束下的最大似然估计  

Maximum Likelihood Estimation of Ratios of Means and Standard Deviations from Normal Populations under Tree-order Restriction

作　　者：李树有[1] 史宁中[2] 赵世舜[1] 

机构地区：[1]吉林大学数学学院 [2]东北师范大学数学与统计学院,长春130024

出　　处：《吉林大学学报（理学版）》2006年第6期893-897,共5页Journal of Jilin University:Science Edition

基　　金：国家自然科学基金(批准号:10431010);辽宁省教育厅基金(批准号:20031066)

摘　　要：考虑3个正态总体均值和方差都是未知参数及所取的3个正态总体样本数不等时,均值和标准差的比在树序约束下的极大似然估计.根据PAVA算法的思想,给出了均值和标准差的比在树序约束下的极大似然估计的计算方法.For three normal populations with unknown means and variances, assume that there is a tree-order restriction between ratios of means and standard deviations. According to the ideas of the PAVA algorithm, when sample numbers of .three normal： populations are different, a procedure of obtaining the maximum likelihood estimators of means and standard deviations under tree-order restrictions is proposed.

关 键 词：半序约束 树序约束 最大似然估计 似然函数 PAVA算法 

分 类 号：O212.1[理学—概率论与数理统计]