两水平无重复因析试验散度效应的估计  

Estimation of Dispersion Effects in Two-level Unreplicated Factorial Experiments

作　　者：李济洪[1] 王钰[1] 张星[1] 

机构地区：[1]山西大学计算中心,太原030006

出　　处：《应用数学学报》2010年第4期710-722,共13页Acta Mathematicae Applicatae Sinica

基　　金：国家自然科学基金(60873128)资助项目

摘　　要：本文对两水平无重复因析试验给出了散度效应的一种新的估计,称为AMH估计,改进了文献中散度效应的较好的MH估计的一个缺陷,给出并证明了AMH估计的无偏条件,证明了AMH估计比MH估计有更小的方差下界.最后通过模拟试验比较了AMH和MH估计的偏度,方差和均方误差.A new estimator of the dispersion effects,called the AMH,in two-level unreplicated factorial experiments is given.The AMH modifies a shortcoming of the MH estimator which was showed that there are many good properties in literatures.The unbiased condition of AMH is persented.We also prove that the AMH has more lower bound of variance than MH.Finally,the bias,variance and mean square error of AMH and MH are compared by simulated experiments.

关 键 词：散度效应模型 MH方法 均方误差 模拟 

分 类 号：O212.1[理学—概率论与数理统计]