分析正交饱和设计的零效应搜索法中s=2的研究  

Procedure of Searching Zero Effects of Analysis Orthogonal Saturated Designs-the Research about s=2

作　　者：张晓琴[1] 乔志勇[1] 

机构地区：[1]山西大学数学科学学院,山西太原030006

出　　处：《山西大学学报（自然科学版）》2008年第3期372-374,共3页Journal of Shanxi University(Natural Science Edition)

摘　　要：文[4]中方法的一个不足之处是只能适用于s≥3,即至少有2个零效应的2水平饱和设计情形,但对s≤2的情形没有进行研究.本文在文[4]的基础上解决只有一个零效应(即s=2)的情形.模拟结果显示,当只有一个零效应时,零效应搜索法也能很好的识别出饱和设计中的显著因子,且对误差方差的估计也很好.The method in [-4] has a shortcoming which is fit for the case s≥3,namely it only slove the case that has at least two zero effects in orthogonal saturated designs ,not the case of s≤2. The case of s= 2 based on the method in [4] is discussed in this paper. The result of simulation shows that the procedure of searching zero effects can also identify the active effects in the case of s=2,and the estimator of the error variance is very well.

关 键 词：正交饱和设计 次序统计量 零效应搜索法 显著效应 W-统计量 

分 类 号：O212.1[理学—概率论与数理统计]