超饱和试验设计中的参数估计问题

发布时间：2018-07-07 15:55

  本文选题：超饱和设计 + 可比估计　； 参考：《数理统计与管理》1997年05期

【摘要】：超饱和试验设计是一种因子个数大于试验次数的试验设计；它是工业统计中的一个新的研究课题，在工业质量控制中有重要的应用。在超饱和设计中，因子效应参数的无偏估计一般不存在，一个因子效应的估计会受到其它因子效应的影响，这种影响是设计本身带来的，称之为交互影响。本文讨论两种参数估计：最小方差可比估计与最小交互影响可比估计。模拟计算的结果显示，在正确地搜寻活动因子的能力方面，，最小交互影响可比估计在与实际情况相接近的模拟条件下强于最小方差可比估计。
[Abstract]:The design of supersaturated test is a kind of experimental design whose number of factors is greater than the number of tests, and it is a new research subject in industrial statistics, which has important application in industrial quality control. In supersaturated design, the unbiased estimation of factor effect parameters generally does not exist, and the estimation of one factor effect is affected by other factor effects, which is caused by the design itself, which is called interaction effect. In this paper, we discuss two kinds of parameter estimation: minimum variance comparable estimation and minimum interaction comparable estimation. The simulation results show that the minimum interaction comparable estimation is stronger than the minimum variance comparable estimate under the simulation condition which is close to the actual situation in terms of the ability to search correctly for the activity factors.
【作者单位】： 清华大学应用数学系
【分类号】：C931.1


本文编号：2105465