基于FDS准则的响应曲面设计预测能力比较OACHSSCDCSTPCD
FDS Criterion-based Prediction Ability Comparison of Response Surface Design
预测能力是实际应用工作者评价设计优劣的重要准则之一,一般地,通过方差散布图可以对一个固定球面上的设计预测能力进行比较和选择,但无法获得对整个设计区域上的比较结果.文章介绍了基于正交表的响应曲面设计(OACD)以及基于极小极大准则的改进设计(OACM),并应用全区域预测方差占比准则——FDS准则对这两类中心组合设计进行预测能力的比较.结果表明,在一个较小方差范围内,OACD得到的FDS结果要优于OACM,而对于大部分给定方差值,OACM的FDS结果要优于OACD.
Prediction ability is an important criterion for practical application workers to evaluate the quality of a design.In general,the design prediction ability of a fixed sphere can be compared and selected by means of the variogram,but the compari-son results on the whole design area cannot be obtained.This paper first introduces response surface design(OACD)based on or-thogonal array and the modified design based on minimax criterion(OACM),and then compares the prediction ability between these two kinds of central combined designs by using the whole region prediction variance ratio criterion—FDS criterion.The re-sults show that in a small variance range,the FDS obtained by OACD is better than that of OACM,but for most given variance values,the FDS of OACM is better than that of OACD.
周祥;陈雪平;严静;林金官
苏州高等职业技术学校(江苏联合职业技术学院苏州分院)教务处,江苏 苏州 215000江苏理工学院 数理学院,江苏 常州 213001南京审计大学 统计与数据科学学院,南京 211815
数学
试验设计最优设计正交设计FDS准则
experimental designoptimal designorthogonal designFDS criterion
《统计与决策》 2024 (002)
41-45 / 5
国家自然科学基金资助项目(11971204);江苏省自然科学基金资助项目(BK20200108);江苏理工学院中吴青年创新人才支持计划(江理人工[2021]137号)
评论