统计与决策2024，Vol.40Issue(4)：44-49,6.DOI:10.13546/j.cnki.tjyjc.2024.04.008
无量纲化方法选择及最优无量纲化方法构建
Selection of Dimensionless Method and Construction of Optimal Dimensionless Method
摘要
Abstract
In order to solve the problem that it is difficult to select the dimensionless method in the comprehensive evaluation process,firstly,the dimensionless methods are preliminarily screened based on the distribution characteristic invariance and vari-ation characteristic invariance,only to find that only the scaling methods can keep the distribution characteristics and variation characteristics unchanged at the same time.Secondly,a statistical test method for the effectiveness of the dimensionless methods is constructed,and the effectiveness of various scaling methods is tested,the results show that not all scaling methods could effec-tively eliminate the dimensional difference between indicators.Then,the measurement model of the dimensionless information loss rate is established;the main steps of dimensionless methods selection are proposed,and an optimal dimensionless model is es-tablished based on the principle that the dimensionless methods should reduce the information loss rate within the indicator while effectively eliminating the dimensional difference between indicators.Finally,a large number of numerical simulation experiments are carried out to select the dimensionless method and solve the optimal dimensionless model.The research results indicate that based on the above steps,the dimensionless method selection can be carried out more scientifically and accurately,that the opti-mal dimensionless method constructed is different from any existing dimensionless methods,which can not only effectively elimi-nate the dimensional difference between indicators,but also prevent the excessive processing of data from resulting in unnecessary information loss.关键词
无量纲化/有效性检验/信息损失速率/数值模拟实验Key words
nondimensionalization/effectiveness test/information loss rate/numerical simulation experiment分类
社会科学引用本文复制引用
高晓红,李兴奇..无量纲化方法选择及最优无量纲化方法构建[J].统计与决策,2024,40(4):44-49,6.基金项目
国家自然科学基金资助项目(11261001) （11261001）
云南省地方本科高校基础研究联合专项面上项目(202301BA070001-054 （）
202301BA070001-059) （）
