统计综合评价中无量纲化方法选择的t检验思路  

T-test Approach for Dimensionless Method Selection in Statistical Comprehensive Evaluation

作　　者：谢忠秋 Xie Zhongqiu(School of Business,Jiangsu University of Technology,Changzhou Jiangsu 213001,China)

机构地区：[1]江苏理工学院商学院,江苏常州213001

出　　处：《统计与决策》2020年第22期32-36,共5页Statistics & Decision

基　　金：江苏省社会科学基金重点项目(16JYA004)。

摘　　要：在综合评价中,对于所运用的无量纲化方法是否合适,人们一直无法作出正确的判断。一个有效的方法就是为其提供规范的统计假设检验。为此,文章运用统计假设检验理论,借助协方差-方差矩阵迹的性质,构建了t检验方法,为判断无量纲化方法是否合适提供了统计判断依据。案例分析表明,所提出的无量纲化方法的检验法则不仅有效,而且可操作,具有较好的推广应用价值。In the comprehensive evaluation,people have been unable to make a correct judgment on whether the dimensionless method is are appropriate or not.An effective method is to provide standard statistical hypothesis tests.For this reason,this paper applies the theory of statistical hypothesis test and the property of covariance-variance matrix trace to structure the t-test approach,which provides a statistical basis for judging whether the dimensionless method is appropriate.The case study shows that the test rules of the proposed dimensionless method are not only effective,but also operable,and have good popularization and application value.

关 键 词：无量纲化 迹 检验统计量 四阶矩 

分 类 号：O21[理学—概率论与数理统计]