针对指标筛选问题,提出了一种基于指标主要信息含量的指标筛选方法.以方差贡献率为权重对因子载荷的绝对值进行加权,取加权结果的最大值作为指标的信息含量;依据最大加权因子载荷标准确定累积信息含量比率,进而通过累积信息含量比率的大小进行指标的筛选.最后,通过一个实例将提出的指标筛选方法与现有因子分析指标筛选方法进行了对比分析,以此说明所提出的指标筛选方法的可行性及有效性.
This paper proposes an indices' screening method based on each index's main information content to solve their reduction problem.First,it uses the proportion of variance to weight the absolute value of loading,and defines its maximum as the information content of an index.Second,by means of the maximum weighted loading standard,it calculates the proportions of the indices' accumulated information contents,whose size is used as the criteria to screen the indices.Finally,using a numerical illustration,the proposed method and the existing factor analysis method are contrastively analyzed,and the feasibility and validity of the proposed method are illustrated.