综合评价指标约简算法综述  

A Review of Comprehensive Evaluation Index Reduction Algorithms

作　　者：余良武 郭文勇[1] 黄家宁 伍哲 曹辰昊 YU Liangwu;GUO Wenyong;HUANG Jianing;WU Zhe;CAO Chenhao(Naval University of Engineering,Wuhan 430033,China)

机构地区：[1]海军工程大学,武汉430033

出　　处：《机械工程师》2023年第9期12-14,17,共4页Mechanical Engineer

摘　　要：针对综合评价问题构建指标体系时求全和求精这一对矛盾体的寻优问题,研究了现有2大类7种指标约简算法:主成分分析和因子分析法、极小广义方差法、极大不相关法、灰色关联分析法、基于差别矩阵的指标约简算法、基于信息熵的指标约简算法、基于粒计算的指标约简算法。研究了这些算法的原理和各自特点及应用场合,为综合评价指标约简方法选择提供支撑。Aiming at the optimization problem of seeking completeness and refinement when constructing an index system for comprehensive evaluation,the existing two categories and seven index reduction algorithms are studied,which include principal component analysis and factor analysis,minimal generalized variance method,maximum irrelevance method,grey relational analysis method,index reduction algorithm based on difference matrix,index reduction algorithm based on information entropy,index reduction algorithm based on granular computing.The principle,characteristics and application of these algorithms are studied,providing support for the selection of the comprehensive evaluation index reduction method.

关 键 词：综合评价 指标约简 算法 原理 

分 类 号：TB11[理学—数学]