基于灵敏度主成分降维和Entropy-TOPSIS 的期刊评价方法  

Journal Evaluation Method According to Sensitivity Based Principal Component Dimensionality Reduction and Entropy-TOPSIS Assessment Method

作　　者：李暄 LI Xuan(School of Business,Sichuan University,Chengdu 610065,China)

机构地区：[1]四川大学商学院,成都610065

出　　处：《价值工程》2021年第22期144-147,共4页Value Engineering

摘　　要：目的/意义:学术期刊是科研成果的重要传播途径,合理的期刊评价可以促进学科发展与学术交流。方法/过程:本文选取2019JCR中的九个指标作为基础数据。本文将指标的变异系数定义为其灵敏度并将每个指标的灵敏度作为指标的权重引入到主成分分析中。变异系数和主成分分析结合的降维法可以弥补主成分分析中的信息缺失问题。在TOPSIS评价方法中,本文引入熵权法对提取出的主成分进行权重计算。基于熵权法和TOPSIS法的评价法可以弥补TOPSIS在计算距离时的赋权问题。结果/结论:本文选取JCR经济学中的97个期刊进行实证分析。结果证实本文所提出的灵敏度主成分降维法和Entropy-TOPSIS评价法可以有效提升核心期刊排名。Purpose/significance:Academic journal is an important way to disseminate scientific research achievements.Reasonable journal evaluation can promote discipline development and academic exchanges.Method/process:Nine indicators in 2019JCR are selected as basic data in this paper.This paper defines the coefficient of variation of the indexes as their sensitivity and introduces the sensitivity of each index as the weight into the principal component analysis.The dimensionality reduction method which combines coefficient of variation and principal component analysis can make up for the lack of information in principal component analysis.In the TOPSIS evaluation method,entropy weight method is introduced to calculate the weight of the extracted principal components.This evaluation method can make up for the weight problem of TOPSIS when calculating the distance.Result/Conclusion:This paper selected 97 journals of JCR in Economics field for empirical analysis.The results confirm that the sensitivity principal component dimensionality reduction method and Entropy-TOPSIS evaluation method can effectively improve the ranking of core journals.

关 键 词：灵敏度主成分降维法 Entropy-TOPSIS评价法 影响因子 期刊评价 

分 类 号：G255.2[文化科学—图书馆学]