基于因子分析与TOPSIS法在学术期刊评价中的改进研究  

Improvement of Academic Journal Evaluation Based on Factor Analysis and TOPSIS Method

作　　者：熊国经[1] 熊玲玲[1] 陈小山[1] 

机构地区：[1]南昌大学经济管理学院,南昌330031

出　　处：《情报杂志》2016年第7期196-200,共5页Journal of Intelligence

基　　金：国家社会科学基金项目"非参数估计的影响因素优化匹配方法与应用研究"(编号:14BTJ016)研究成果之一

摘　　要：[目的/意义]用多属性评价方法评价学术期刊会遇到指标间多重共线性以及指标权重不确定的而影响到评价结果的难题;基于此,给出了一种综合评价改进方法,对期刊评价的权重合理分配以及消除期刊评价中主观因素的影响具有一定的积极意义。[方法/过程]其本质思路是通过因子分析法消除指标间的相关重叠性,确定期刊学术影响力的主因子,以熵值法的差异系数对主因子的方差贡献率和因子得分系数进行调整,从而确定各指标的权重、构造加权规范矩阵,并运用TOPSIS法比较备选样本期刊与最优解的相对贴近度。[结果/结论]通过实证研究,发现该方法既可以有效消除指标间的多重共线性,也可以合理为指标赋予权重,得到更为客观的评价结果。[Purpose/Significance] Multi-attribute evaluation method for periodical evaluation index will face two big problems of multi-collinearity and index weight uncertainty. This paper puts forward a new method to improve the evaluation. [ Method/Process] The essen-tial idea is to use factor analysis method to eliminate overlaps between indicators,determine the main factor of the journal academic influ-ence,and to adjust the variance contribution ratio of the main factors and the factor score coefficient by the difference coefficient of the method of entropy value,thus to determine the weight of each index and construct the weighted matrix,the TOPSIS method is used to com-pare the relative closeness of the alternative sample journals and the optimal solution. [ Result/Conclusion] An empirical study shows that the method can effectively eliminate the multicollinearity between the various indicators,and the weights allocated are more reasonable.

关 键 词：期刊评价 多属性评价 因子分析 熵值法 

分 类 号：G203[文化科学—传播学]