TOPSIS在期刊评价中的应用及在高次幂下的推广  

Application of TOPSIS in Journal Evaluation and its Expansion at Higher Power

作　　者：俞立平[1] 潘云涛[2] 武夷山[2] 

机构地区：[1]宁波大学 [2]中国科学技术信息研究所

出　　处：《统计研究》2012年第12期96-101,共6页Statistical Research

基　　金：国家社科基金后期资助项目(10FTQ003):科技评价中多属性评价方法基本理论研究

摘　　要：标准TOPSIS评价是在2次幂下进行的,本文以学术期刊评价为例,将标准TOPSIS推广到其他幂次并分别进行评价,同时比较了不同评价结果的一致性、区分度、打分倾向、数据分布特点等。研究表明,标准TOPSIS是一种对较好期刊区分度较好、对弱势期刊区分度较差,对较好期刊打分倾向偏高、对弱势期刊打分倾向偏低的评价方法。应该根据不同的评价目的选择不同幂次的TOPSIS评价,本文的研究方法也可用于其他多属性评价方法的选取。Standard evaluation of TOPSIS is used at power of 2. This paper expands the standard TOPSIS to higher power and makes evaluation using academic journal as an example, and then compares consistency, discrimination, grade tendency, data distribution of different evaluation results. The results show standard TOPSIS method has good discrimination and high grade tendency for better journals, and bad discrimination and low grade tendency for worse journals. The different power of TOPSIS method should be chosen according to different evaluation objectives. The method of this paper is also used to choose other multiple attribute evaluation methods.

关 键 词：TOPSIS 幂次 学术期刊 区分度 打分倾向 

分 类 号：C812[社会学—统计学]