学术期刊评价中熵权法的系统误差研究
Systematic Error Research of Entropy Weight Method in Academic Journal Evaluation
俞立平 1张连桥2
作者信息
- 1. 浙江工商大学统计与数学学院,浙江杭州 310018;浙江工商大学统计数据工程技术与应用协同创新中心,浙江杭州 310018
- 2. 广州大学人文学院,广东广州 510006
- 折叠
摘要
[目的/意义]熵权法作为重要的学术期刊评价方法,其系统误差问题值得关注.[方法/过程]本文将误差理论引入到学术期刊多属性评价中,分析了熵权法系统误差的产生机制,并提出了解决思路.[结果/结论]学术期刊熵权评价系统误差问题必须引起足够的重视,广泛存在于评价指标选取、权重赋权、数据标准化、结果公布等诸多环节.降低熵权法系统误差必须全方位介入,除了合理选择评价指标、进行权重合理性分析外,重点应该注意标准化方法问题,采用sigmoid函数标准化是一种行之有效的手段.熵权法作为评价方法的局部对其他评价方法的系统误差会产生影响,总体上影响相对较小但仍然应该重视.[创新/局限]熵权法系统误差的分析范式可广泛应用于其他学术评价方法中.
Abstract
[Purpose/significance]The entropy weight method is an important evaluation method for academic journals,and its system-atic error is worthy of attention.[Method/process]This paper introduces the error theory into the multi-attribute evaluation of aca-demic journals,analyzes the generation mechanism of the system error of the entropy weight method,and puts forward a solution.[Re-sult/conclusion]The research shows that the error problem of the entropy weight evaluation system of academic journals must be paid enough attention to,which widely exists in many links such as evaluation index selection,weight empowerment,data standardization,and result publication.To reduce the system error of the entropy weight method,we must intervene in all aspects.In addition to choos-ing the evaluation index reasonably and analyzing the rationality of the weight,we should pay attention to the standardization method.Using the sigmoid function to standardize is an effective means.As a part of the evaluation method,the entropy weight method will have an impact on the systematic errors of other evaluation methods.The overall impact is relatively small,but it should still be taken seriously.[Innovation/limitation]The analytical paradigm of the systematic error of the entropy weight method can be widely used in other academic evaluation methods.
关键词
熵权/系统误差/sigmoid/指标选取/自然权重/标准化方法Key words
entropy weight/systematic error/sigmoid/index selection/natural weight/standardization method引用本文复制引用
基金项目
国家社会科学基金(21FTQB016)
浙江省自然科学基金重点项目(Z21G030004)
出版年
2024
NETL
NSTL
万方数据