科技评价中兼顾均值与区分度的标准化方法研究——动态最小均值逼近标准化方法

Standardization Method Considering Mean and Discrimination in the Science and Technology Evaluation:Standardization Method of Dynamic Minimum Mean Approximation

[目的/意义]针对科技评价中数据标准化后评价指标均值不相等会导致自然权重以及评价指标区分度较低问题,该文将其放在一个系统框架里进行研究。[方法/过程]该文提出了一种新的动态最小均值逼近标准化方法,并以中国知网《中国学术期刊影响因子年报(人文社会科学)》中图书馆情报与文献学20种CSSCI期刊为例进行了实证研究。[结果/结论]动态最小均值逼近标准化方法能够保证评价指标再次标准化后均值相等,并且能够增加标准化指标的区分度,同时也能增加评价结果的区分度;动态最小均值逼近标准化方法可与动态最大均值逼近标准化方法结合使用;采用动态最小均值逼近标准化方法会影响评价结果;要特别注意专家人为打分值的均值与区分度标准化问题;指标均值与区分度兼顾标准化问题有待进一步探索。

[Purpose/Significance]Aiming at the problem that unequal mean of evaluation indexes after data standardization in academic evaluation will lead to natural weight and low discrimination of evaluation indexes,this paper conducts a research of them in a systematic framework.[Method/Process]The paper proposes a new standardization method of dynamic minimum mean approximation,and takes 20 CSSCI journals of library intelligence and philology in impact factor annual report of Chinese academic journals(humanities and social sciences)of CNKI as an example.[Result/Conclusion]The results show that the standardization method of dynamic minimum mean approximation can ensure the mean equality after the re-standardization of evaluation indexes,and can increase the discrimination of standardized indexes and evaluation results;The standardization method of dynamic minimum mean approximation can be combined with the standardization method of dynamic maximum mean approximation;The evaluation results will be influenced by the standardization method of dynamic minimum mean approximation;Special attention should be paid to the standardization of the mean and discrimination of the artificial scores of experts;Taking into account at the same time the standardization with the consideration of the mean and discrimination of indexes needs to be further explored.