主管:中国科学院
主办:中国优选法统筹法与经济数学研究会
   中国科学院科技战略咨询研究院

中国管理科学 ›› 2020, Vol. 28 ›› Issue (1): 162-169.doi: 10.16381/j.cnki.issn1003-207x.2020.01.014

• 论文 • 上一篇    下一篇

动态综合评价中的数据预处理方法研究

徐林明1, 李美娟2   

  1. 1. 福建工程学院管理学院, 福建 福州 350118;
    2. 福州大学经济与管理学院, 福建 福州 350108
  • 收稿日期:2019-01-09 修回日期:2019-07-10 出版日期:2020-01-20 发布日期:2020-01-19
  • 通讯作者: 李美娟(1979-),女(汉族),福建古田人,福州大学经济与管理学院,教授,博士,博士生导师,研究方向:综合评价方法及其应用,E-mail:758300@qq.com. E-mail:758300@qq.com
  • 基金资助:
    国家自然科学基金资助项目(71872047,71801050);福建省社会科学规划项目(FJ2019B086,FJ2018B029,FJ2017C017);福建省软科学项目(2019R0083,2019R0013);科技部创新方法工作专项(2017IM010200);福建工程学院科研启动项目资助课题(GY-S18002);福建省高校领军人才资助

A Data Preprocessing Method in Dynamic Comprehensive Evaluation

XU Lin-ming1, LI Mei-juan2   

  1. 1. School of Management, Fujian University of Technology, Fuzhou 350118, China;
    2. School of Economics&Management, Fuzhou University, Fuzhou 350108, China
  • Received:2019-01-09 Revised:2019-07-10 Online:2020-01-20 Published:2020-01-19

摘要: 数据预处理是综合评价的基础,通常包括指标类型一致化处理和无量纲化处理,指标值预处理结果会对后续的评价结论产生较大的影响。针对利用现有静态数据预处理方法对动态综合评价中的三维数据进行处理时会消除数据中隐含增量信息的问题,提出一种动态综合评价的数据预处理方法来解决该问题。通过介绍现有常用评价指标预处理方法,并分析这些预处理方法的特点,扬长避短,提出全局改进归一化方法,作为动态综合评价中指标一致化及无量纲化方法,对指标值进行预处理。最后通过算例比较分析,说明各方法特点,验证全局改进归一化方法是一种有效的动态综合评价数据预处理方法。

关键词: 动态综合评价, 无量纲化方法, 一致化, 预处理方法, 评价指标

Abstract: Data preprocessing is the basis of comprehensive evaluation and it usually includes uniformization of indicator types and nondimensionalization. Since the preprocessing result of indicator values can greatly influence the subsequent evaluation conclusion, whether its processes are appropriate or not has a direct impact on the reasonability of the final conclusion. Focusing on the problem that implicit incremental information will be eliminated when the three-dimensional data in dynamic comprehensive evaluation are processed by existing static data preprocessing methods, a data preprocessing method used in dynamic comprehensive evaluation is put forward to solve this problem.
By introducing and analyzing the characteristics of six preprocessing methods of evaluation indexes, including range conversion method, linear proportional transformation, normalized processing, vector specification, standard sample transformation, efficacy coefficient method, a global improved normalization method (GIN) is proposed as a method for uniformization and nondimensionalization of indicators in dynamic comprehensive evaluation and preprocessing of indicator values.
The idea of GIN method is as follows:it comprehensively considers and uniformly preprocess each index value in every moment, so as to retain the incremental information implied in the original data. The subtraction consistency method is employed for index type consistency processing. After processing, the larger the index value means the better. The difference between the original index values is retained after the dimensionless processing, and the pre-processed index values are comparable. Finally, through comparison and analysis of calculating examples, it illustrates the features of various methods and verifies that the global improved normalization method is an effective data preprocessing method in dynamic comprehensive evaluation.
The GIN method can retain the incremental information of the original data, and can be used to process the three-dimensional data in the dynamic comprehensive evaluation, which is a useful supplement to the comprehensive evaluation method and its application research.

Key words: dynamic comprehensive evaluation, nondimensionalization, uniformization, preprocessing method, evaluation indicator

中图分类号: