上海交通大学学报(自然版) ›› 2017, Vol. 51 ›› Issue (2): 180-.

• 无线电电子学、电信技术 • 上一篇    下一篇

一种效能评估中缺失数据的填充方法

王军1,2,李建勋1,韩山3,王兴2   

  1. 1. 上海交通大学 自动化系, 上海 200240; 2. 中国洛阳电子装备试验中心, 河南 洛阳 471000;
    3. 中国航空无线电电子研究所, 上海 200233
  • 出版日期:2017-02-28 发布日期:2017-02-28
  • 基金资助:

    国家重点基础研究发展规划(973)项目(613319032015JD07),国家自然科学基金资助项目(61673265),电子信息系统复杂电磁环境效应国家重点实验室开放课题(CEMEE2013K0301A)资助

A Missing Data Filling Method for Effectiveness Evaluation System

WANG Jun1, 2,LI Jianxun1,HAN Shan3,WANG Xing2   

  1. 1. Department of Automation, Shanghai Jiao Tong University, Shanghai 200240, China;
    2. Luoyang Electronic Equipment Test Center of China, Luoyang 471000, Henan, China;
    3. China Aeronautical Radio Electronics Research Institute, Shanghai 200233, China
  • Online:2017-02-28 Published:2017-02-28

摘要:

 针对评估数据缺失的问题,提出了一种新的缺失数据填充方法.从信息熵的角度对评估系统进行描述,并提出一种建立在区间值信息系统上的加权广义信息熵,对这种信息熵的非负性、单调性和极值性进行证明.在此基础上,通过研究评估系统数据缺失前后信息熵的变化情况,建立一种基于加权广义信息熵的缺失数据填充方法,并从知识粒度的角度论述加权广义信息熵的机制.通过实例分析与现有的部分算法进行比较,进一步证明文中算法对评估系统缺失数据填充的有效性.

关键词: 效能评估, 数据缺失, 数据填充, 加权广义信息熵

Abstract:

To solve the data missing problem in effectiveness evaluation of weapon systems, a novel method was proposed to fill the missing data. The effectiveness evaluation system was described from the perspective of information entropy while the so called weighted generalized information entropy was found on the intervalvalued information system. Besides, the nonnegativity, monotony and extremality of the new information entropy were discussed. Based on the new theory, and according to the variation of the information entropy before and after data missing of the effectiveness evaluation system, a novel method for filling the missing data based on weighted generalized information entropy was proposed, and the theory of weighted generalized information entropy was explained from the perspective of knowledge granularity. The proposed algorithm performs better in  comparison with other algorithms, which verifies that the algorithm is valid in filling missing data of effectiveness evaluation system in consequence.

Key words: effectiveness evaluation, data missing, data filling, weighted generalized information entropy

中图分类号: