• 论文 •    

基于粗糙集的不完备信息系统空值估算方法

李聪,梁昌勇,杨善林   

  1. 合肥工业大学 管理学院,安徽合肥230009
  • 出版日期:2009-03-15 发布日期:2009-03-25

Null values estimation method based on rough set for incomplete information systems

LI Cong, LIANG Chang-yong, YANG Shan-lin   

  1. School of Management, Hefei University of Technology, Hefei 230009, China
  • Online:2009-03-15 Published:2009-03-25

摘要: 针对基于相似关系模型的空值估算方法存在的不足,从群体决策的角度提出了一种改进方法。通过定义属性的向域和对象的简单多数规则率,将其结合用于空值估算,从而无需原方法的投票策略就能实现不完备信息系统的完备化,并消除了原方法可能出现的对象无相容类问题和投票策略潜在的多众数、无众数问题,得到的估算结果在覆盖率和准确率两方面均优于原方法。通过算例及应用示例表明了本文方法的有效性。

关键词: 粗糙集理论, 不完备信息系统, 空值, 群体决策

Abstract: To deal with shortcomings in null value estimation method based on similarity relationship model, an improved method was proposed from the perspective of group decision. The direction-area of attributes and simple-majority-rule ratio of objects were defined, and they were combined with null values estimation. Thus completion of incomplete information system was realized without using the voting tactic of original method, moreover the problems such as “objects have no tolerance class” and latent “multimode” or “no mode” of voting tactic were eliminated, and the estimation results of proposed method were better than those of original method on coverage and accuracy. The results of a numerical example and an instance demonstrated the effectiveness of proposed method.

Key words: rough set theory, incomplete information systems, null values, group decision

中图分类号: