计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (34): 145-148.DOI: 10.3778/j.issn.1002-8331.2008.34.045

• 数据库、信号与信息处理 • 上一篇    下一篇

扩展正区域的属性约简方法

王俊祥1,胡 峰2   

  1. 1.重庆文理学院 基础学院,重庆 402168
    2.重庆邮电学院 计算机科学与技术研究所,重庆 400065
  • 收稿日期:2007-12-18 修回日期:2008-03-21 出版日期:2008-12-01 发布日期:2008-12-01
  • 通讯作者: 王俊祥

Reduction algorithm under extended positive region

WANG Jun-xiang1,HU Feng2   

  1. 1.College of Fundamental Science,Chongqing University of Arts and Sciences,Chongqing 402168,China
    2.Institute of Computer Science and Technology,Chongqing University of Posts and Telecommunications,Chongqing 400065,China
  • Received:2007-12-18 Revised:2008-03-21 Online:2008-12-01 Published:2008-12-01
  • Contact: WANG Jun-xiang

摘要: 扩展了Rough集正区域和边界的定义,在得到信息系统最大正区域的前提下,给出了认知正区域、认知属性核和认知属性约简的定义,并给出了从经典属性约简到认知属性约简转换的高效算法。此外,在认知正区域的定义下,由于决策表的不相容性,在变精度模型下实现属性约简的增量处理是相当困难的,结合提出的高效算法,解决了这一问题。最后,仿真实验说明了算法的有效性。

关键词: 粗集, 决策表, 认知核属性, 认知属性约简, 增量式

Abstract: The definitions of cognitive positive region,cognitive attribute core and cognitive attribute reduction are proposed,by extending the notions of positive region and boundary region in classic rough set theory.Furthermore,a high efficient algorithm of transforming classic attribute reduction to cognitive attribute reduction is presented.It is difficult to implement incremental attribute reduction in VPRS model due to the non-tolerance of decision table,which is solved by combining high efficient transforming algorithm.Experimental results illuminate the validity and effectiveness of algorithms.

Key words: rough set, decision table, cognitive core attribute, cognitive attribute reduction, incremental