华南理工大学学报(自然科学版) ›› 2006, Vol. 34 ›› Issue (9): 50-55.

• 计算机科学与技术 • 上一篇    下一篇

基于粗糙集属性变分区的属性约简

邓九英1 毛宗源1 徐宁2   

  1. 1.华南理工大学 自动化科学与工程学院,广东 广州 510640;2.广东教育学院 计算机科学系,广东 广州 510303
  • 收稿日期:2005-12-05 出版日期:2006-09-25 发布日期:2006-09-25
  • 通信作者: 邓九英(1962-),女,在职博士生,广东教育学院副教授,主要从事智能控制、数据挖掘与仿真技术方面的研究 E-mail:djyl111@126.com
  • 作者简介:邓九英(1962-),女,在职博士生,广东教育学院副教授,主要从事智能控制、数据挖掘与仿真技术方面的研究

Attribute Reduction Using Attribute Variable Partition of Rough Set

Deng Jiu-ying1  Mao Zong-yuan1  Xu Ning2   

  1. 1.School of Automation Science and Engineering,South China Univ.of Tech.,Guangzhou 510640,Guangdong,China2. Dept. of Computer Science.Guangdong Institution of Education,Guangzhou 510303,Guangdong.China
  • Received:2005-12-05 Online:2006-09-25 Published:2006-09-25
  • Contact: 邓九英(1962-),女,在职博士生,广东教育学院副教授,主要从事智能控制、数据挖掘与仿真技术方面的研究 E-mail:djyl111@126.com
  • About author:邓九英(1962-),女,在职博士生,广东教育学院副教授,主要从事智能控制、数据挖掘与仿真技术方面的研究

摘要: 应用粗糙集的方法,分析决策系统中不同的属性分类方法,以及不同分类方法引起的属性重要性与属性相对约简极小子集的变化情况,寻求属性分类方法与属性约简结果相互影响的内在因素,给出高效的属性分类方法和合理确定约简子集的策略,生成策略对应软件的实现算法,并运用软件实现算法来选取相对约简子集.试验结果显示了该策略及算法的有效性.

关键词: 属性值分区, 相对约简, 极小子集, 属性重要性, 核属性

Abstract:

By means of the methodology of rough set,different attribute classification methods for decision systems are first analyzed in this paper,and the attribute significance as well as the extremely small subset of the attribute relative reduction caused by a classification variety is discussed.Next, the intrinsic factor to affect the interaction between the attribute classification method and the attribute reduction results is revealed. Then.a scheme for performing effective attribute classification and a strategy for determining reasonable attribute reduction subsets are in-troduced,and an algorithm is presented to implement the corresponding software according to the strategy . By using the presented algorithm,a relative reduction subset is finally selected. Experimental results verify the effectiveness of the proposed strategy and algorithm.

Key words: attribute-value partition, relative reduction, extremely small subset, attribute significance, core attribute