计算机工程 ›› 2012, Vol. 38 ›› Issue (12): 139-142.doi: 10.3969/j.issn.1000-3428.2012.12.041

• 人工智能及识别技术 • 上一篇    下一篇

区间值信息系统的启发式属性约简

梁春华 1,2,曲开社 2,张海云 2   

  1. (1. 太原理工大学财经学院,太原 030024;2. 山西大学计算机与信息技术学院,太原 030006)
  • 收稿日期:2011-08-01 出版日期:2012-06-20 发布日期:2012-06-20
  • 作者简介:梁春华(1978-),女,讲师,主研方向:粗糙集理论;曲开社,教授;张海云,讲师、博士研究生
  • 基金项目:
    国家自然科学基金资助项目(61070100, 70971080)

Heuristics Attribute Reduction in Interval-valued Information System

LIANG Chun-hua 1,2, QU Kai-she 2, ZHANG Hai-yun 2   

  1. (1. Institute of Finance and Economics, Taiyuan University of Technology, Taiyuan 030024, China; 2. School of Computer & Information Technology, Shanxi University, Taiyuan 030006, China)
  • Received:2011-08-01 Online:2012-06-20 Published:2012-06-20

摘要: 针对区间值信息系统中的属性约简问题,引入α-极大相容类的概念,定义区间值信息系统的属性间依赖度和信息熵,提出相应属性内(外)重要度的度量方法,给出一种统一的启发式属性约简算法,通过实验验证该算法的有效性,并分析不同相似水平α对约简结果的影响。

关键词: 区间值信息系统, 近似约简, α-极大相容类, 属性重要度, 信息熵, 属性依赖度

Abstract: Interval-valued attribute reduction is the core problem in interval-valued information systems research. By introducing a kind of α maximal tolerance class, this paper defines the attribute dependency and conditional entropy among attributes in interval-valued information systems, and proposes two types of attribute importance measurement, which has inner attribute importance and outer attribute importance. A uniform heuristic attribute reduction algorithm is given. The validity of the algorithm is illustrated by some experiments and the effect of different level of similarity α for reduction is investigated as well.

Key words: interval-valued information system, approximation reduction, α-maximal consistent class, attribute importance, information entropy, attribute dependency

中图分类号: