Journal of South China University of Technology (Natural Science Edition) ›› 2007, Vol. 35 ›› Issue (5): 81-85.

• Computer Science & Technology • Previous Articles     Next Articles

Mining Algorithm for Multi-Level Fuzzy Association Rules Based on SOFM Network

Li Xue-jun1  Li Long-shu1  Xu Yi2   

  1. 1. Key Laboratory of Intelligent Computing & Signal Processing of the Ministry of Education , Anhui Univ. , Hefei 230039 ， Anhui , China;2. School of Computer Science and Engineering , Anhui Univ. , Hefei 230039 , Anhui , China
  • Received:2006-11-20 Online:2007-05-25 Published:2007-05-25
  • Contact: 李学俊(1976-)，男，讲师，博士生，主要从事Web挖掘、机器学习、智能软件方面的研究. E-mail:xjli@ahu.edu.cn
  • About author:李学俊(1976-)，男，讲师，博士生，主要从事Web挖掘、机器学习、智能软件方面的研究.
  • Supported by:

    国家自然科学基金资助项目( 60273043 ) ;安徽省自然科学基金资助项目(050420204) ;安徽省教育厅自然科学研究项目(KJ2007B153 )

PDF

588

Abstract:

In order to represent a complex large concept hierarchy tree , this paper proposes a more general coding scheme which applies concept hierarchy into the mining of fuzzy association rules. As it is difficult to determine the membership function subjectively , a self-organizing feature map (SOFM) network is introduced to determine the membership function of sample data. Based on the improved coding scheme and the SOFM network , fuzzy set is then introduced to design a new algorithm of mining multi-level fuzzy association rules. Experimental results show that the proposed algorithm is of high efficiency and scalability and can effectively mine multi-level fuzzy association rules that are meaningful and easily understandable.

Key words: self-organizing feature map network, concept hierarchy, fuzzy set, association rule