Abstract: Firstly, a method of determining the membership function of sample data based on the self-organizing feature maps network(SOFM) was presented. Based on this method, according to the concept of similarity, with the calculating formula of similarity degree and the Apriori algorithm, a new algorithm of mining fuzzy similar association rules was proposed. Compared with existing similar algorithms, the new algorithm overcomes the defect of determining membership functions subjectively, especially when the data structure is more complex it is very difficult. Meanwhile, this algorithm reduces the redundant rules mined.

Key words: data mining, fuzzy association rule, SOFM network, membership function

摘要： 提出了一种基于自组织特征映射(SOFM)网络的自动确定样本数据隶属度函数的新方法,并在此基础上根据相似性的概念,给出了相似度的计算公式,结合Apriori算法,提出了一种挖掘模糊相似关联规则的新算法。与现有的同类算法相比,现有的方法均需人为地确定隶属度函数,带有一定的主观性,尤其当数据结构较复杂时,隶属度函数难以确定;该算法克服了这一缺点,同时减少了冗余规则。

关键词:  , 数据挖掘, 模糊关联规则, SOFM网络, 隶属度函数