基于兴趣度的关联规则挖掘算法  

Association Rules Mining Algorithm Based on Interest Measure

作　　者：马建庆[1] 钟亦平[1] 张世永[1] 

机构地区：[1]复旦大学计算机与信息技术系,上海200433

出　　处：《计算机工程》2006年第17期121-122,149,共3页Computer Engineering

摘　　要：分析了Apriori核心算法,举例说明了其设计思想上的不足,并重新定义了关联规则形式和引进了兴趣度的概念。主要定义了合理的兴趣度,即基于可信度和支持度方差的兴趣度InterestR=[(CR?SRH)/4]*(CR+SRH),并因此而设计了基于此兴趣度定义的关联规则挖掘算法,并对算法做了适当的分析。Based on the analysis of Apriori algorithm, this paper points to the defects of its design thought. After adopting the concept of interest measure and new definition of association rules, it defines the more reasonable interest measure that bases on the difference between the square of confidence and the square of support, also InterestR = [（CR -SRH）/4]^＊（ CR ＋SRH）. Therefore, with the new definition of interest and association rules, it designs a new association rules mining algorithm and also gives some analysis on it.

关 键 词：兴趣度 关联规则挖掘算法 关联规则 可信度 支持度 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]