摘要： 关联规则是数据挖掘中重要的课题之一。传统的由频繁项目集产生关联规则的 方法由于要考虑频繁项目集的每一个非空真子集，当频繁项目集的长度较长时代价较大。文 中提出用K项子频繁项目集导出关联规则以后得到的有用信息以指导K+1项父频繁项目集产生 强关联规则，由于利用了前面已有的有用信息，当频繁模式长度很长时，或者可信度较高时 ，效率尤其显著。

关键词: 频繁项目集, 强关联规则, 子项目集, 父项目集

Abstract: One of the most important data mining problems is mining association rules. Traditional method for generating association rules from frequent itemse ts is expensive because it have to consider each real sub set of the frequent it emset which is not null. The paper considers how to use the obtained useful rule information of frequent sub K-itemsets to generate strong association rules for the frequent parent (K+1)-itemsets, the performance will be better when the fre quent itemset is very long or the confidence is very high, because it uses the o btained useful information.

中图分类号: 

• TP311.13