摘要： 对关联规则和约束关联规则的算法进行了研究和分析，基于候选集的约束算法需要反复扫描数据库，并产生大量的候选集，在挖掘低支持度、长模式的规则时效率低下。针对算法的缺陷，该文提出了一种Conf-H-Mine算法，采用Conf-H-Struct结构存储事务集合，不产生候选集，优化了关联规则的挖掘。实验结果证明了该算法的有效性。

关键词: 数据挖掘, 关联规则, 项目约束挖掘

Abstract: After analyzing and studying constraint-based data mining algorithms, there are great flaws in constraint algorithm based on candidate sets, the algorithm needs multiple scanning, produces lots of candidate sets, and has low efficiency when mining low support threshold, long rules. This paper introduces a new algorithm Conf-H-Mine which produces no candidate sets, and optimizes association rules mining. Experimental results show the algorithm is effective.

Key words: data mining, association rule, constraint-based mining

中图分类号: 

• TP301.6