摘要
提出了一种新颖的关联规则挖掘算法QAIS,与经典两阶段式关联规则挖掘算法不同的是,它只需扫描一遍事务数据库,不需要生成候选集,并且可以方便的应用在增量式关联规则挖掘算法中,该算法经合成数据验证是有效的.同时针对关联规则生成过程中出现大量冗余规则的问题,还讨论了冗余关联规则去除的问题.
A novel association rules mining algorithm QAIS is proposed in this paper. It is different from classical two phrase mining algorithm. QAIS algorithm requires to pass over the transaction database only one time. It can mine association rules more efficiently by not generating candidate item sets and it can be used conveniently for incremental algorithm for maintaining association rules. The performance of the algorithm is verified on the basis of synthetic data. At the same time, aiming at emergence of a great deal redundancy rules while generating association rules, a method on reducing the redundancy of frequent item sets is discussed as well.
出处
《佳木斯大学学报（自然科学版）》
CAS
2004年第2期151-156,177,共7页
Journal of Jiamusi University：Natural Science Edition
关键词
关联规则
数据挖掘
频繁项集
冗余规则
association rule
data mining
frequent item set
redundancy rule
