多支持度关联规则在网络使用挖掘中的应用

The Application of Association Rules Based on Multiple Minimum Supports for Web Usage Mining

关联规则方法是网络挖掘中一种备受人们关注的技术,但大多数方法在整体上采用统一的最少支持度,使得其在实际应用中受到限制。论文从这一问题出发,拟定了一个事务集,论述多支持度的关联规则方法。并把该方法应用到网络使用挖掘中,根据页面的出现频率确定其最少支持度阀值,发现并论证了不平衡数据项之间的有效规则。从执行时间和搜索空间上验证了算法的有效性。

Association rule mining as a technique of web usage mining has attracted many people.Most of these methods adopt one uniform minimum support on the whole data base.However,this is seldom the case in applications.This paper addresses the problem and designs a data base to discuss the approach of mining association rules with multiple minimum supports.Moreover we apply this method into web usage mining and find out effective rules between frequent items and infrequent items whose minimum supports are specified by their concurrent frequent in data base.Experiment results show the validity of the technique on execution time and search space.