基于K-means算法的Web访问用户关联规则挖掘算法

Web Users Association Rule Mining Algorithm Based on K-means Algorithm

针对传统关联规则算法中事务扫描的重复性以及最小支持度设定的不确定性,导致关联规则挖掘算法扫描事务数据库运行效率低下的问题,提出一种基于K-means的Web访问用户关联规则挖掘算法,该算法利用K-means算法聚类的效果,将Web访问用户数据集聚类为不同的小数据集,采用不同的最小支持度,分别对Web访问用户聚类小数据集进行关联规则挖掘。分析和实验结果证明,该算法可有效提高传统关联规则挖掘算法的效率,同时也可有效避免传统关联算法中扫描中的重复性。

Using traditional association rule algorithm, the repetitiveness of scanning things and the uncertainty of minimum support lead to the low efficiency of scanning the transaction database. The paper proposes a web users association rule mining algorithm which uses the clustering effect of K-means algorithm to cluster the web users＇ data into different small data sets, and uses different minimum support to carry out association rule mining on the small data sets. Analysis and experiment resuhs show that the algorithm can effec- tively improve the efficiency and avoid the scanning repetitiveness of traditional association rule mining algorithm.