摘要
关联规则是数据挖掘中一个非常重要的任务,有许多针对于关联规则的挖掘算法,然而需要提高算法的有效性来处理现实世界中的数据集。基于聚类的关联规则挖掘算法法通过扫描数据库创建聚类表,将收集的事务记录放入聚类表中,通过局部聚类表的约束来产生频繁项集,不仅可以剪枝候选项集,降低数据扫描的时间,而且确保挖掘结果集的正确性。实验结果表明,基于聚类的关联规则挖掘算法比Apriori算法有更高的执行效率。
Association rule is one of the significant tasks in data mining, and there are many mining algorithms about association rule. But in order to deal with realistic data sets, we need to improve efficiency of the algorithm. The association rule algorithm based on cluster scans data bank, creates clustering table and records the collected affairs in the clustering table. Through re- strain of partial clustering table we can obtain frequent item set, which cuts candidate item set, lowers data scan time and en- sures the accuracy of mining result set. According to test results, association rule algorithm based on cluster is much more effi- cient than apriori algorithm.
出处
《太原大学学报》
2011年第3期113-116,共4页
Journal of Taiyuan University
关键词
数据挖掘
关联规则
聚类
data mining
association rule
cluster
