基于属性分组的高效挖掘关联规则算法

挖掘频繁项集在数据挖掘中有着重要的作用。目前,关于频繁项集的挖掘问题已经提出了一些算法,虽然实现了一次扫描数据库即可以发现所有的频繁项集,但是当属性数目很多时,算法的执行效率下降很快。论文首次提出了利用属性分组作为挖掘关联规则的工具,给出了基于属性分组的频繁项集挖掘算法,用矩阵来存储数据库属性间的信息并提取频繁项集,而且不产生候选项集。经实验验证该算法是快速有效的。

An Efficient Algorithm of Mining Association Rules Based on Dividing Attributes

Mining frequent itemsets plays an important role in data mining and researchers present some algorithms for mining frequent itemsets so far.Although the primary algorithms of finding frequent itemsets realizes the consequence that scanning the database only once,but when it comes to more attributes in database,it becomes more difficult to be resolved.This article puts forward a new tool of mining association rules using dividing attributes for the first time in this area and presents the algorithm based on dividing attributes.The algorithm generates all the frequent itemsets using a matrix for storing information between attributes in database,without generating candidate itemsets.Experimental results show its effectiveness and efficiency.