利用模式指导树的并行频繁项集挖据方法

Algorithm for frequent items mining based on tree and pattern guidance

关联规则挖掘的主要任务是根据对事务的统计找出项之间的关系。传统的挖掘算法要求项具有逻辑属性,并在挖掘过程中产生大量的中间项集,成为算法的瓶颈。给出一种基于关联路径树的表格数据组织形式,并采用模式指导的方式进行频繁项集挖掘,该方法不要求项具有逻辑属性,初始模式不同的项集组合迭代可以分配到不同的CPU完成,提高了算法的执行效率。该算法对美国1984年国会选举数据进行了实验,结果完全正确。

The primary task of association rule mining is to find the relationship in terms of transactions.The logic attri- butes are requested in traditional association rule mining and huge internal item sets are generated.This problem is one of the bottlenecks of the mining algorithms.This paper presents a method to mine frequent items sets based on association path tree under the guidance of patterns and without logical data.The iterate program of finding frequent item sets can run on multiple CPUs for improving the algorithm performance.The approach is tested by House-Votes-84 and the correct results are obtained.