一种高效的并行频繁集挖掘算法

Efficient Parallel Frequent Itemsets Mining Algorithm

针对Apriori算法在挖掘超大规模数据集时存在的效率低下问题,在数据集分块和事务数据库布尔化映射基础上,提出一种直接利用布尔矩阵向量运算挖掘频繁集的并行频繁集挖掘算法(PFIM)。仿真实验分析表明,PFIM算法比Apriori算法的挖掘时间缩短了近90%,该方法可用于挖掘超大规模数据库,具有良好的并行性和可伸缩性。

Aiming at inefficient problem of Apriori algorithm when mining very large database, this paper presents an efficient Parallel Frequent Itemset Mining algorithm（PFIM） based on database dividing and computing of Boolean matrix mapped from original database. Experimental results show that PFIM algorithm cuts down ninety percent mining time of Apriori, so it is suitable for mining very large size database and it has good characteristics of parallel and expandable.