频繁闭合项目集的并行挖掘算法研究

频繁项目集挖掘因其在数据挖掘领域中的基础地位和广泛应用备受学术界和产业界的关注，用挖掘频繁闭合项目集代替挖掘频繁项目集是近年来提出的一个重要策略。不同于以往提出的挖掘所有频繁项目集的并行算法，本文针对频繁闭合项目集的特性及并行挖掘的特点，给出了共享存储器模型上(Shared Memory)基于频繁模式树(FP-tree)的挖掘频繁闭合项目集的并行算法(FCIPM)思想，提出了频繁闭合项目集直接判断法，性能分析表明所提技术对算法的性能提高起到了关键作用。

A Parallel Algorithm for Mining Frequent Closed Itemsets

Frequent itemsets mining has been studied extensively in data mining research. Because the huge number of frequent itemsets is found,an interesting alternative has been proposed recently .instead of mining the complete set of frequent itemsets .only find frequent closed itemsets .which has the same power as mining the complete set of frequent itemsets,but it will substantially reduce the number of frequent itemsets. In this paper,we propose an efficient parallel algorithm,FCIPM(Frequent Closed Itemsets Parallel Mining), based on shared memory for mining frequent closed itemsets, with the development of three techniques: (DA FP-tree based bottom-up no tree-projection method, (2)itemsets pruning, (3)frequent closed itemsets checking directly. Our performance study shows the advantage of these techniques and that the FCIPM may has good performance in terms of runtime,memory usage and scalability.