摘要
研究分布式存储结构下频繁闭合模式挖掘的并行化问题,针对频繁闭合模式的特点,提出了两阶段并行判断频繁模式闭合性的方法,基于串行算法FPclose和两种FP-tree的并行构造方式,分别给出了两个频繁闭合模式挖掘并行算法DP-FP和DL-FP,性能分析表明,这两个算法具有较大的并行化,较小的I/O开销与良好的负载平衡。
In this paper, we addressed the problem of parallel mining for frequent closed patterns on distributed memory multi-processors. The two phases strategy was proposed to parallel check if a frequent pattern was closed. Based on FPclose algorithm and two methods for parallel constructing FP-tree, two parallel algorithms, DP-FP and DL-FP were presented for frequent closed patterns mining. Our performance study shows the two algorithms achieved parallelism maximization, disk I/O minimization and good workload balancing.
出处
《微电子学与计算机》
CSCD
北大核心
2007年第10期161-163,共3页
Microelectronics & Computer
