科技通报
科技通報
과기통보
BULLETIN OF SCIENCE AND TECHNOLOGY
2014年
4期
131-133
,共3页
关联规则挖掘%并行算法%FP-forest
關聯規則挖掘%併行算法%FP-forest
관련규칙알굴%병행산법%FP-forest
association rules mining%parallel algorithm%FP-forest
在研究传统挖掘频繁项集并行算法的基础上,提出一种以FP-forest为基础的高效能并行挖掘算法FPPMA(FP-forest based high-effect Parallel Mining Algorithm)。此算法中,各运算节点只需要和核心节点之间传递少量信息,而无需和其他运算节点通信,减少了通信费用。此外,运算节点不需要同步和交换数据就可独立挖掘出全局频繁项集。
在研究傳統挖掘頻繁項集併行算法的基礎上,提齣一種以FP-forest為基礎的高效能併行挖掘算法FPPMA(FP-forest based high-effect Parallel Mining Algorithm)。此算法中,各運算節點隻需要和覈心節點之間傳遞少量信息,而無需和其他運算節點通信,減少瞭通信費用。此外,運算節點不需要同步和交換數據就可獨立挖掘齣全跼頻繁項集。
재연구전통알굴빈번항집병행산법적기출상,제출일충이FP-forest위기출적고효능병행알굴산법FPPMA(FP-forest based high-effect Parallel Mining Algorithm)。차산법중,각운산절점지수요화핵심절점지간전체소량신식,이무수화기타운산절점통신,감소료통신비용。차외,운산절점불수요동보화교환수거취가독립알굴출전국빈번항집。
Aiming at solving the problems in traditional parallel methods, a high-effect parallel algorithm based on FP-for-est is proposed. In this algorithm, each computing node only needs to pass few information with the kernel node and connec-tion with other computing node don’t needed,so the communication is reduced . In addition, computing nodes can mine global frequent itemsets independently without synchronization and data exchanging.