并行多层关联规则的挖掘算法研究  

Research on parallel mining algorithm of multilevel association rules in cluster

作　　者：王华秋[1] 王越[1] 曹长修[1] 

机构地区：[1]重庆大学自动化学院

出　　处：《计算机应用》2006年第2期373-375,共3页journal of Computer Applications

基　　金：重庆市教委基础研究项目(020612);重庆市科委攻关项目(20020828)

摘　　要：在分析了单数据库多层关联规则算法SMAM算法的基础上,为了提高现有算法的效率,改善内存的使用率,提出了两种并行多层关联规则算法PMAM-L和PMAM-LG。在无共享的工作站机群上进行性能测试,证明PMAM算法是有效的。In practical applications, multilevel concept of association rules mining is often encountered. To overcome problems existing in efficiency of candidate sets pruning and execution time of current algorithm, based on designing and analyzing the algorithm of mining association rules in single database, two parallel multilevel association rules algorithms, PMAM-L and PMAM-LG were proposed. Finally the results of test on share-nothing cluster of workstation show that the proposed algorithms are efficient.

关 键 词：多层关联规则 并行算法 候选集 剪枝 机群 

分 类 号：N311.13[自然科学总论]