首页 | 本学科首页   官方微博 | 高级检索  
     

快速发现关联规则挖掘算法的并行化方法
引用本文:陆楠,陆春一,周春光.快速发现关联规则挖掘算法的并行化方法[J].深圳大学学报(理工版),2005,22(4):334-339.
作者姓名:陆楠  陆春一  周春光
作者单位:1. 深圳大学信息工程学院,深圳518060
2. 吉林大学计算机科学与技术学院,长春,130021
基金项目:深圳大学科研基金资助项目(200128)
摘    要:分析挖掘关联规则主要并行算法及性能.针对算法中负载平衡和时间响应问题,提出一种高效可行的挖掘关联规则的NA(Ntransactionalgorithms)并行算法,给出了NA算法的策略.通过前期实验结果表明,这种快速发现关联规则的并行算法在计算大项集过程中不需要同步和交换数据,在任意情况下,可独立计算局部大项集.

关 键 词:数据挖掘  关联规则  并行算法  概念格  频繁项集  负载平衡
文章编号:1000-2618(2005)04-0334-06
收稿时间:2005-06-06
修稿时间:2005-09-14

The parallel method on fast finding mining algorithms of association rules
LU Nan,LU Chun-yi,ZHOU Chun-Guang.The parallel method on fast finding mining algorithms of association rules[J].Journal of Shenzhen University(Science &engineering),2005,22(4):334-339.
Authors:LU Nan  LU Chun-yi  ZHOU Chun-Guang
Affiliation:1. College of Information Engineering Shenzhen University Shenzhen 518060 P. R. China;2. Department of Computer Science Jilin University Changchun P. R. China
Abstract:Parallel algorithms of mining association rules were introduced. A new and highly efficient algorithm of NA(N-transaction algorithms)was developed to balance the load and response of time in the algorithm. An analysis on performance of algorithm and its strategy was made. By our early experiment running, results show that the parallel algorithms for fast finding association rules do not need synchronization and data exchange in computing big item set. In addition, this new algorithm suffices to compute local big itemset by itself.
Keywords:data mining  association rules  parallel algorithm  lattice theory  frequent item set  load balance  
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号