| 基于VOPP并行编程环境的最大频繁项集生成方法 |
| |
| 引用本文： | 郑晓艳,石连栓,孙济洲. 基于VOPP并行编程环境的最大频繁项集生成方法[J]. 计算机应用研究, 2009, 26(4): 1305-1307 |
| |
| 作者姓名： | 郑晓艳 石连栓 孙济洲 |
| |
| 作者单位： | 天津大学,计算机科学与技术学院,天津,300073;天津工程师范学院,计算机系,天津,300222;天津工程师范学院,计算机系,天津,300222;天津大学,计算机科学与技术学院,天津,300073 |
| |
| 基金项目： | 天津市自然科学基金资助项目（06YFJMJC01600） |
| |
| 摘 要： | 频繁项集挖掘是关联规则挖掘的核心内容,提出了一种挖掘最大频繁项集的并行算法CDTR。它对CD (counting distribution)算法进行了改进,根据一种新的分布式共享内存环境下面向视图并行编程思想,将数据库划分成视图。为了实现动态任务分配,对数据库进行了预处理。实验结果显示CDTR能够高效地生成最大频繁项集,大大提高了分布式共享内存系统的效率。
|
| 关 键 词： | 分布式共享内存系统 面向视图并行编程 关联规则 最大频繁项集 |
Algorithm for finding frequent itemsets based on VOPP |
| |
| ZHENG Xiao-yan,SHI Lian-shuan,SUN Ji-zhou. Algorithm for finding frequent itemsets based on VOPP[J]. Application Research of Computers, 2009, 26(4): 1305-1307 |
| |
| Authors: | ZHENG Xiao-yan SHI Lian-shuan SUN Ji-zhou |
| |
| Affiliation: | (1.College of Computer Science & Technology, Tianjin University, Tianjin 300073, China; 2.Dept. of Computer, Tianjin University of Techno-logy & Education, Tianjin 300222, China) |
| |
| Abstract: | Mining frequent item sets is a crucial issue in data mining applications. This paper proposed a novel and powerful parallel algorithm for mining maximal frequent item sets, called CDTR. CDTR improved the counting distribution algorithm based on VOPP (view-oriented parallel programming), a novel style for parallel programming on cluster computers. It divided search space into views and preprocessed database to help dynamic tasks allocation. Experiments show that CDTR finds maximal frequent item set efficiently and improves the performance of distribution shared memory system. |
| |
| Keywords: | distribution shared memory view-oriented parallel programming association rules maximal frequent itemsets |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《计算机应用研究》浏览原始摘要信息 |
|
点击此处可从《计算机应用研究》下载全文 |
|
