首页 | 本学科首页   官方微博 | 高级检索  
     

一种新的MDP算法的研究
引用本文:高集荣,田艳,杨永红,党运峰. 一种新的MDP算法的研究[J]. 微型机与应用, 2012, 31(5): 8-10,16
作者姓名:高集荣  田艳  杨永红  党运峰
作者单位:中山大学信息科学与技术学院,广东广州,510006
基金项目:广东省自然科学基金项目
摘    要:提出了一种高效的挖掘数据仓库中多维关联规则的MDP算法。MDP算法通过构造一种扩展的前缀树MDP-tree,将数据仓库中的有效信息压缩存储,再使用基于MDP-tree的MDP-mining方法快速发现有趣的关联规则。MDP算法仅需要扫描一次数据仓库,就可以构造出MDP-tree,进而得到所有的关联规则。该算法还具有频繁模式查找简捷、二次查找迅速等优点。通过实验验证了MDP算法的高效性和稳定性,与传统的多维关联规则算法相比有更好的性能。

关 键 词:数据挖掘  多维关联规则  FP-growth算法  MDP算法  频繁模式

Research of a new MDP algorithm
Gao Jirong,Tian Yan,Yang Yonghong,Dang Yunfeng. Research of a new MDP algorithm[J]. Microcomputer & its Applications, 2012, 31(5): 8-10,16
Authors:Gao Jirong  Tian Yan  Yang Yonghong  Dang Yunfeng
Affiliation:(Department of Information Science and Technology ,Zhongshan University ,Guangzhou 510006 ,China )
Abstract:A novel algorithm is proposed for mining multi-dimensional association rules: MDP algorithm, which is based on FP-growth algorithm. An extended prefix tree, which is called MDP-tree, is constructed to store compressed, crucial information of a data cube. An efficient method, MDP-mining, is also developed to mine association rules. The MDP-tree can be built by scanning a data cube only once. Therefore, it can save lots of time. Besides, efficiency of MDP algorithm is achieved with many other techniques. The performance study shows that the MDP algorithm is efficient and scalable for mining multi-dimensional association rules, and is about an order of magnitude faster than the traditional algorithms.
Keywords:data mining  multi-dimensional association rules  FP-growth algorithm  MDP  frequent pattern
本文献已被 CNKI 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