摘要
算法MARC(Mining Association Rulesusing Clustering)将聚类技术应用到关联规则的发现上,MARC利用聚类技术压缩交易数据库,从而减少开采算法需要处理的数据量以提高开采效率,同时算法提出了聚类汇总转换的概念用以减轻压缩数据带来的信息丢失。在几个实际数据集上的实验表明该算法可以达到高精度和高性能。
MARC algorithms are proposed to apply clustering analysis to other fields. It integrates clustering into association rules discovery to reduce the size of data sets. It also uses CS (Clustering Summary) transformation to alleviate the loss of information brought by the compression. MARC only needs to scan the database one time. The experiments with several real data sets have demonstrated that MARC can achieve quite well precision and high performance.
出处
《计算机科学》
CSCD
北大核心
2007年第8期180-183,214,共5页
Computer Science
基金
国家科技攻关计划项目(编号:2002BA901A02)
湖北省科技攻关项目(编号:2004AA210B01)
关键词
数据开采
聚类分析
关联规则
Data mining, Clustering analysis, Association rules