关联规则挖掘算法介绍  

Introduction of Mining Association Rules Algorithm

作　　者：陈玉婷[1] 王斌[1] 刘博[1] 宋斌[1] 李颉[2] 

机构地区：[1]北京邮电大学电信工程学院,北京100876 [2]中国人民解放军国防大学研究生院

出　　处：《计算机技术与发展》2006年第5期21-25,共5页Computer Technology and Development

摘　　要：数据挖掘是一个多学科交叉融合而形成的新兴的学科,它利用各种分析工具在海量数据中发现模型和数据间的关系。而在大规模事务数据库中,挖掘关联规则是数据挖掘领域的一个非常重要的研究课题。文中介绍了关联规则挖掘的研究情况,描述了经典Apriori算法的实现,并对该算法进行了分析和评价,指出了其不足和原因。描述了FP树挖掘最大频繁项集的算法,通过实例对该算法进行了性能评估,并得到结论:数据库中潜在的最大频繁模式越多,运行时间越长。Data mining is an emerging subject that composed and amalgamated by multiple subjects. It is an analytic process designed to explore data in search of consistent patterns and/or systematic relationships between variables. Mining association rules in business transaction datahases is one of the important topic of research on data mining. This paper introduced the research complexion of the association rules mining algorithm, describes the classical Apfiori algorithm,analyses and evaluates it. The author emphasizes FP tree mining maximum frequent item sets algorithm specially. And evaluates performace of the algorithm through instance. At the end, the paper gives the conclusion：the more maximum frequent item pattern in the database, the longer run time is needed.

关 键 词：数据挖掘 关联规则 频繁项集 FP树 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]