适合于高效更新的关联规则挖掘算法  

Practically High Efficiency Calculating Method in Association Rules and Datamining

作　　者：周海岩[1] 

机构地区：[1]太原师范学院计算机系,山西太原030012

出　　处：《小型微型计算机系统》2004年第4期634-637,共4页Journal of Chinese Computer Systems

基　　金：山西省高校科技研究开发项目 ( 2 0 0 2 0 5 6)资助

摘　　要：实用的关联规则挖掘算法 ,为了发现事先未知的关联规则 ,用户需要通过对最小支持度和最小可信度这两个阈值的不断调整来逐步聚焦到那些真正令其感兴趣的关联规则上去 ,这将是一个动态的交互过程 .因此 ,迫切需要高效的更新算法来满足用户对较快的响应时间的需求 .基于这种思想 ,并深入分析了已有的诸关联规则挖掘与更新算法且指出其共同存在的问题与不足 ,在此基础上 ,提出一种当数据库数据不变时 ,仅扫描数据库一次 ,即可反复调整最小支持度和最小可信度进行关联规则挖掘与更新的高效、实用的算法 ,特别在对关联规则进行更新时 ,该算法对最初和前次挖掘过程中所得到的信息加以充分的利用 ,从而对关联规则进行更新时算法的执行效率得到进一步的提高 .It introduces a calculating method in association rules and data mining. To discover the association rules unknown in advance,the user will have to gradually focus on to those association rules interested in by continual adjusting between the two threshold values with minimum support and minimum credulity.this is an alternatively dynamic process.thus an updated high efficiency calculating method is required urgently to satisfy the user's need for faster respond time span. On the basis of this consideration, and on the basis of the analysis of existing association rules,data mining and updating methods,pointing out their common problems and deficiencies,this paper raised a practically high efficiency calculating method which is cable to carry on association rules and updating calculation by scanning the database only once and repeatedly adjusting the minimum support and minmum credulity .In particular,during the updating of association rules, the efficiency of the method is further increased by taking full use of the information of the begining and the process of dataming .The paper analyzes and discusses on the calculating method.

关 键 词：知识发现 数据挖掘 关联规则 增量式更新 频繁项目集 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论] TP18[自动化与计算机技术—计算机科学与技术]