挖掘多支持率分类规则的虚拟投影算法  

PSEUDO PROJECTION ALGORITHM FOR MINING OF CLASSIFICATION RULES

作　　者：刘君强[1] 孙晓莹[1] 王勋[1] 

机构地区：[1]杭州商学院计算机信息工程学院,杭州310035

出　　处：《计算机应用与软件》2003年第9期8-10,共3页Computer Applications and Software

基　　金：浙江省自然科学基金(60 2 1 4 0 ) ;国家 863计划 (2 0 0 2AA1 2 1 0 64) ;浙江省教育厅科技计划(2 0 0 2 0 635)

摘　　要：本文首先提出了一种挖掘频集的高效算法PP。它采用了一种基于树的模式支持集表示 ,避免了反复扫描数据库和递归建造个数与频繁模式数相同的模式支持集 ,其效率比Apriori和FPGrowth高 1～ 3个数量级。PP被进一步扩展成发现分类规则的有效算法CRM PP。CRM PP将多支持率剪裁集成到频集发现阶段 ,将二阶段挖掘法改进为单阶段挖掘法。CRM PP的效率也比基于Apriori和FPGrowth的二阶段算法高 1～In this paper,an efficient algorithm,called PP(Pseudo Projection),is proposed to discover frequent patterns.PP represents subsets of transactions that support patterns by a tree based structure which avoids repetitive scans of databases and recursive materializations of transaction subsets.PP is one to three orders of magnitude efficient than Apriori and FPGrowth.Then,PP is extended into another efficient algorithm,called CRM PP,to mine classification rules.CRM PP pushes multiple minimum supports threshold into the discovery stage of frequent patterns,and generates rules in the same stage.CRM PP is also one to three orders of magnitude efficient than algorithms derived from Apriori and FPGrowth.

关 键 词：数据挖掘 虚拟投影算法 分类规则 数据库 模式支持集 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]