项约束频繁项集挖掘的新方法  

New method for mining frequent itemsets with item constraints

作　　者：李英杰[1] 

机构地区：[1]浙江林学院信息工程学院,浙江临安311300

出　　处：《计算机工程与应用》2009年第3期161-164,共4页Computer Engineering and Applications

摘　　要：项约束频繁项集挖掘是项约束关联规则挖掘的关键步骤。对项约束频繁项集挖掘的内涵进行讨论,认为一个项集X本身满足项约束条件B是不够的,数据库中支持X的全部事务均满足B才能称"项集X满足条件B"。据此,将Direct算法改进为Di-rect*,在Direct*中负项被作为一个独立的项来看待。项约束是简洁性约束,但目前已有的算法没有充分利用其简洁性,提出利用项约束简洁性的MSEB算法。实验表明:对稠密数据库,MSEB的效率较高,并且Direct*和MSEB两个算法均是正确的。Mining frequent itemsets with item constraints is a key step for mining association rules with item constraints.The meaning of mining frequent itemsets with item constraints is discussed first,the paper affirms that an itemset＇ X satisfy an item constraints B by itself is not enough.It can be said X satisfy B only if all transactions that support the X satisfy the B.Based on this point,the algorithm Direct is modified to be Direct^＊.A negative item is looked as an independent item in Direct^＊.Item constraint is succinct,while the algorithms proposed do not utilize the succinctness adequately till now.A new algorithm MSEB is proposed and it utilizes the succinctness adequately.Experiments show the correctness of the two algorithms：Direct^＊ and MSEB, which also show the efficiency of MSEB.

关 键 词：数据挖掘 频繁项集 项约束条件 简洁性 稠密数据库 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]