基于频繁闭项集的Web日志挖掘算法  

Web log mining based on closed frequent itemsets

作　　者：秦东霞[1] 周航[1] 张栋梁[1] 吴文欢[1] 

机构地区：[1]周口师范学院计算机科学与技术学院,河南周口466001

出　　处：《周口师范学院学报》2012年第2期97-100,共4页Journal of Zhoukou Normal University

基　　金：周口师范学院青年科研基金资助项目(No.zknuqn201031A)

摘　　要：介绍了Web日志挖掘的模型,分析了使用关联规则挖掘Web日志时遇到的规则数量大且存在冗余等问题,提出了基于频繁闭项集的挖掘办法来减少规则数量.同时引入最小关联规则的概念,从而避免了冗余规则的产生.最后用实验验证了算法的有效性,并以周口师范学院校园网为例,对该网站日志数据进行分析,得到了有价值的规则,并对该网站提出了相应的建议和意见.In this study, we introduced the model of Web log mining, and at the same item, a new method of Web log min ing is proposed. Firstly,we found that the traditional Web log mining approaches suffer from one major deficiency： a training data set often generates a huge set of rules,which is also redundancy. Secondly,we put forward the frequent closed itemsets based on the way to reduce the number of mining rules. The paper introduced the concept of minimum association rules, so as to avoid the generation of the redundancy rules. And then, our extensive experiments on learning database repository show that the algorithm is effective with higher quality and no redundancy. Finally, we take the campus web-log mining of zhouKou normal university for a ~tata source. We extracted valuahle rules and put forward recommendations and methods for improving the site.

关 键 词：WEB日志挖掘 频繁闭项集 最小关联规则 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]