【作者】 陈振；

【导师】 郑诚；

【作者基本信息】 安徽大学 ， 计算机应用技术， 2005， 硕士

【摘要】 Web挖掘是数据挖掘的一个重要的研究方向。Web挖掘突破传统意义上的数据格式的限制,随着Web的发展,它会变得越来越重要,成为web用户更好的获得信息的一个有效的途径。 如何对网上用户分类是web挖掘领域应用最多的任务之一。对网上用户分类可以使网站针对不同的用户提供不同的服务,这样,用户就能方便快捷的获得自己所需的信息。如何有效的找出关于web用户的分类模式知识是web挖掘研究的热点问题,具有重要的理论意义和实用价值。 本文针对数据挖掘技术在web挖掘中的应用进行了研究工作。在第一章绪论中介绍了本论文的研究背景、研究意义,以及论文的主要工作。第二章对数据挖掘技术的定义,结构,过程以及一些分析方法进行了综述。第三章分析Web挖掘当前研究现状,指出了Web挖掘的困难,着重讨论了Web内容挖掘技术。第四章讨论了Web使用挖掘的过程,模式发现的常用算法,详细分析了日志预处理的各个过程。 第五章中对关联规则的经典算法Apriori进行分析,并指出它的一些缺陷,存在的问题,同时,也讨论FP-growth算法,指出用FP-growth算法来挖掘频繁项集可以避免产生大量候选集,并且算法效率有很大的提高。在此基础上,分析了传统分类关联规则挖掘算法产生的问题,提出一种用于web挖掘的分类关联规则挖掘算法。 第六章运用不同的方法来对web用户兴趣预测。试验表明用于web挖掘的分类关联规则挖掘算法是有效的。在全文的分析基础上,提出一种基于web挖掘的网站个性化推荐系统结构。 第七章是全文的总结和研究工作展望。 论文主要做了以下工作: ● 通过对经典关联规则挖掘算法Apriori进行分析,同时,也讨论FP-growth算法,指出用FP-growth算法来挖掘频繁项集可以避免产生大量候选集,并且算法效率有很大的提高。在此分析基础上,分析了传统分类关联规则更多还原

【Abstract】 Web mining is an important research direction in data mining. It breaks through the restriction of traditional data behaving format, and with the internet development, it becomes more and more important. It is an effective way for web users to get information.Web users classification is one of the most commonly used tasks in web mining. It can help web site provide personalized services for web users and so the users can get information conveniently .How to find all those classification patterns of web users is a research issue in web mining and has great theoretic significance and practical value.The paper makes some researches of the application of data mining to web data. It introduces the background, the significance and the structure of the paper in the first chapter. In the second chapter, the paper introduces the concept, the structure and the process of data mining techniques. The third chapter describes web mining, including the state and the difficulties in the web mining. In this chapter, the" paper summarizes web usage mining, including process and pattern finding arithmetic and puts emphases on the preprocessing of web log.In the fifth chapter, we analyze the traditional association rules arithmetic and point out some deficiencies of the arithmetic. Then we analyze the FP-growth arithmetic. Based on the analysis, we put forward a classification association rule finding arithmetic for web mining.In the sixth chapter, we attempt to apply some arithmetic, including ID3, traditional classification association rule finding arithmetic and the association rule finding arithmetic for web mining. The result shows that the association rule finding arithmetic for web mining is effective.In the seventh chapter, we summarize the whole paper and make aprospect of our researches. Main works of the paper:1. FP-growth arithmetic dose not produce the candidate items and the efficiency of the arithmetic is better than Apriori arithmetic. Based on the analysis, the paper puts forward a classification association rule finding arithmetic for web raining.2. We attempt to apply some arithmetic, including ID3, traditional classification association rule finding arithmetic and the classfication association rule finding arithmetic for web mining to predict the web users’ interests. We compare the results and conclude that the classification association rule finding arithmetic for web mining is effective.3. We put forward a framework of a personalized recommended system for web sites and the framework integrates web usage mining, web content mining and web usage mining.更多还原