基于访问行为的个性化推荐网络购物系统设计与实现

发布时间：2018-05-09 09:31

本文选题：数据挖掘 + 个性化推荐　；参考：《电子科技大学》2014年硕士论文

【摘要】：随着Internet的快速发展,网络已经成为人类生活的一部分,也成为人们获取信息的一个重要途径。由于网络信息量的不断增加,人们不得不花时间从海量的信息中搜寻去自身所需的相关信息。仅仅依靠人们的搜索时很难在短时间内找到自己的目标的。虽然搜索引擎可以帮助人们解决这个矛盾,但是由于搜索引擎缺乏智能以及个性化的推荐,因此不能从根本上解决这个难题。解决这一难题的思路就是本文提出的基于Web Agent的用户访问行为的个性化系统。本文重点以用户浏览行为的角度分析了用户对网页的兴趣度,对于当前被广泛使用的用户个性化模型只依赖页面内容而建立的方式,具有一定的启发作用。本文在总结了其他学者们的研究成果基础上,对Web Agent和Web数据挖掘进行了研究,并根据用户行为的特性,提出了一种基于关联规则的协同过滤算法,并对本基于Web Agent的个性化推荐系统进行了设计和原型实现。本文所做的工作如下:1.提出了获取客户端用户行为数据的方法并对行为数据进行数据挖掘,将用户的访问行为和其有兴趣的页面结合起来构建了一个基于Web Agent的用户访问行为的个性化推荐原型系统。2.使用Web数据挖掘对用户的行为数据进行了分析,由于服务器记录的用户数据具有冗余性,而且服务器和客户端的日志文件也存在着差异,本文就此使用Web数据挖掘,就用户行为数据进行了筛选和排除。本文设计的基于Web Agent的个性化推荐原型的数据处理包括三个流程:在线监听、离线学习和在线推荐。在线监听的作用是将用户的各种行为数据、注册信息等数据结合起来成为有价值的数据挖掘的数据来源,通过对这些数据进行挖掘和分析,就能发现其中的关联规则。离线学习包括为内容数据的预处理、结构数据的预处理和使用数据的预处理。当预处理结束之后,再选用合适的工具对这些模式和数据进行分析,从海选的数据中选拔出有用的规则。在线推荐模块式根据使用者的喜好向用户其推荐他们可能会感兴趣的产品,然后在根据用户的反应,系统给予统一的评价,通过挖掘规则提取出的模式与当前用户会话比较,生成用户需要的个性化页面。
[Abstract]:With the rapid development of Internet, the network has become a part of human life, and it has also become an important way for people to obtain information. Because of the increasing amount of information in the network, people have to spend time searching for the necessary information from a large amount of information. It is difficult to find a short time only by relying on people's search. Although the search engine can help people to solve this problem, the search engine can't solve this problem fundamentally because of lack of intelligence and personalized recommendation. The idea to solve this problem is the personalized system of user access behavior based on Web Agent. This paper focuses on the use of this paper. In the perspective of user browsing behavior, the user's interest in web pages is analyzed. It has a certain inspiring effect on the way that the widely used user personalization model relies on the content of the page only. This paper has studied the Web Agent and Web data mining on the basis of the research results of other scholars, and based on the users. A collaborative filtering algorithm based on association rules is proposed, and the personalized recommendation system based on Web Agent is designed and implemented. The work done in this paper is as follows: 1. the method of obtaining the client user behavior data and data mining for the behavior data are proposed, and the user's access behavior and the user's access behavior are presented. Interested pages are combined to build a personalized recommendation prototype system based on Web Agent user access behavior..2. uses Web data mining to analyze user behavior data. Because the server records are redundant, and there are differences between the server and the client's log files. This uses Web data mining to screen and exclude user behavior data. The data processing of personalized recommendation prototype based on Web Agent includes three processes: online monitoring, offline learning and online recommendation. The role of online monitoring is to combine the user's various behavioral data, registration information, and so on. Data mining is a source of data mining. By mining and analyzing these data, the association rules can be found. Off-line learning includes preprocessing for content data, preprocessing of structural data and preprocessing of data. Select the useful rules from the data selected by the sea. The online recommendation module recommends the products that they may be interested in according to the user's preference, and then gives a unified evaluation based on the response of the user, and compares the patterns extracted by the mining rules to the current user conversations to generate the personalized pages needed by the user. Noodles.

【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP391.3

【参考文献】