一种网页推荐系统的设计与实现

发布时间：2019-04-21 17:38

【摘要】：面对大量的信息和信息的多元化,人们可获得的信息越来越多,但是快速有效而又准确地获取自己所需的信息却是相当困难的。搜索引擎的出现解决了用户的检索需求,但却不能适时地为用户主动提供其需要的信息。当用户不确定所需信息的关键词描述,无法进行搜索；又或者用户自己不愿进行搜索,只是希望获得一些信息的推荐时,就需要一个比搜索引擎更主动的系统来满足这些需求,智能推荐服务系统因此诞生,推荐技术也日益成为一个重要的研究课题。本文针对用户的信息推荐需求,设计并实现了一种网页推荐系统。本系统以基于奇异值分解的协同过滤算法为关键技术,用文本分类中的特征选择方法来预处理网页信息,并运用特征选择方法对收集到的网页以及用户潜在兴趣页面进行处理,同时引入聚类分析算法解决服务器负荷大和推荐结果可能不准确的问题,最后通过用户反馈和用户使用时间段对推荐结果列表进行优化,使推荐结果的准确度提高,使得推荐结果更加符合用户的需求,更加人性化。本文设计的网页推荐系统是而向用户的,结合用户兴趣并采用基于奇异值分解的协同过滤推荐算法实现的混合推荐系统。本推荐系统综合考虑用户正在浏览的页面、用户的兴趣和用户历史访问记录,结合用户使用推荐系统的时间段来为用户提供网页推荐服务。最终实现的推荐系统可以发现用户的兴趣和爱好,并向目标用户推荐符合其兴趣爱好的信息或物品,参考用户反馈和用户使用时段来进行智能推荐。论文首先对网页推荐系统的研究背景、研究现状和研究内容以及在实际中的应用进行了综述。在对网页推荐系统进行了需求分析的基础上,论文提出了网页推荐系统的总体架构和概要设计,对模块进行了详细设计和编码实现,并针对采用的关键技术和方法进行了详细的介绍和说明。论文的最后对网页推荐系统进行了实验测试,并指出了网页推荐系统需要进一步改进和完善的方面。
[Abstract]:In the face of a large number of information and the diversity of information, people can obtain more and more information, but it is very difficult to get the information they need quickly, effectively and accurately. The emergence of search engine solves the user's need of retrieval, but it can not provide users with the information they need at the right time. When the user is uncertain about the keyword description of the required information, the search can not be carried out; Or if users do not want to search themselves, just want to obtain some information recommendations, they need a more proactive system than the search engine to meet these needs, intelligent recommendation service system was born. Recommendation technology has increasingly become an important research topic. In this paper, a web recommendation system is designed and implemented according to the information recommendation requirements of users. This system takes the collaborative filtering algorithm based on singular value decomposition as the key technology, uses the feature selection method in text classification to pre-process the web page information, and uses the feature selection method to process the collected web page and the user's potential interest page. At the same time, cluster analysis algorithm is introduced to solve the problem that the server load is heavy and the recommendation result may not be accurate. Finally, the recommended results list is optimized by user feedback and user usage time, so that the accuracy of recommendation results can be improved. So that the recommendation results more in line with the needs of users, more user-friendly. The web recommendation system designed in this paper is a hybrid recommendation system which combines user's interest and adopts collaborative filtering recommendation algorithm based on singular value decomposition (SVD). This recommendation system takes into account the pages that the user is browsing, the user's interest and the user's historical visit record, and combines the time period of the user's use of the recommendation system to provide the web page recommendation service for the users. The final implementation of the recommendation system can discover the interests and interests of users and recommend to the target users information or items in accordance with their interests, reference to user feedback and user use time to carry out intelligent recommendations. Firstly, this paper summarizes the research background, research status and application in practice of web recommendation system. On the basis of analyzing the requirement of web recommendation system, this paper puts forward the overall structure and outline design of web page recommendation system, and designs the module in detail and implements the coding. The key technologies and methods adopted are introduced and explained in detail. At the end of the paper, we test the web recommendation system, and point out the aspects that need to be further improved and perfected.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.3

【参考文献】