个性化推荐系统研究软件工程专业论文.docx

II II 南京邮电大学硕士研究生学位论文 Abstract Abstract With the rapid development of network technology, the problem of an explosion of information growth making it overload in the network becomes more and more serious. Its difficult for users to find useful information in the network. Some rarely concerned information is also submerged in the sea of information and becomes the isolated island of information. The traditional search engine cant help people to effectively solve this problem and personalized recommendation systems emerge as the times require. Finding their individual preference resources object for users is the essence of recommendation system. Inevitably, recommendation systems are confronted with many problems, such as the accuracy of recommendation, cold start and user interest changes with time and other issues. How to effectively solve these problems become many scholars’ research goals. In recent years, the maturity of Web2.0 technology makes the social tags system widely used. Social tagging systems introduce a novel platform for users participation and search network resources more efficiently. Because social tags can provide highly abstract information about not only item contents but also personalized preferences, the use of labels can improve the accuracy of personalized recommendation. In order to solve the limitations of the traditional personalized recommendation system, this dissertation studies how to use the social tag to improve the performance of the recommendation system. Aiming at the problem of user preferences changes over time in the personalized recommendation system, and combining with using tag frequency and label time on user-object-tag tripartite graphs, we propose a recommendation a lgorithm based on tag time-weighted network. Simulation experiments are performed in the Delicious and MovieLens two data. Experimental results demonstrate that the usage of tag time-weighted can significantly improve accuracy and diversification of recommend