The Design And Implementation Of News Personalized Recommendation System Based On User's Interest

With the development of Internet, it has become an important source of information. However, with the increasing accumulation of information, it becomes very difficult for the users to find the information which they really need. Although there are many search engines (such as google, baidu) to meet the needs of users, they return a general result to all users, which can't meet the users'special request based on different dimensions, such as time, background, purpose and so on.In the competition of the growing Internet market, the click-trough rate is a measure of the success of a website. The high click-trough rate not only allows the website to develop in a good environment, but also increases the visibility of the website.The significant click-trough rate is built on a stable and large user group. If the website gives different pages to users based on their interest, it not only help users find the content they need quickly and accurately, but also make users have a sense of being attached importance to, which can increase the number of users.Personalized recommendation system is in the context of this demand generated. Personalized service is the inevitable result of the expansion of the network information, and is the trend of website development in future, and it will be the focus of the Internet research area.To investigate the Personalized Recommendation System, this paper designs and implements a news personalized recommendation system for the website, www.idoican.com.cn . Firstly, This system deals with the news with data mining, and divides the news into several parts in use of the technology of automatic classification, keyword extraction ,subjects finding, among which the subjects finding uses the association rules, finally we get a news theme set. Every news theme is associated with a number of news reports, which not only provides the news navigation function for users, but also provides the recommendation basis for later personalized recommendation service. Based on the analysis of user behavior, the system builds the users'interest models in use of the theory of memory model, and recommends to users some news articles which they may be interested in, and then generates the personalized news pages for users. And the system updates the users'interest models through tracking and recording the users' click. Because the interests of every user are different, every user is able to get the personalized service in line with their own interests. In this way, we will get the satisfactory results for different dimensions of needs.Experiments show that this system can appropriately divide the daily news, and accurately recommend to the virtual users in the experiments news pages which they are interested in .The experiments result is satisfactory.The paper has some reference value to the study of some application areas of the personalized recommendation system.The main structure of this paper is as follows:Chapter 1 describes the background and significance of the research on the Personalized Recommendation System, and describes the research information on the personalized recommendation system at home and abroad, and then outlines the main contents of this article, gives the organizational structure of the paper.Chapter 2 introduces the concept, classification, and related technologies of the Personalized Recommendation System, and then briefly discusses the concept and algorithms of association rules.Chapter 3 first outlines the system, and introduces the system design constraints, the realization of the principle of the recommendation, the workflow and relationship between modules, flow chart of the overall work system and the database table structure.And then it introduces the detailed design and implement of the system with the example of the theme management module and the user management module:Chapter 4 introduces the design and implement of the theme management module, reviews the association rules, and describes the process of subjects finding with Apriori algorithm using a practical example, and gives the pseudo-code of the Apriori algorithm.Chapter 5 introduces the design and implement of the user management module, including the source of the user model, the model building process, the interaction between the user and the system and the implementation code.Chapter 6 introduces the environment and technical support of the system, describes the operation of the theme of discovery and establishing and updating the user interest model, and then gives the background management interface and foreground display interface, and introduces, and the analysis about the result of the experiment.Chapter 7 makes a summary of the research, and describes a few ideas about how to improve the system.