Design And Implement Of News Recommendation System Based On User Characteristics

As a new kind of media, network has deeply changed people’s traditional way of life.People use network services in various forms, and the utilization ratio of netnews has beenstable at a high level. Currently, mainstream news sites present the same content for differentusers. Users can only find the news they wanted by using navigation and search function.Thecontents recommended by news sites almost are the most popular or latest news, which don’tmeet users’ own interests. Users need to look for the news they wanted in mass amounts ofnews, costing a lot of time and effort.Personalized news recommendation system can help users to get the news they wanted,and dig out the news may be interested to users, which greatly reduced the time wasted onsearching. So, news recommendation system has a good prospect of application. At present,news recommendation system are mostly implimented by users’ subjective score, answeringquestions or manual settings, which putting too much emphasis on the role of users, causinggreatly abated on user experience. Also, the system availability is at a low level.In this paper, we describe the research status of personalized recommendation technology,and focus on the causes of why traditional news recommendation system can’t meet the users’needs. We discuss and analyse the user characteristics on news-oriented field, detailed reportthe source of user data, the stability and change of user characteristics. Also, we give aconcrete model of user characteristics.We present a news recommendation system based onuser characteristics. The recommendation system obtains user characteristics through clickrecords of users, then, analyses and predicts what users may be intersted in. On the basis ofuser model, this system extracts recommended rules and recommends news to users. Becauseof different users having different user characteristics, this system achieves personalized newsrecommendation services for different users, and improves the recommendation quality andaccuracy, enhances the user experience.