Research On Adaptive User Modeling For RSS Web Information Services

Because of the rapid development of Internet technology with the problem of overloading information, people have to spend much more time and energy on searching and skimming needful information. How to get what you need from the excessive information has become an urgent issue. The appearance and application of the personalization system give a new way to solve this issue, which provides different services to different users. User modeling and model updating technology have already become the core of individuation service, and also become one of current research hotspots. RSS-based information service has been widely used in many areas, which changes the way of obtaining information from people searching information actively to information finding people automatically. Furthermore, it can fast provide the newest information to users. In this paper, we mainly focus on researching the application of adaptive user modeling in RSS-based information services, with our object of improving its quality.The user model of the article bases on the vector space model (VSM), which represents as the duality group of users'interests and their weight, and the users'interests adopt mixed users'interest model, consisted of long-term and short-term interest model. Because of users'interest changing in an uncertain way, and in order to trace the changing of user model, the article makes use of an approach to update user profile which bases on Rocchio feedback. Then a method combining single feedback and times feedback of Rocchio has been introduced, with the purpose of tracking the concept drift of users'interests and implementing online and offline learning. Online learning increased system plasticity, and offline learning assured system stability. What we have done is to find a balance on the plasticity and stability problem.Finally, through designing and implementing a user adaptive information service platform--RSS News Filtering system, the paper attempts to providing personalization RSS information service. Meanwhile, the topic-oriented search engine is as a new information source, and this platform makes use of Google's custom search engines to provide much more related information for users successfully. Experimentation data indicates that it is available to use the method of user modeling and updating.