Research On Interest Model Of User Behavior

With the rapid development of Internet, when people get wealth of information, they also get much problems from much information at the same time. The traditional information service cannot meet the demand of users now. Therefore, the research and application of personalized service technology which supplys the personal service for different users is paid more attention. The key of personal service technology is to find the demand of users, and it can be supplied by Interest Model of User Behavior.However, the existing Model of User Behavior still has some weak points, for example, the very big system expenses, the neglect of the decay of user interest, and the long-term interest of user. And most of the systems mines users' logs by servers, and it overlook the information of clients. By analysis the users' interest model, we proposed a method of building interest model based users' behavior, which research the relevance of users' behavior and their interest, that is to say, pick up the information of users'behavior, then quantization and analysis to choose a effective page.By the theory research, I developed the interest model system based users' browse behavior which include picking up information, data processing, data analyse modeling and so on. The level of picking up information pick up and analysis the information of users' behavior to build the interest of users' browsers pages. The level of data processing cluster analysis the pages to select the users' interested pages by clustering algorithm, and quantize it by word frequencies. The level of data analysising analysis the pages to build the users' interest model, and the model can be adjusted by the dency of users' interest, the users' feedback and the users' new interest.At last, the experiment of prototype system show that, the Interest Model of User Behavior can describe the interest of users correctly. And the interest attenuation factor can react the changes of users' interest. The system has application value in customized services.