Research on User Profiling Technology for Personalized Demands

User profiling technology is discussed based on the diversity and randomness of user demands. The key technologies, which include profile initialization and updating, are researched. The user profile is expressed in form of vector space model. It´s built using centroid-based classification method. User´s browsing behaviors are analyzed to get feedback implicitly, calculating the documents degrees. A periodic adaptive learning mechanism based on Rocchio algorithm is put forward. The profile´s capability to dynamic track user needs is verified using satisfaction as evaluation indicator in experiments.