User Model Evolution Algorithm: Forgetting and Reenergizing User Preference

User model is one of the fundamentals to support personalized services. Not only the model has to represent user preferences precisely, but it must reflect the changes of user preferences over time as well. Existing user modeling approaches have several deficiencies such as lack of dynamics and adaptation, so the evolution of these models is much de-pendent on user´s explicit feedbacks. Particularly, the existing approaches cannot update user models automatically to adapt the drift of user preferences over time. Aiming at solving these problems, the Forgetting and Reenergizing User Preference (FRUP) algorithm based on forgetting and reenergizing mechanism of memory in psychology is presented in this paper. Firstly, the characteristics of memory are analyzed through ZGrapher. Then, FRUP algorithm which can trace the user preference through Ebbinghaus forgetting curve is presented. Finally, the preferences in FRUP algorithm are divided into long-term, medium-term and short-term for describing different memory patterns accurately in different cases. The forget-ting mechanism of memory in FRUP algorithm shows proper adaptivity to evolve the user model. The reenergizing mechanism can trace the changes of user preference in time.