Representation and Evolution of User Profile in Web Activity

In the Web personalization of Web service and construction of adaptive Web site, how to represent uses profile is one of the key issues. Aiming at solving the existing problems in representation of user profile includes incapability of effectively representing userpsilas recent interest, lack in dynamically effectively updating. This paper presents a new method of representation of user profile. By computing the interest degree with association degree between topics, updating of interest degree is preformed. By introducing memory model, it can predict which Web pages are still concerned by the user. By classifying the operations on the presented user profile, the construction and evolution can be done. In case study a userpsilas browsing activities is traced and corresponding user profile is built and evolves, thus to prove the feasibility of this method.