Personalized Information Service On XML

With the rapidly increasing information on Internet, traditional information service can not satisfy customers any more. This paper discusses current user modeling techniques and the ability of corresponding personalized recommendation. It puts forward a new compositive user model to develop these models'advantages. Our user model is modified automatically according to users'visiting behaviors; therefore the services the personalized recommendation applied are more and more effective. Our test website is to provide essays on computer science domain. The essays'contents are memorized in XML format for its good semantic functions and excellent constructions. We put our compositive user model and personalized recommendation technologies into the test website to construct a personalized service system. Experiment results indicate that it improves the personalized service effectively.