Research On The Method Of Aggregation Of User Model For Cross System Personalization

With the gradual increase of the information resources on the internet, users'demands for personalized services are increasingly heightened. The existing personalization systems have their own user models respectively. Even if they save the same or the similar user information, the information couldn't be shared. Therefore, the user has to register his/her information repeatedly when he/she visits several different systems. Cross system personalization shares personalization information across different systems in a user-centric way, which can overcome the aforementioned problems.The method of aggregation of user models which come from different source is studied in this paper. The method represented in this paper is for cross system personalization. The material contents are as follows.Firstly, the representation of personal preferences for users is not exactly at present. Aimed to this characteristic, a user model which can be called Domain ontology-based user model and the structure of it are represented. The process of Learning and updating of user model are presented. It also analyses the process of aggregation of user models, which come from different source.Secondly, since the computing method of ontology concepts similarity is not exactly and comprehensively, a calculation model for general similarity of ontology concepts is introduced. Aimed at the problems of user clustering algorithm in the process of aggregation for user model which come from different sources, a computing algorithm of user similarities is introduced based on user inner product space model. This paper proposes a new user clustering algorithm based on user inner product space model, and in order to test the correctness and effectiveness of the proposed algorithm, this paper compares it with other algorithms from different aspects. Finally, the measurement of similarity of ontology concepts is not exactly in ontology mapping. Aimed to this characteristic, a new method for ontology mapping is presented. It is based on general similarity of ontology concepts. This paper also analyses the mapping tactics between different ontology. The ontology mapping tactics has been used in the process of aggregation of user model, which has been proved to be feasibly by experiments.