Research On Mobile Personalized Service System

Human society has entered the information age, the ever-increasing amount of information has brought the joy of sharing information to us. It has also brought puzzles of "Getting Lost of Information" and "lack of knowledge". Personalized service is an inevitable result of the growth of internet information, with the current surge in the number of mobile user terminals, mobile personalized service focus of our attention. Mobile personalized service as the core technology for mobile users modeling research has gradually become the focus of our attention.Basing on Mobile Terminal Information System, I analyzed the necessary of Personalized Information Service. I also introduce the background, significance and the current state of research on personalized service. Then, I give a introduce on classification system of Personalized Information and some related personalized user modeling technology.This thesis proposal a Naive Bayesian classifier based personalized mobile user model. The model learns from training sets of different fields, dig through the user's browsing history and predict the user interests. By combining Bayesian statistical classification and frequency, the model can update the user model to get improvement of accuracy. Meanwhile, action model which is abstracted from user action is also combined into this user model. User model is constructed implicitly and automatically. Compared with the traditional system of personalized service, the method can predict user interests accurately. Basing on the characteristics of mobile personalized services, I furthermore revise the model with a forgotten factor and long-term interests and short-term interests of users.A series of experiments have been implemented, I give the experimental data and simulation results in figures. The experiments show that the user model proposed in this thesis has with better accuracy, effectiveness and practicality. Finally, I summarize the characteristics of the user model, and discuss the future work on improving this research.