User Modeling Technology Based On The Structural Features Of Web Pages And Phrases

Along with the Internet development and the rapid growth of Internet information, the personalized service receives the users' favour day by day. As the core technology of personalized service, user modelling has also attracted more attention gradually. As a central concern in personalized service, the Users' Interest Construction Model Technology mainly aims to probe into how to organize the users' interest sources effectively and how to indicate,upgrade,store and calculate the users' interest. This article has analyzed and elaborated each technology in user modelling systematicly and completely, summarized the user model expression method, and analyzed the information source of user model, and has carried on the induction and the classification to the user modelling technology.Firstly, the paper puts forward the key techniques to build Personalized User Profile, such as Web Mining, User Action Mining, Machine Learning. Behind the background knowledge having been introduced, in view of the special nature of the web pages, the paper presents an user modeling technology based on the structural features of web pages. It through to promote vocabulary frequency of words in certain web page marks, put the structural features of web pages in the calculation of user profile. Then, the paper proposes an user modeling technique based on phrases, the prases have been put in user model.The experimental results show, the user modeling technology based on the structural features of web pages and the user modeling technology based on phrases all can build more effective user profile.