Study And Design Of User Model In Personalized Service

With the development of computer technology, Internet has become an indispensable information source for people. However, comparative to the immense, disorderly Internet information space, the information, which each user is truly interested in is extremely finite, is merely a drop in the bucket. How to quickly, accurately seek the needed information from the vast information resource has already become a difficult problem that puzzles the network users. Under such demand driving, the personalized information service technology undertaken considerable development, and became a focus research in the domain of information service. Personalized service gives different services strategy and different services content according to different users, the key question of which lies in that the system must know the user's interest, and accurately establish the user interest model. Whether or not user model can accurately reflect the user's interest decides the quality of the personalize information service system.The primary works of this thesis are as follows:(1) The rise of personalized services is described firstly. To be the core of personalized service, user modeling has recently attracted much attention. The techniques of user modeling are induced, mainly aims to probe into how to organize the users' interest sources effectively and how to acquired, indicate, upgrade the users' interest, as well as the main modeling technology and so on.(2) A method that using the interest document submitted by user to construct user model is proposed based on analysis the traditional supervised modeling, also the process of modeling is narrated in a detailed way, main include the presentation and preprocessing of the demonstration documents.(3) Introduced the process of automatic modeling based on user's browsing information. The user's behavior data is collected through a design of browser plug-in, which can catch user's behavior data in client-side. There is close correlation between user's web browser behavior and user's interest for the web page, in this paper, we described the correlation by linear regressive equation, then through experimentation we proved that this user interest model was tenable, reasonable and effective.(4) Proposed the user model updating algorithm that used in this paper based on analyzing the existing updating algorithm.(5) Designed a user model structure, which can describe user's interest accurately and effectively, also it has a better usability and flexibility, and can track the user's interest change, easy to update the user model.(6) We make experiment on the construct and update user model, which mainly includes user demonstration documents which provides by the user to study initial user interest, and using user browsing information update user model. According to the experimentation and analyses, it is proved that the user model establishment and the renewal realization in this article are realizable, and the user model is reliable, it has the practical application value in personalization system.