| Nowadays, Internet has become one of the most important ways to obtain information, and search engine is the most commonly used tool when searching information. Search engine is quite powerful, but at present the services provided by most search engines, fail to satisfy users' comprehensive and demanding needs, particularly in the aspect of personalized services. General search engines lack intelligence, which leads to many search results provided irrelevant to users' search intentions. Therefore, to help Internet users to obtain information according to their personal interests, retrieving information and screening irrelevant information automatically, that is, providing personalized information services in network environment is of great significance.The key to realize personalized service is correctly retrieving and analyzing the visit information of web users, and then accurately describe their interest. Whether the user model can accurately reflect the real needs of users or not will directly affect the actual results of the system. Only accurately obtaining the users'needs can the system offer desirable resources to them. In this paper, by retrieving and analyzing the browsing information generated in the course of user visit to the Internet, we explored the way to acquire, express and build the user model characterized by client-based information mining. Using RDF file format to store user preference information, we designed a user interest model based on classification methods.On the ground of above work, we designed and implemented a personalized information recommendation system of interest-based model. The system, based on B/S model, can recommend information consistent with user interest and preference. Experimental results show that the system gains better recommendation accuracy and can dig deeper for getting user interest to meet user preferences, overcoming the problem of poor quality recommendations in large measure. |
PDF Full Text Request