A Personal Service For Web Searching Based On Relevant Feedback And User's Interest Modeling

The information time of human society has been coming. More and more information shared by more and more people. But, at the same time,some questions such as "Lost in Information"and "Lost of Knowledge"appeared. All hope to find information quickly and individually. Personal research on search engine's searching results can solve these questions. We proposes to acquire users'interests through interacting with them, Then filter and reset the search result based on users'interests. It means we acquire users'interests by feedback of the initial search results from users. Then the searched results of search engine will take on personal feature. This article studies and analyzes some important technology such as feature extraction, Chinese words segmentation and genetic algorithm and so on. This system has some main function including Web text pretreatment,text feature extraction, and similar calculation of Web features and user interests. Web text pretreatment washes the webs in order to delete some irrelevant items. Text feature extraction extracts the features of Web and organizes Web VSM (Vector Space Model). Similar calculation of Web features and user interests is the most important ways to analyze the relevance of Web and user's requirements. My paper introduces the course of invention and reality of this PASE (Personal Agent of Search Engine) system and the methods and technologies of personal service. We compare the relevant of PASE to some popular search engines and draw a conclusion that our system is superior to the latter in personal.