Research On Modeling User Webpage Browsing Interest

The rapid development of internet has involved us in the era of information explosion, and internet data's features of hugeness,speed and dynamic become more and more visible.The cyber brings a mass of information. Meanwhile, it brings new challenge too.How to search the information which fits different users and provide Web Information Personalization Service for different users makes the hot research. While the research on user interest model which can reflect user's intention well becomes the highlight of internet application in the personalization information service research. To enhance the expression accuracy and ability of user interest model, and improve the service quality of personalization information service system make significant contribution to the further development of internet.Firstly, the research background and significance were presented; secondly data mining, Web mining,association rules, clustering technology, personalized recommendation technology and Vector Space Model (VSM) were introduced. Focusing on the limitations of Vector Space Model (VSM), this dissertation introduces the user interest model based on keywords dependency relationship.The model represents semantic relation through a graphic structure in which the keywords are defined as nodes and the relationship among keywords is defined as lines among nodes.This dissertation also introduces the calculation ideas and methods of keywords (nodes) weights and dependency relationship (lines among nodes) weights.Based on the theoretical content, we developed user webpage browsing interest mining system prototype. The system prototype's structure is four-layered, and the four layers are Layer User Interface (Layer representation), Layer Data Acquisition, Layer Data Pretreatment and Layer Logic (Analysis Function).At last, this dissertation proved the validity of the modeling method through a specific experiment.This dissertation aims to propose one new user interest modeling method to improve the present technology, thereby to provide novel ideas and methods for the research of Web mining and personalized service.