A Quantized Analyzing Method for Finding a User's Interested Web Pages Based on the User's Browsing Actions and Query Words

How to capture a user's interested Web pages (IWPs) is very important in a Web personalized service. In this paper, a method to capture a user's IWPs by combining a quantized analysis of the user's browsing actions and query words has been presented. To distinguish the differences among users and at different time of a user, a set of parameters have been introduced when calculating the contribution of browsing actions to the interest value on a Web page for a user. The parameters are adjusted periodically according to the current Web pages browsed by the user. The interest value is used to adjust if a Web page is interested by the user. Then, the method has been improved by introducing the analysis of the user's query words in order to determine more precisely if a border interested page is really an interested one. The method has been implemented using the technologies of Hook and BHO. Experiments have been given to show its feasibility and validity.