Combining Browsing Behaviors and Page Contents for Finding User Interests

This paper proposes a system for finding a user's interests based on his browsing behaviors and the contents of his visited pages. An advanced client browser plug-in is implemented to track the user browsing behaviors and collect the information about the web pages that he has viewed. We develop a user-interest model in which user interests can be inferred by clustering and summarization the viewed page contents. The corresponding degree of his interest can be calculated based on his browsing behaviors and histories. The calculation for the interested degree is based on Gaussian process regression model which captures the relationship between a user's browsing behaviors and his interest to a web page. Experiments show that the system can find the user interests automatically and dynamically.