Based On Web Page Interest Degree Of User Interest Model System Studies

Entering the age of Web2.0, the amount of information on the internet grows rapidly with the emergence of blog, SNS, microblog, etc. It becomes difficult for us to find useful information among such large scale of information. Search engineer is a common way for us to do this, which filters the webpages with users’ input keywords. After amazon releases good recommendation services, people find that intelligent recommendation can bring great success. The academia and IT industry gradually focus on the research on intelligent and personalized recommendation.User interest model is a mathematical representation model for users’ different interest, which is one of the methods to realize personalized recommendation. It is obtained by analyzing users’ browsing content and behavior to extract content characteristics and interest rate. After setting up the interest model, content can be intelligent recommended to users by comparing the matches between the content and the interest model. In the aspect of content characteristics extraction, Vector Space Model is adopted. On the other hand, an interest rate metrics algorithm is proposed to measure users’ interest towards the content. In the updating procedure of users’ interest model, the shifting of users’ interest is often out of consideration in many cases, which will affect the accuracy of the recommendation, so time slicing mechanism and time decaying mechanism are imported to improve the overall system performance.An experiment system is built up according to the proposed user interest model to verify the algorithms.2000articles on the themes of the Expo, Manchester United are collected from Sina.com and used as the experiment data. The system collects users’ browsing behavior, calculate users’ interest model and then recommend new articles to them. Experimental results show that the interest rate metrics is effective and the according interest model is confirmed.