Research On User Interest Model Based On Web Browsing Behaviors

With the rapid development of internet, web information is growing exponentially, which leads to people need to spend a lot of time to find the information they interested. Personalized recommendation is a very effective method to solve the problem of information overload. User interest model is the core component of personalized recommendation, a reliable interest model can better improve the user’s internet experience and information efficiency, so interest model become the hot topic of the academic and IT industry. This dissertation is mainly focused on the evaluation of the user’s interest and the representation and update mechanism for user interest model based on text information processing. The aim is to establish effective interest model and design the news recommendation system of practical application significance at the same time. The main research contents are as follows:(1)Aiming at the deficiency of the evaluation of the user’s interest when building user interest model, this dissertation brings out a quantification method of interest rate based on hybrid behaviors, which highlights the specificity of page browsing time and evaluates interest rate based on other browsing behaviors in the case of abnormal page browsing time. The interest rate is applied to the traditional VSM model to make the extended interest model bring the content and behavior.(2)Aiming at the large storage and increased amount of calculation of system after the establishment of the system, this dissertation brings out time segmentation mechanism to reduce the amount of calculation and storage of interest model. In addition, this dissertation brings out interest drift mechanism to reflect the change of user interest based on time segmentation mechanism.(3)A news recommendation system is built up according to the proposed user interest model. 2000 articles are collected from internet and used as the data set. The system collects users’ browsing behavior continuously, analyzes the user interest, updates the user interest model, and then recommends articles to user.The experimental results show the rationality and validity of the model based on hybrid behaviors interest rate.