Reading behaviour based user interests model and its application in recommender system

User interests model is a central issue for information recommender systems. Basically, a user interests model aims to describe a user's interests precisely in order to improve the quality of information access. However there are three challenges for developing a user interests model, i.e. initial user information acquisition, user profile representation and description of user interests evolution. In this paper, we try to solve the problems by employing tags. We propose a user interests model which creates users' profiles based on tags named interest tags and we implement a mechanism to dynamically update users' profiles through analyzing their reading behaviour. The proposed user interests model can distinguish between long-term interests and short-term interests. We also present how the model can be integrated into a recommender system to help analyze users' interests. We conduct experiments on massive text information collected from the Web and the results show the effectiveness of our model.