Design of personalized recommendation system based on LBS in mobile classroom project

Collaborative filtering is one of the most successful approaches to building recommendation system. However, it still has some known disadvantages. One of them is called cold start problem caused by lack of user´s historical data. Another problem appeared because mobile technology develops in high speed. The phenomenon that users change their taste according to their position quickly in mobile environment is more and more common. The changeable preference is called short-term interest. But the collaborative filtering algorithm usually ignores this short-term interest. It leads to decrease of accuracy of recommendation. In this paper, a design of personalized recommendation system based on LBS for the Mobile Classroom Project is proposed. Considering the actual conditions, this system solved the problem of cold-start using clustering method and some other solutions. Also, it focus on user´s short-term interest to a certain extent. Compared with the traditional collaborative filtering, it works better.