A novel collaborative filtering recommendation algorithm based on user location

With information technology and the Internet developing fast, people gradually walk out of the time of information deficient and enter the era of information overload. Whether information consumers or information producers are faced with big challenge: how to obtain or sell the information. Recommendation system is a key to this problem. Traditional recommendation system focuses on connecting user interest and items, and recommends items which match user interest. However, all these algorithms ignore the context which users are in. In terms of this problem, this paper presents a novel collaborative filtering recommendation algorithm based on user location context. Firstly, this algorithm defines user location attenuation function to calculate the relations between user locations, then combines this function with traditional Pearson similarity method to get similarity between users, finally, uses the traditional collaborative filtering recommendation algorithm to realize preference prediction and recommendation. Experiments show that this algorithm which has location information taken into account can improve recommendation quality for traditional collaborative filtering recommendation algorithms.