Community cooperation in recommender systems

Recommender systems have been widely used in commercial and research oriented systems. In this paper, we propose to develop an intelligent, Internet-based movie recommender system, to help moviegoers choose movies. Our system, COOP-R uses a hybrid recommendation technique based on collaborative and content based filtering. As opposed to previous work using the neighbourhood paradigm, our collaborative filtering approach uses the community of chosen friends, thus allowing better control of the overall recommendation, and takes advantage of the influential and popular friends that have some authority in the movie community. We believe that our system allows more social interaction among moviegoers. We discuss the design and implementation of COOP-R, report on its performance evaluation, and present a comparative study to traditional collaborative filtering systems. Our results indicate that COOP-R exhibits a better precision when compared to traditional collaborative based system