Enhanced collaborative filtering for rating prediction on movies

Abstract

Previous rating system averaged all of each user’s rating that was represented in a numerical value. And it contained limitation that rating after computed in the system was not accurate to all the users. In this paper, we raised the problems that previous rating system had and established hypothetical problems that caused poor accuracy. And we made filtering algorithm for rating system to overcome existing problems by considering enhanced collaborative filtering after we examine related works briefly. We presented the result of experiments that proved our approach performed better than previous rating system and some recommendation systems. We also proposed some future works to improve our system, capability of extending to other area and limitations that were not solved in our studies.