MovieReSys

This is my bachelor thesis project, a movie recommender system, implemented in Java. Slope-one algorithm is modified to try to forecast user preference considering the effect of time. Item-based, User-based collaborative filtering and the original Slope-one algorithms are compared with the modified Slope-one. With the MovieLens database, it shown that the modified Slope-one has a better prediction. However, the cost of time of the modified Slope-one is much larger then others.