This document describes a movie recommendation system that uses both collaborative filtering and content-based filtering approaches. It summarizes that collaborative filtering makes recommendations based on ratings from similar users, while content-based filtering recommends movies similar in attributes like director, cast, etc. It then describes using a hybrid approach that learns item similarities from user ratings to combine the two methods. The system is evaluated on a MovieLens dataset and achieves a mean square error of 1.076 for rating predictions.