hybrid-movie-recsys

Movie recommender system build with hybrid methode. Taking advantages from Collaborative Filtering and Content-based Filtering. So the recommended movie is the popular one but still special for each user since it is similar with user's preference.

Dataset used for this project is a public dataset from Movie Lens (http://grouplens.org/datasets/).

Collaborative Filtering

Build using Matrix Factorization with Stochastic Gradient Descent

Content-based Filtering

Similarities from each movie calculated using cosine similarity.