Top-N Movie Recommendations Using Machine Learning

Abstract

In this paper we explore recommendation algorithms using machine learning. Specifically, our goal is to predict top-N movie recommendations using different models to give us predicted ratings for a movie. As recommendation research has shown there are several metrics to measure when evaluating top-N recommendations such as accuracy (RMSE/MAE), hit rate, coverage, and diversity. In this research, we are focusing on rating ranking and movie genre coverage. We utilize collaborative filtering, content filtering, hybrid recommenders, and finally include neural nets to generate predictions.