Recommender systems based on matrix factorization and the properties of inferred social networks

Recommender Systems (RS) are a vital part of companies with an active participation on the web. These companies require strategies that allow them to take advantage of product ratings from users in order to provide future recommendations to other users. In the last decade, several algorithms have been developed for movie recommendation, with Matrix Factorization algorithm being one of the most popular algorithms. Our approach is to evaluate the performance of this recommendation algorithm in scenarios where underlying social networks, which characterize certain types of interactions between users, can be inferred. In particular, the MovieLens dataset is used, which consists of approximately 100,000 ratings by 671 users on 9066 movies, during the period from 29 March 1996 to 24 September 2018.