Chapter 19 - Neural hybrid recommendation based on GMF and hybrid MLP

Abstract

Recently, recommender systems have become exceptionally widespread, providing users with personalized appropriate items or service recommendations. On the other hand, deep learning techniques achieved great success in many research areas including computer vision, image processing, pattern recognition, and natural language processing. The application of deep learning in recommender systems have been explored with promising results. This chapter focuses on the use of deep neural networks for learning the interaction function from data. We propose a hybrid recommendation approach that combines collaborative filtering (CF) and content-based filtering (CBF) in an architecture based on two models: (1) generalized matrix factorization (GMF) and (2) hybrid multilayer perceptron (HybMLP). Extensive experiments on two well-known datasets MovieLens-1M and Yelp show significant improvements in our approach compared to the existing methods, especially for the cold start situation. These empirical evaluations show that the use of deeper layers of neural networks offers better recommendation accuracy.