A Hybrid Collaborative Recommendation System Based On Matrix Factorization And Deep Neural Network

Abstract

The paper explores a modified recommender system that is established based on the combination of matrix factorization and deep neural network that work on the implicit feedbacks of users and also auxiliary information of both users and items. Recent works show the effectiveness of deep neural network on recommendation systems. Proposed models aim at discovering additional relationships by using auxiliary information to explore the internal relationship between users and also the relationships of items among themselves. Experiments show 0.5556 and 0.8036 in NDCG and HR with the model which is an improvement compared to other popular collaborative filtering methods.