Recent Developments in Recommender Systems

With greater penetration of online services, the use of recommender systems to predict users' propensity for continuous engagement becomes crucial in ensuring maximum revenue. There are many challenges, such as the cold start problem and data sparsity, that are continuously being addressed by a myriad of techniques in recommender systems. This paper provides insights into the trends of the techniques used for recommender systems and the challenges they address. With the insights; deep learning, matrix factorization or a combination of both can be used in addressing the data sparsity challenge.