Real-time accurate recommendation of large-scale recommender systems is a challenging task. Matrix factorization (MF),.as one of the most accurate and scalable techniques to predict missing ratings, has become popular in the collaborative filtering (