Deep Time-Aware Matrix Factorization

The appearance of recommendation system solves the problem of information overload. Traditional recommendation systems generally consider the preferences of users, but ignore external conditions, such as the timeliness and popularity of goods.In this experiment, the time factor is added to form a triple, like User-Item-Time, and the neural network is used for training. Compared with the matrix factorization experiment which integrates time factor, the prediction effect is better when the movie popularity is integrated into the recommendation.