A New Hybrid Method for Predicting Recommendations for Collaborative Recommender Systems

In this article analyzes the current state of models and methods for building recommender systems. The main classes of problems solving recommender systems are identified. The features of the application of the method of collaborative (general) filtering are shown. A mixed numerical-categorical clustering method has been developed to search for user groups using numerical rating and demographic characteristics of users; a hybrid method for searching user groups has been developed using the sparsity coefficient of the user-subject matrix.