MODERN RECOMMENDER SYSTEMS’ REVIEW

Nikolaev Konstantin Sergeevich

Nowadays users have to sort through large amount of diverse information in order to find the most relevant one. Recommender systems improve the efficiency of this process; they deal with huge number of items and offer content portion that users would be interested in. The paper describes one of the most popular recommender systems used for production purposes and of great practical importance in industry. This work aims to review the existing researches and analyze models’ features presented there. The contributions of the paper are the following: analysis of mentioned recommender systems, describing their application areas, strengths and weaknesses. The conclusion focuses on the practical value of explorations and opportunities for other industries.

Keywords:recommender systems, recurrent neural networks, reinforcement learning, cold start, machine learning