ABSTRACTS OF ARTICLES OF THE JOURNAL "INFORMATION TECHNOLOGIES".
No. 9. Vol. 28. 2022

DOI: 10.17587/it.28.456-464

V. G. Lyalikova, Assistant Professor, M. M. Bezryadin, Assistant Professor,
Voronezh State University, Voronezh, 394018, Russian Federation

Review of Modern Recommendation Systems

An analysis was made of modern approaches to building recommender systems, such as content-based filtering, knowledge-based filtering, collaborative filtering, and hybrid recommender systems. The advantages and disadvantages of each of the systems are revealed. The main ones are the problem of a cold start, poor predictions for atypical users, computational resource intensity, and trivial predictions. The features of the use of recommender systems in the most famous companies such as Amazon, YouDo, Facebook, Youtube, Yandex.Zen are considered.
Keywords: recommender system, collaborative filtering, hybrid filtering, knowledge-based filtering, content-based filtering, singular value decomposition, estimate accuracy of recommender systems

P. 456–464

To the contents