Irina Beregovskaya, Mikhail Koroteev
Recommender systems are one of the most applied methods in machine learning and find applications in many areas, ranging from economics to the Internet of things. This article provides a general overview of modern approaches to recommender system design using clustering as a preliminary step to improve overall performance. Using clustering can address several known issues in recommendation systems, including increasing the diversity, consistency, and reliability of recommendations; the data sparsity of user-preference matrices; and changes in user preferences over time. This work will be useful for both beginners in the field of recommender systems and specialists in related fields that are interested in examining the applicability of recommender systems. This review is focused on the analysis of the scientific literature on the topics of recommender systems and clustering models that have appeared in recent years and contains a representative list of the literature for the further exploration of this topic. In the first part, a brief introduction to the so-called classic or traditional recommendation algorithms is given, along with an overview of the clustering problem.