Modern Directions of Research in the Field of Recommender Systems

The constant growth in the volume of content generated by electronic services has caused the problem of finding the necessary information in a limited time. Recommender systems are a useful tool that, among other things, solves the problem of speeding up the search for the necessary information. Web applications make extensive use of recommender systems to provide users with relevant content based on their preferences or interests, thus making it easier for users to access the information they seek. At the same time, the presence of a business effect from the introduction of such systems also shows the importance of their development and operation, but at the same time, the question of the degree of influence of algorithmic improvements in recommendation systems on target business metrics remains open. In various domains (recommendations of music, books, video content, product recommendations in online stores and marketplaces, etc.) various types of recommender systems are used, which are based on a wide range of technologies, including machine learning models and computational algorithms. The purpose of this work is to identify the main modern directions of research in the field of recommender systems, as well as a description of unsolved problems and challenges of the field.

recommender system, collaborative filtering, content-based filtering, cold start, machine learning