Методи оцінки якості рекомендаційних систем

Abstract

This article describes some types of recommendation systems and methods for their evaluation. Recommendation systems are an integral part of services that provide the user with a choice among a variety of alternatives. Using a service, most people do not even think about what gives them more and more new recommendations. These systems greatly facilitate the use of such services, since they provide recommendations based on user information. Particular attention is paid in the article to methods for assessing the quality of reference systems. Evaluation of advisory systems is necessary in order to understand how effective this or that system is. The effectiveness of the system is very important, since the more effective it is, the better and more qualitative recommendations it provides. The article is divided into two semantic parts: the first one deals with what is the recommendation system, which types of recommendation systems exist, the advantages and disadvantages of each species are also described, the second describes methods for evaluating the recommendation systems. In the second semantic part, two methods for evaluating recommendatory systems are described: the method for assessing the quality of recommendations and the method for evaluating the quality of work of algorithms.