Abstract
This article is devoted to the study of modern methods for modeling the interests social network users, as well as the development and implementation of our own method that allows you to model interests in the form of a set of terms and categories that are understandable to any user. The developed method is based on determining the topics of user messages using topic modeling algorithms and regularizing the resulting distribution, taking into account the user's connections in the social graph. Also, a method was developed that allows you to automatically match the received topics with a set of concepts and categories from the WikiPedia encyclopedia. To implement the developed method, a system for modeling the interests of users of the Twitter social network has been built. The article describes the architecture of the implemented system, and also conducted experimental testing of the constructed system on a test dataset and describes the results obtained.