Identifying interests of users based on social network profile analysis

With the development of social networks and the increase their role in people's lives, the role of recommendation systems used in social networks is increasing. The methods of news recommendations used in most social networks to the user use either the clearly indicated interests of the user (subscription to the community) or only the explicit connection between the users (their friendship), disregarding the interests of the user. In this paper an alternative method is proposed for finding communities of interest to the user to offer him content that he has not subscribed to yet. The paper describes an algorithm for constructing an interest graph based on data from the social network VKontakte (vk.com) and an algorithm for creating a recommender system that predicts the interest of users in a particular content based on the graph of interests. By using the graph of interests and various metrics, the system takes into account various prerequisites for the user's interest, which do not take into account the usual methods. The result of the work is a system of building recommendations, which showed a good generalizing ability on new data.