Validation and Enrichment Profiles of Users of Social Networks

In recent years, the popularity of social networks in the world is continuously growing. This process is directly related to the phenomenon of socialization of user data, information becomes publicly available. This fact gives unprecedented opportunities to solving research and business problems. Nevertheless, initially only a certain personal information is known about the users of the network, which is often not filled in completely or is intentionally indicated incorrectly. As a rule, these data are not enough to form an accurate and correct portrait of the user of the social network. The purpose of this work is to study and develop algorithms of validation and enrichment of the attributes of the social network VKontakte. Within the framework of the work, a number of tasks are solved, such as the collection and processing of data from the social network, the applicability of the machine learning methods to achieve the goal, the search and selection of optimal solutions. The results of the work are the developed models of enrichment and validation of fields gender and age of user profiles. The age determination model has the following values of quality metrics: MSE=21.55, MAE=2.13, the gender determination model has accuracy=0.896 and F-score=0.904. Keywords: social networks, user data, machine learning.