DETECTION OF COMMUNITIES IN SOCIAL NETWORKS

Abstract

One of the main threats is malicious programs (bots), fake accounts capable of imitating human behavior. At the moment, bots create a lot of problems, both for ordinary users and for those who use social networks to conduct a marketing campaign or conduct social research. Using bot profiles in social networks greatly distorts information about the real benefits and interests of portal users. Therefore, it is necessary to determine which users of the social network are programmed, and to be able to divide the flow of data into that generated by bots and by humans. The threatening scale of the use of social bots requires the creation of effective algorithms for their detection. Internet platforms and social services themselves are not too concerned about this problem. As a result, both ordinary users who organize various communities and companies that promote goods, brands, and services through social networks suffer. Thus, the task of recognizing malicious accounts in social networks and combating them remains relevant in the issue of cyber security. The solution to the problem will be the development of network analysis methods that are designed to identify and classify communities in social networks, assess their connectivity, degree of trust, as well as develop effective algorithms for detecting malicious accounts. The purpose of the work is to develop an improved algorithm for detecting malicious accounts in social networks, which is based on the study of modern methods.