Application and Analysis of the Effectiveness of Clustering Methods in Identifying Communities in Social Media

Graduation qualification work is executed by Valeria Annenkova, a 2nd year student of the group №IAUP-22-1 in the direction of training 38.04.05 Business Informatics educational programme "Information analytics in enterprise management". Theme: "Application and Analysis of the Effectiveness of Clustering Methods in Identifying Communities in Social Media". Supervisor: Elena Borisovna Zamyatina, lecturer of the Department of Information Technologies in Business, Candidate of Physical and Mathematical Sciences. Scope of work: 62 pages The work consists of 3 chapters. In the first chapter, a review of scientific literature was conducted, which included definitions of basic concepts, study of the essence and features of social networks, consideration of key concepts of social network analysis, as well as methods of identifying communities based on description. The second chapter is devoted to the compilation of the database. To conduct the research, a programme for collecting data on communities was developed, which was used to compile a database containing information on 20,000 communities of the VK social network. The database was divided into two samples: test and non-test. Using a neural network, the communities' descriptions were used to identify their topics. The final stage of the study was clustering of communities using various cluster analysis tools. As a result, groups of communities united by thematic features were obtained. The results of the study allow us to draw conclusions about the degree of effectiveness of various clustering methods in social media analysis and provide recommendations for their application in the practice of social network data analysis.