Business Application of Clustering Models for Social Networks

The development of information technology has already simplified communication between people, and this process is not going to stop. Every day, each of us is in contact with a huge number of people, directly or indirectly. Interactions between people form a social network. Currently, this term is almost always understood as a platform on the Internet, although this term covers much more cases of people's interaction in society. Analysis of social networks can tell a lot about the characteristics of its elements, as well as their interaction with other elements of this network. Clustering a social network graph can be used by different companies to identify clusters of customers and their characteristics for more targeted provision of services or sales of goods. This is one of the most effective ways of obtaining information about customers, requiring only a certain processing power and a relatively small amount of time. However, there is still no understanding of which algorithm is best used in this or that case, which forces researchers to apply inefficient algorithms or to spend time comparing the effectiveness of different approaches to clustering. This study aims to develop recommendations for the selection of algorithms of recommendations, depending on the type of task facing the researcher. The results of the work include an overview of existing clustering methods, selected classes of tasks that are encountered in real life, as well as recommendations on the choice of algorithms in this or that case, based on their characteristics. Also, these recommendations have been practically tested on a real business objective, with a full description of the need to solve this problem, the process of analyzing the social network, as well as its results and their interpretation.