Choice of a method for segmenting the client base in the conditions of the information asymmetry

The subject of the research is the approaches to clustering goods and services in relation to behavioral activity, as well as the development of methods for clustering customers based on the results of their behavioral activity. An important feature of the technique is the ability to work not only with numeric and categorical features, as well as the ability to determine the number of clusters. As a result of the study, the following tasks were solved: various methods of classification and clustering of clients according to the characteristics of behavior are considered and the main categories are explained; substantiated the need for clustering customers based not only on numerical, but also categorical variables, which makes it possible, if necessary, to switch from methods of controlled classification (for example, RFM analysis) to methods of hierarchical clustering; investigated and proposed for use metrics of distances between elements within a cluster, as well as indicators of combining elements into a cluster and separating individual clusters among themselvesthe methods of hierarchical clustering with the use of groups with broken symmetry are investigated, the choice of methods of hierarchical clustering for various cases is justified; the problem of automated estimation of the number of clusters is being solved; a clustering technique based on numerical and categorical features using hierarchical clustering methods is proposed. Information asymmetry is indicated as the main reason that generates changes in the behavioral activity of customers in the work, which is expressed in different degrees of awareness of the groups of sellers and groups of buyers-users of the product about the state of the market. As a result of dividing consumers of products into clusters, it is possible to establish the reasons for the violation of symmetry in the behavior of groups of consumers of goods and services.