Clustering of participants in the MaxBonus loyalty system using Kohonen’s self-organizing maps

The purpose of this article is to investigate customer clustering based on big data of the consumer loyalty system. The object of the research is the retail chains that are the clients of the Maxbonus system. Self-organizing Kohonen maps, implemented by the selforgmap function of the Deep Learning Toolbox module of the Matlab system, were used as a clustering tool. As a result of clustering 990 customers, classes invariant to the partition were identified, and it was shown that their number stabilizes with an increase in the number of classes in the Kohonen self-organizing map. The average indicators of representatives of each class have been determined. This result indicates the efficiency of the approach for clustering customers of customer loyalty systems. The prospect of the work is to include in the number of input parameters clustering the volume of purchases by product categories. Also promising is the transition to clustering retail chains participating in the Maxbonus system.