Detecting Change in Customer Behaviour by Means of Concept Drift

The study of consumer behavior is the subject of widespread interest. Identifying customer preferences is one of the key areas. The authors have tried to solve the problem by using different tools and approaches. The main feature of these studies is that the analyst must take into account the dynamics of the data, so a loose definition of Concept Drift appeared. The purpose of this work is to create an information system based on the theory of Concept Drift, which would be able to help marketers of retail trade system to make operational and strategic decisions. To achieve this, a set of tools such as RStudio, Neurosimulator 5.0 and MS Visual Studio have been employed. The system allows to forecast the customer preferences for a few next months and show what factors have the most influence on their choice. This system will undoubtedly facilitate the job of the marketer in the field of data analysis and provide a competitive advantage for the business.