The problems of identifying individual customers are formulated based on the analysis of large amounts of data on cash receipts in a large supermarket. Models of behavior of various categories of individual customers in the supermarket are developed. A computational algorithm is proposed for solving the problems of identifying individual customers using panel data from cash receipts. The algorithm is universal, since it does not use any personal data about the buyer, but is based on an analysis of only buying activity, calculated on the basis of panel data on cash receipts. The algorithm allows you to identify groups of customers, as well as with a certain probability, an individual customer. As an example of the application of the developed models and computational algorithms, commodity checks from the X5Retail Group company supermarket chain for a certain period of time were used.
Димитриенко Ю.И., Котельникова А.В. Задачи идентификации индивидуальных покупателей на основе анализа больших объемов панельных данных о кассовых чеках. Математическое моделирование и численные методы, 2019, № 4, с. 100–116.