Challenges and Perspectives of Recommender Systems in Vending Machines

Abstract

This paper describes the features of using recommendation algorithms for vending machines. A vending machine is an automated machine with a significantly limited range of products (snacks, beverages, cigarettes), and it also does not provide any additional information about the customer, which imposes some limitations on the predicting relevant products. In order to increase the demand for vending products and make a profit, this paper proposes to consider how modern recommendation algorithms can be applied in this area. One of the important stages in building recommendation system is preparing data for analysis, so this paper describes an approach for grouping data by sales intensity and also considers the set cover problem.