Abstract Developing new marketing methods to understand customers’ needs is of vital importance in a highly competitive market. The customer segmentation aims to divide the customers into disjoint groups of self-similar individuals to apply specific marketing strategies to each group. Assigning each customer to the most appropriate group is a challenging task in large datasets. In this paper, we propose a graph-based approach to the customer segmentation problem by representing each customer as a vertex and two vertices are connected by a weighted edge according to the recency, frequency and monetary values associated with each customer. The main advantage of this graph representation is that we design a procedure to build a reduced graph with fewer vertices and edges, such that clustering this smaller graph is equivalent to clustering the customer dataset in the original graph. It clearly provides an advantage in dealing with large instances, and we demonstrate this equivalence by proving two theorems. We applied our methodology to synthetic datasets randomly generated and to a real-world dataset available in the UCI Machine Learning repository. In all cases, we compared the results of our method to the results of clustering algorithms of the literature in terms of the silhouette index. In particular, we considered the results for the real-world dataset reported in Rungruang et al., Expert Systems With Applications 237 (2024) 121449 . The computational results show that our method outperformed the other methods in the majority of the instances.
Filho et al. (Wed,) studied this question.