Clustering algorithm for electronic services customers: A case study of the banking industry

Today, recognizing and retaining customers is one of the major challenges of customer-oriented organizations, especially in the field of banking, which has attracted the attention of many researchers. With the increasing growth of customers and the use of electronic devices that have led to the production of large volumes of data, customer behavior analysis can be considered as a competitive factor for them. In this paper, considering the varied density and data growth that leads to computational overhead, a combined approach is used of the RFM method , density-based clustering algorithm and Map-Reduce( which is an efficient and low-cost framework for running synchronous algorithms.) it is used. The results show that the proposed algorithm is more accurate than VDMR-DBSCAN. Also, the output of the algorithm is in the form of 5 clusters, the results of which can help managers in identifying valuable customers, and This method leads to increased revenue and reduced unnecessary costs that occur due to lack of recognition and incorrect segmentation of customers.