×

Bank customer classification algorithm based on improved \(K\)-means clustering. (Chinese. English summary) Zbl 1424.62112

Summary: In order to improve the accuracy of bank customer classification and maximize the bank’s revenue, this paper proposed a bank customer classification algorithm based on improved \(K\)-means clustering. The algorithm defined the maximum similarity (AMS) between classes and determined the optimal number of clusters based on AMS. When the calculated current AMS value was smaller than the previous AMS value, it selected the initial clustering center according to the distance principle; when the calculated current AMS value was larger than the previous AMS value, it considered the cluster center corresponding to the minimum AMS value as the initial cluster center. According to the optimal number of clusters and initial clustering center, the algorithm achieved bank customer segmentation. The simulation results show that the proposed algorithm can jump out of the local optimum and improve the accuracy of customer classification.

MSC:

62H30 Classification and discrimination; cluster analysis (statistical aspects)
62P05 Applications of statistics to actuarial sciences and financial mathematics
PDFBibTeX XMLCite
Full Text: DOI