A multi-attribute data mining model for rule extraction and service operations benchmarking

Purpose Customer differences and similarities play a crucial role in service operations, industries need to develop various strategies for different customer types. This study aims understand the behavioral pattern of customers banking industry by proposing hybrid data mining approach with rule extraction operation benchmarking. Design/methodology/approach The authors analyze identify best using modified recency, frequency monetary (RFM) model K -means clustering. number clusters is determined two-step quality analysis based on Silhouette, Davies–Bouldin Calinski–Harabasz indices evaluation distance from average solution (EDAS). best–worst method (BWM) total area orthogonal vectors (TAOV) are used next sort clusters. Finally, associative rules Apriori algorithm derive customers' behavior patterns. Findings As result implementing proposed financial industry, were segmented ranked into six analyzing 20,000 records. Furthermore, frequent patterns recognized demographic characteristics transactions customers. Thus, types classified as highly loyal, high-interacting, low-interacting missing Eventually, appropriate interacting each type proposed. Originality/value propose novel multi-attribute operations benchmarking combining tools multilayer decision-making approach. has been implemented large-scale problem services industry.