Customer Value Analysis in Banks Using Data Mining and Fuzzy Analytic Hierarchy Processes

The customer value is a starting point for customer relationship management to realize and evaluate the value of the customers for every organization. Value is the base of all of the marketing activities because all parties expect to get value in their transactions. This study focusses on the Tejarat Bank branches in Iran and systematically integrates several data mining techniques and management issues in order to analyze customer value. First, we applied the fuzzy analytic hierarchy process for weighing attributes and then imported the demographic, frequency, money and trust attributes to the K-means. Using the proposed scoring model, we created the customer value pyramid. Finally, in order to analyze the obtained pyramid classes and perform the learning process from the data, we utilized a decision tree, support vector machine, random forest classification techniques, along with six chosen characteristics and introduced the most appropriate model according to the applied attributes. Including learning patterns for classifying new customers, with respect to the importance of defined factors, would be useful in this case. According to results, we achieved a model attribute with the best characteristics in accuracy, precision, recall, F-measure and class error measures.