Providing a Model to Analyses of Customer’s Behavior and A Rating Model in Banking Services Using Data Mining and Neural Network

Abstract

The banking industry is one of the sensitive and dependent industries. Because the main part of the working capital of banks is provided through the investment of customers. Therefore, the main goal of banks is to attract new customers and retain old customers by providing optimal services and improving customer relationship management. Therefore, in this study, using data related to private banking services in Iran, we used a combination of data mining methods and artificial neural network to evaluate the performance of each in predicting customer behavior from the point of view of churn or loyalty. So, behavior of 414 customers which have accounted in private banks in Iran was evaluated based on 17 services indicators between 2014 and 2019 using SVM, ANN, SVM & ANN, SVM & FPA and ANN & FPA algorithms. The results show that the combined ANN & FPA algorithm with 94.79% accuracy has the highest performance in predicting customer’s behavior.