Rating the Actual Customers of Banks based on Credit Risk using Multiple Criteria Decision Making and Artificial Intelligence Hyperbolic Regression

This study wants to investigate the rating of the actual customers of banks based on credit risk using multiple criteria decision making and artificial intelligence hyperbolic regression. This is an applied research. The statistical population of the study includes the credit customers of Agriculture Bank in west branches of Mazandaran province, Iran in 2012-2016. A total of 100 cases have been evaluated. AHP method has been used in the case of elites' comments on the prioritized seven key factors using the corresponding weighting matrix. To model the classes of creditworthy and non-creditworthy customers and to predict the appropriate model based on the evidence in the customer's credit file, artificial intelligence hyperbolic regression has been employed. Using AHP method, the rating include customer revenue, credit in the market, customers' job, the duration of the relationship with the bank, the type of collateral, the value of collateral, average account balance to facilitate the credit risk of the actual customers, respectively. Using artificial intelligence hyperbolic regression, prioritization is based on the amount of credit in the market, customer revenue, the value of collateral, the duration of the relationship with the bank, the type of collateral and customers' job.