Discovering Hidden Patterns in Loan Reimbursement

Loans are the major resources at banks. However, in some cases the cost that they incur to banks soar and finally makes them detrimental, as a result of irregular or delaying reimbursement or not paying at all. Due to the low wage rates in Iranian banks and the Central Bank of Iran (CBI) regulations in determining interest rates for deposits and loans, banks are becoming more and more dependent to the loans and their related profits. Therefore, banks have to look for customers with low risk for punctual payment. According to defect loan reimbursement in past years, banks have to specify severe prerequisites and limited contracts in granting loans to their customers. Contravening banking regulations and lack of consistent customers' accreditation banks are getting into heavy losses. Evaluating situations of the granted loans in EN Bank of Iran during a six-month period, based upon the profiles and loans history and the trend of payments useful patterns are discovered; designing a practical model of loan payment in Iran, the future default or failure to regain the granted loans is predicted and sensible methods of granting loans in Iran are developed. In order to extract hidden patterns in data statistical methods and data mining tools with focus on decision tree techniques are applied.