CREDIT SCORING IN IRAN WITH CSVM AND LOGIT MODEL

Features and specifications such as The status of the client's existing checking account, The duration of the credit period in months, The client's credit history, The purpose for the credit, The credit amount requested, The client's savings account/bonds balance and ...., And the different methods used to determine good customers from the bad accounts, but the large volume of outstanding bank loans. This paper investigates the practice of credit scoring and introduces the use of the clustered support vector machine (CSVM) for credit scorecard development in Iran. Accordingly, with a sample of 3000, this study shows that the CSVM can achieve comparable levels of classification performance while remaining relatively cheap computationally. Classifications by model CSVM factors affecting lack of timely repayment of credit facilities include: The client's housing arrangements (i.e. own their home, rent, or live for free), the applicant's monthly income, the duration of the credit period in months, How to refund, the credit amount requested.