Abstract
During the last two decades credit cards have became one of the main ways for accomplishing financial transactions. The number of credit card owners have increased rapidly. Unfortunately, at the same time the cases where the owners cannot fulfil their obligations to the banks have also been increased. This fact forced credit institutions and banks to search for methodologies that will allow them to accurately evaluate the credibility of each credit card applicant. Multi‐criteria decision aid methods as well as machine learning algorithms can be used to accomplish this task. The present paper proposes a new intelligent decision support system for credit card evaluation, based on a machine‐learning algorithm, namely the Composite Rule Induction System and the Rough Sets. The major advantage of the algorithm and the system is the incorporation of qualitative variables, which have an essential role in credit card evaluation. The system is applied on a real case study concerning credit card evaluation by a leading Greek commercial bank and the obtained results are compared to the results of multi‐criteria decision aid methods as well as other machine learning algorithms. Copyright © 2003 John Wiley & Sons, Ltd.