NEURAL FRAUD DETECTION IN CREDIT CARD OPERATIONS

This paper presents an on-line system for fraud detection of credit ca rd operations based on a neural classifier. Since it is installed in a transactional hub for operation distribution, and not on a card-issui ng institution, it acts solely on the information of the operation to be rated and of its immediate previous history, and not on historic da tabases of past cardholder activities. Among the main characteristics of credit card traffic are the great imbalance between proper and frau dulent operations, and a great degree of mixing between both. To ensur e proper model construction, a nonlinear version of Fisher's discrimin ant analysis, which adequately separates a good proportion of fraudule nt operations away from other closer to normal traffic, has been used. The system is fully operational and currently handles more than 12 mi llion operations per year with very satisfactory results.