An empirical study of credit card fraud detection system in Taiwan

In the automatic workflow of credit card risk management, the most complex parts are the examine system and the detection system. Because of the slack institution of credit card in Taiwan, the detection system is still the major role in credit card risk management. Even though the learn by example alarming system can reach the effect in monitory, but its hit ratio is still pool, therefore our study tries to use neural network detection system to arise the effect. The result shows that the neural network system can provide the highest accuracy.