Resumo

The prevention of credit card fraud is an important application for forecasting and computational intelligence approaches. Computational intelligence is an association of computational methodologies that includes as its main member the neural networks, fuzzy systems, evolutionary computation and probabilistic computation approaches. This paper presents an efficient system based on computational intelligence methodologgies for fraud detection of real credit card data operations in e-commerce transactions. The new proposed method called Fcontrol  of fraud detection and classification integrates concepts of neural networks, fuzzy systems, and evolutionary computation. The proposed method provides quality solutions in terms of efficiency and forecasting successful. The Fcontrol  system has been installed on an Intel Pentium IV 2.4 MHz (bi-processor and RAM of 4 Gbytes) at Ciashop E-Commerce company and is currently in use for fraud detection for 250 companies with e-commerce services.