Credit Card Fraud Detection Using Fuzzy Logic and Neural Network

The credit card fraud is dramatically increasing due to rise and rapid growth of E-commerce. Due to the huge number of transactions, credit card fraud detection is a big challenge for banks to minimize their losses and for customers to feel secure. In this paper fuzzy database is used to detect credit cards fraud. Fraud detection involves monitoring user's behaviour to estimate, detect or avoid undesirable behaviours. To correctly identify a transaction as legitimate or fraudulent has been considered a data mining problem. In this paper we discuss the fuzzy logic method, fuzzy rules, membership functions, fuzzification and defuzzification. Later, this method is implemented on the dataset using fuzzy logic toolbox in Matlab and the results are compared with the results of the artificial neural network method ANN. Our results indicates that the ANN method is 33% more accurate than the fuzzy logic.