Automatic Bank Fraud Detection Using Support Vector Machines

With the significant development of communications and computing, bank fraud is growing in its forms and amounts. We try in this paper to analyze the various forms of fraud to which are exposed banks and data mining tools allowing its early detection using data already accumulated in a bank. We propose the use of supervised learning methods called support vector machines to build models representing normal and abnormal customers behaviors and then use it to check new transactions. We also propose a hybridization of the two SVM methods, binary and single class to enhance the detection of fraudulent transactions. The obtained results from databases of credit card transactions show the power of these techniques in the fight against banking fraud comparing them to others in the same field.