Journal of Industrial Engineering and Management Studies

Abstract The ever-increasing volume and number of transactions in the bank make the fraud monitoring and detection process very complicated, costly, and time-consuming. In recent years, the development of new technologies has opened many ways for fraudsters and criminals to commit fraud. In this research, data mining methods are investigated in order to detect fraud in bank transactions. In order to detect fraud in bank card transactions, which are very unbalanced data types, the optimization of the support vector machine algorithm with hyperparameter techniques is presented and simulated on the Kaggle website data set, which includes bank card transactions, in the Python software environment has taken. The presented model benefits from bank transaction data and has the ability to extract complex patterns. As an effective optimization method, grid search technique intelligently adjusts the parameters of the support vector machine algorithm. The results of the model evaluation show that the support vector machine has a significant improvement in the detection of fraud patterns according to the criteria of accuracy and correctness. The combination of support vector machine and grid search technique as an innovative solution can help to improve the security of bank transactions in the digital age. In this research, hyperparameter optimization and smote balancing methods were used to reduce the number of false alarms. The proposed model can be commercialized and connected to the electronic banking system, online or offline, to detect fraudulent actions in transactions. The proposed model can be commercialized and connected to the electronic banking system, online or offline, to detect fraudulent actions in transaction.