Design and Development of an Efficient System for the Detection of Fraudulent in the Banking Industry

Many people and businesses have lost a lot of money due to fraudulent online transactions. The development of cutting-edge technology and the expansion of global communication have both contributed to a rise in online fraud. In this study, an open data set on credit card transactions from Kaggle was used in this research. The data underwent extensive preprocessing, which included standardizing, and balancing classes, to provide the best possible data quality for training the model. Five models were tested: CNN, KNN, Random Forest, AdaBoost Classifier, and Logistic Regression. The Random Forest model demonstrated the highest performance compared to all other models, with an Area Under the Curve of 0.98, as well as the best accuracy, precision, recall, and F1 score, at 0.98% and 0.99%, respectively. Because credit card fraud methods are always changing, it is essential that we have better fraud detection systems, and the results offered here provide a solid foundation for that effort.