AI Models Enhance Banking Fraud Detection Accuracy

The rise of online banking, exacerbated by global events like the COVID-19 pandemic, has led to a surge in fraudulent financial activities. Detecting these sophisticated scams is a critical challenge for financial institutions. This research investigates the effectiveness of artificial intelligence, specifically machine learning algorithms, in identifying fraudulent banking transactions. The study focuses on developing robust machine learning models and innovative data preprocessing techniques tailored for banking data. Key aspects include strategies for handling highly imbalanced datasets, which are common in fraud detection where legitimate transactions vastly outnumber fraudulent ones, as well as feature transformation and engineering to extract more predictive signals. The findings demonstrate that an artificial neural network model effectively improves fraud detection accuracy. Furthermore, a logistic regression algorithm achieved an impressive AUC value of approximately 0.946, with stacked generalization pushing this even higher to 0.954. These results underscore the significant potential of AI to enhance the security and integrity of digital banking operations.