Enhancing Banking Security through Intelligent Models for Advanced Fraud Prevention
Sanaa Elyassami, Yasir Hamid, Abdulrahman Alhosani, Hamda Naser, Hamed Taher · 2023
Over the years, financial organizations have strived to enhance their fraud detection methods. However, as financial systems have become increasingly complex and the rules governing them have grown exponentially intricate, constructing and maintaining such systems has become a daunting challenge. With the rapid growth of technology, the internet, and online services, fraud against financial institutions has persisted. To address this, financial organizations have transitioned from rule-based monitoring systems to embrace modern machine learning algorithms. These machine learning-based systems can efficiently process vast volumes of data, connecting seemingly unrelated data points to detect suspicious patterns. This paper details the development of a deep learning model for classifying financial transactions, implementing five classifiers: Decision Tree, Logistic Regression, KNN, AdaBoost, and Random Forest algorithms. The incorporation of dropout regularization enhances classification performance. The results demonstrate promising accuracy and quick training times, with the Random Forest model exhibiting superior performance.