Digital Twin-Enabled IoT Automation System for Monitoring and Control of Fraudulent Transactions in the Banking Industry
The banking industry faces significant challenges due to fraudulent transactions, resulting in substantial financial losses and decreased customer trust. Traditional fraud detection methods have limitations, including high false-positive rates and slow detection The aim of this study is to introduce a digital twin-enable.d IoT automation system to improve fraud detection and control in the banking industry This work proposes a dynamic solution system that involves creating a virtual replica of the banking transaction process and integrating real-time data from transaction logs and user behavior. Machine learning algorithms, implemented using MATLAB code, are used to analyze data and detect anomalies, with automated responses for blocking or flagging suspicious transactions. Simulations using real-world transaction data and MATLAB code demonstrated that the system flagged one anomalous transaction and blocked one fraudulent transaction. This showcases a 30% reduction in false positives, a 40% improvement in detection speed, and a 25% increase in prediction accuracy compared to conventional methods. The model demonstrated 92% accuracy, 50% precision, and 6.25% recall, indicating enhanced detection accuracy and faster response times, significantly improving fraud prevention in banking.
Download Paper