Hybrid Quantum-Classical Method for Bank Account Fraud Detection using Quantum Encoding and Quantum Machine Learning

Fraud detection has attracted interest because it can ensure corporate sustainability, minimize losses, and catch fraudsters. We introduce a novel classification/terminology of Devastating Evolutionary Dynamic Entities (DEDE), where fraud and its sub-types fall under. The paper emphasizes the ineffectiveness of traditional fraud detection methods and calls for the adoption of state-of-the-art fraud detection models immediately. Due to hardware constraints, a hybrid quantum-classical setting was used. Although hybrid methods do not outperform classical methods under current hardware limitations, they still perform moderately well under limited samples, hardware, and features. Overall, we believe that quantum computing has great potential for fraud detection.