Exploring the Role of Quantum Machine Learning for Credit Card Fraud Detection: A Hybrid Approach

Detecting fake credit card payments is a demanding problem. It occurs because valid transactions are very frequent then once false and signals of fraud are difficult to recognize. QML presents a strong approach to possibly enhance together the correctness and quickness of identification by attaching unique features of quantum machines. In this research we examine three QML simulations— the Variational Quantum Classifier (VQC), Sampler Quantum Neural Network (SQNN), Estimator Quantum Neural Network (EQNN)—with two specific data sets. We also experiments with multiple converters and model architecture. Our conclusions show that VQC steadily produces highly solid output, obtaining F!-score of 0.88. SQNN also exhibits solid execution, while the EQNN has complication in understanding the data. Moreover we implement QUBO-based classification model on a neural-atom quantum processor, exploring several system configurations in hybrid model approaches. Examine a real quantum device, using setups of upto 24 atoms, display a significant strength against system noise. Notably, in some cases, a low level of sound even linked to a little better outcomes. These novelties highlight the significance of well- planned QML model development and aim to significant potential of atom-based quantum system or improving fake transaction detection in digital security.