Daily Archives: October 9, 2024

Abstract-In recent times, the exponential growth in the usage of credit cards has increased fraudulent activities, which impacts financial institutions significantly. A large number of machine learning (ML) techniques are used to detect fraudulent transactions in order to thwart such threats. This paper represents a review of state-of-the-art ML algorithms used for credit card fraud detection and further analyzes their performance with regard to accuracy and privacy. Besides, a hybrid approach combining ANN with federated learning is proposed. This approach has the potential to not only increase the detection accuracy but also mitigate data privacy issues. The given model has had promising results for real-time application in credit card fraud detection while keeping users’ data private. Keywords— Artificial Neural Networks, Credit Card Fraud Detection, Federated Learning, Machine Learning, Privacy-Preserving, Blockchain. Credit card fraud has been an exploding problem with the large-scale growth of digital transactions, posing significant risk exposure to financial institutions. In this paper, we conducted a comprehensive review of various ML techniques applied to credit card fraud detection, touching on both aspects of accuracy and concerns over data privacy. We herein present a novel hybrid model based on the paradigm combination of ANN and FL for overcoming challenges arising from accuracy and privacy protection in detection. The advantages of the model are the usage of pattern recognition ability on ANN and its preservation of data privacy through decentralized learning. It has promising uses and outcomes since high detection accuracy and user privacy persistence were noted in achieving this characteristic. This makes this type of model suit fraud detection applications applied real-time. Keywords: Credit card fraud detection Machine learning Artificial neural networks Federated learning Privacy.