Ajibade-Ajibosin, Boluwatife, Ani, Uchenna Daniel, Kyriacou, Theocharis ORCID: https://orcid.org/0000-0002-5211-3686 and Turner, Mark (2024) Improving Credit Card Fraud Detection with Combined Feature Extraction and Class Balance Optimisation Techniques. In: Jaatun, M. G., (ed.) Proceedings of the International Conference on Cybersecurity, Situational Awareness and Social Media. Cyber Science 2024. : Springer Proceedings in Complexity . Springer, pp. 27-44

Full text not available from this repository.

Abstract

Credit card fraud is a persistent and evolving challenge that poses significant financial harm to cardholders, financial institutions, national, and global economies. The use of Machine Learning (ML) methods has heavily enhanced the detection of credit card fraud, offering improvements to other traditional credit card fraud detection approaches such as human manual checks and rule-based methods. However, there are limitations that impact the performance and efficiency of the credit card detection using ML. This work addresses some of the current challenges associated with employing ML for credit card fraud by developing a robust credit card fraud detection model. The proposed model employs Synthetic Minority Over-sampling Technique (SMOTE) to address the class imbalance issue typically akin to credit card fraud datasets. A Recursive Feature Elimination with Cross-Validation (RFECV) scheme was utilized as a feature selection technique to select the optimal subset of features that can improve the accuracy of the credit card fraud detection model. Following rigorous evaluation, the proposed credit card fraud detection system demonstrates exceptional performance above other known existing systems across critical metrics including accuracy, recall, precision, F1 score, and ROC-AUC. Thus, is considered a better solution for more accurate detection of credit card fraud instances.

Item Type:	Book Section
Status:	Published
DOI:	10.1007/978-981-96-0401-2_2
School/Department:	York Business School
URI:	https://ray.yorksj.ac.uk/id/eprint/13093

University Staff: Request a correction | RaY Editors: Update this record