Issue |
SHS Web Conf.
Volume 218, 2025
2025 2nd International Conference on Development of Digital Economy (ICDDE 2025)
|
|
---|---|---|
Article Number | 02017 | |
Number of page(s) | 9 | |
Section | Finance Tech Advances: Impacts and Innovations | |
DOI | https://doi.org/10.1051/shsconf/202521802017 | |
Published online | 03 July 2025 |
A Machine Learning Approach to Credit Card Transaction Fraud Prediction
College of Arts and Sciences, Boston University, Boston, MA, 02215, United States
* Corresponding author: Zixuan05@bu.edu
Credit card fraud has a significant impact on the financial industry and is now a growing concern. Machine learning can minimize bias against legitimate transactions and enable accurate identification of fraud. This study explores machine learning techniques to address category imbalances in credit card fraud detection datasets to mitigate economic losses while improving model performance. The results show that logistic regression outperforms other classifiers, including support vector classifiers (SVC), K-nearest neighbor classifiers (KNN), and decision trees, achieving an optimal balance between fraud detection and minimizing false positives. By conducting data processing techniques such as feature scaling and dataset balancing, the model shows an effective identification of fraudulent transactions that rarely exist in a vast number of legitimate transactions. In addition, simple neural networks trained on oversampled data reveal higher recall rates but at the cost of higher false positives, highlighting the tradeoff between accuracy and fraud detection sensitivity. These findings underscore the importance of choosing models that can both effectively detect fraud and minimize disruption to legitimate transactions, which also provide valuable insights for financial institutions seeking to enhance their fraud detection systems.
© The Authors, published by EDP Sciences, 2025
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.