Credit Card Fraud Detection with Machine Learning: An Integrated Use of Class Imbalance and Explainability Approaches

Abstract

Credit card fraud poses a critical threat to financial institutions. Existing studies rarely address class imbalance management, explainable AI, and counterfactual analysis within a unified framework; this study aims to fill that gap. A publicly available European credit card dataset comprising 284,807 transactions (0.172% fraud) was used, with a stratified 80/20 train-test split. Nine configurations combining three classifiers (Logistic Regression, Random Forest, XGBoost) with three balancing strategies (original, SMOTE, cost-sensitive) were evaluated using the F2 score, which prioritises false negative costs. XGBoost with the original strategy achieved F2=0.8524 and PR-AUC=0.8783; SMOTE and cost-sensitive strategies fell below the original strategy average (F2=0.7909). Multi-layer SHAP analysis consistently identified V14 (|SHAP|=1.74) and V4 (|SHAP|=1.72) as primary determinants across global, local, and interaction levels. Counterfactual analysis showed that shifting only three features (V14, V10, V4) reversed the fraud decision. Results suggest that a strong model architecture can simultaneously satisfy performance and explainability objectives.

Keywords

• credit card fraud
• class imbalance
• SMOTE
• XGBoost
• explainable artificial intelligence
• SHAP
• counterfactual analysis

Makine Öğrenmesi ile Kredi Kartı Sahtekarlığı Tespiti: Veri Dengesizliği ve Açıklanabilirlik Yaklaşımlarının Bütünleşik Kullanımı

Abstract

Kredi kartı sahtekarlığı finansal kurumlar için kritik bir tehdit oluşturmaktadır. Literatürde veri dengesizliği, açıklanabilir yapay zekâ ve karşı-olgusal analiz bütünleşik ele alınmamaktadır; bu çalışma bu boşluğu gidermeyi amaçlamaktadır. 284.807 işlem içeren Avrupa kredi kartı veri seti (%0,172 sahtekarlık) üzerinde tabakalı örneklemeyle %80/20 eğitim-test ayrımı uygulanmıştır. Üç sınıflandırıcı (Lojistik Regresyon, Rastgele Orman, XGBoost) üç dengeleme stratejisiyle (orijinal, SMOTE, maliyet duyarlı) birleştirilen dokuz konfigürasyon, hatalı negatiflere öncelik veren F2 skoru üzerinden değerlendirilmiştir. XGBoost orijinal stratejiyle F2=0,8524 ve PR-AUC=0,8783 elde etmiş; SMOTE ve maliyet duyarlı stratejiler ortalama F2 açısından (0,7909) gerisinde kalmıştır. Çok katmanlı SHAP analizi V14 (|SHAP|=1,74) ve V4 (|SHAP|=1,72)’ü global, lokal ve etkileşim düzeyinde tutarlı biçimde birincil belirleyiciler olarak ortaya koymuştur. Karşı-olgusal analiz, yalnızca üç özelliğin (V14, V10, V4) kaydırılmasının fraud kararını tersine çevirdiğini göstermiştir. Bulgular, güçlü model seçiminin performansı ve açıklanabilirliği eş zamanlı karşılayabileceğine işaret etmektedir.

Keywords

• kredi kartı sahtekarlığı
• veri dengesizliği
• SMOTE
• XGBoost
• açıklanabilir yapay zekâ
• SHAP
• karşı-olgusal analiz

References

1. Aburbeian, A. H. M., & Ashqar, H. I. (2023). Credit card fraud detection using enhanced random forest classifier for imbalanced data. arXiv preprint arXiv:2303.06514. https://doi.org/10.48550/arXiv.2303.06514
2. Arrieta, A. B., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., García, S., Gil-López, S., Molina, D., Benjamins, R., Chatila, R., & Herrera, F. (2020). Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI. Information Fusion, 58, 82–115. https://doi.org/10.1016/j.inffus.2019.12.012
3. Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32. https://doi.org/10.1023/A:1010933404324
4. Chang, V., Ali, B., Golightly, L., Ganatra, M. A., & Mohamed, M. (2024). Investigating credit card payment fraud with detection methods using advanced machine learning. Information, 15(8), 478. https://doi.org/10.3390/info15080478
5. Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002). SMOTE: Synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 16, 321–357. https://doi.org/10.1613/jair.953
6. Chen, T., & Guestrin, C. (2016). XGBoost: A scalable tree boosting system. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 785–794. https://doi.org/10.1145/2939672.2939785
7. Confalonieri, R., Coba, L., Wagner, B., & Besold, T. R. (2021). A historical perspective of explainable artificial intelligence. WIREs Data Mining and Knowledge Discovery, 11(1), e1391. https://doi.org/10.1002/widm.1391
8. Dal Pozzolo, A. (2015). Adaptive machine learning for credit card fraud detection [Doctoral dissertation, Université Libre de Bruxelles]. ULB Institutional Repository.

9. Dal Pozzolo, A., Caelen, O., Le Borgne, Y. A., Waterschoot, S., & Bontempi, G. (2014). Learned lessons in credit card fraud detection from a practitioner perspective. Expert Systems with Applications, 41(10), 4915–4928. https://doi.org/10.1016/j.eswa.2014.02.026
10. Fawcett, T. (2006). An introduction to ROC analysis. Pattern Recognition Letters, 27(8), 861–874. https://doi.org/10.1016/j.patrec.2005.10.010
11. Lundberg, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, 30, 4766–4777.
12. MLG-ULB. (2016). Credit card fraud detection [Dataset]. Kaggle. https://www.kaggle.com/datasets/mlg-ulb/creditcardfraud
13. Mothilal, R. K., Sharma, A., & Tan, C. (2020). Explaining machine learning classifiers through diverse counterfactual explanations. Proceedings of the 2020 ACM Conference on Fairness, Accountability, and Transparency, 607–617. https://doi.org/10.1145/3351095.3372850
14. Pes, B., & Lai, G. (2021). Cost-sensitive learning strategies for high-dimensional and imbalanced data: a comparative study. PeerJ Computer Science, 7, e832. https://doi.org/10.7717/peerj-cs.832
15. Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). "Why should I trust you?": Explaining the predictions of any classifier. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 1135–1144. https://doi.org/10.1145/2939672.2939778
16. Szeghalmy, S., & Fazekas, A. (2023). A comparative study of the use of stratified cross-validation and distribution-balanced stratified cross-validation in imbalanced learning. Sensors, 23(4), 2333. https://doi.org/10.3390/s23042333
17. Tayebi, M., & El Kafhali, S. (2025). Generative modeling for imbalanced credit card fraud transaction detection. Journal of Cybersecurity and Privacy, 5(1), 9. https://doi.org/10.3390/jcp5010009
18. The Nilson Report. (2023). Card fraud losses worldwide. HSN Consultants, Inc.