Abstract
Credit card fraud datasets are extremely imbalanced. This study systematically evaluates SMOTE, Borderline-SMOTE, ADASYN and SMOTE-Tomek together with random forest, XGBoost and LightGBM classifiers. Experiments were conducted with stratified cross-validation on a public dataset of 284,807 transactions. XGBoost combined with SMOTE-Tomek gave the best result with a PR-AUC of 0.88. Applying oversampling only within training folds was shown to be critical to prevent data leakage.
Dengesiz Veri Kümelerinde SMOTE Türevleriyle Kredi Kartı Dolandırıcılığı Tespiti
Kredi kartı dolandırıcılığı veri kümeleri aşırı derecede dengesizdir. Bu çalışmada SMOTE, Borderline-SMOTE, ADASYN ve SMOTE-Tomek yöntemleri; rastgele orman, XGBoost ve LightGBM sınıflandırıcılarıyla birlikte sistematik olarak değerlendirilmiştir. Deneyler 284.807 işlem içeren açık veri kümesinde tabakalı çapraz doğrulama ile yürütülmüştür. SMOTE-Tomek ile birleştirilen XGBoost modeli 0,88 PR-AUC ile en iyi sonucu vermiştir. Aşırı örneklemenin yalnızca eğitim katmanlarında uygulanmasının veri sızıntısını önlemede kritik olduğu gösterilmiştir.
Full Text
Declarations
- Ethics Approval
- This study does not require ethics committee approval.
- Conflict of Interest
- The authors declare no conflict of interest.
References 6
- 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
- Chen, T., & Guestrin, C. (2016). XGBoost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 785–794). https://doi.org/10.1145/2939672.2939785
- Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5–32. https://doi.org/10.1023/A:1010933404324
- Lundberg, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, 30.
- Hastie, T., Tibshirani, R., & Friedman, J. (2009). The elements of statistical learning (2nd ed.). Springer. https://doi.org/10.1007/978-0-387-84858-7
- Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., et al. (2011). Scikit-learn: Machine learning in Python. Journal of Machine Learning Research, 12, 2825–2830.
How to Cite
License
© 2024 Merve Şahin, Aigerim Nurlanovna. This article is distributed under the terms of the CC BY 4.0 license, which permits unrestricted use, distribution and reproduction in any medium, provided the original work is properly cited. License text