Application of Cost-Sensitive Threshold Calibration in Fraud Detection
DOI:
https://doi.org/10.54097/n0bz6e28Keywords:
Cost-Sensitive Learning; Threshold Calibration; Credit Card Fraud Detection; Class Imbalance; Bayesian DecisionAbstract
Detecting credit card fraud in practice must confront two core challenges: extreme rarity of the positive class and losses that differ sharply between the two types of mistake. The conventional practice of employing a fixed 0.5 cutoff while optimising for accuracy is ill-suited to this setting, as it cannot reconcile the economic damage of a missed fraud with the operational expense of a false alert. The present work develops a threshold calibration framework driven by expected cost minimisation. A misclassification cost matrix is first constructed. Three base classifiers—Logistic Regression, Random Forest, and XGBoost—provide per-transaction fraud likelihoods, which are then passed through Platt Scaling to improve reliability. A Bayes minimum-risk decision rule yields the optimal static cutoff, and an additional dynamic mechanism that ties the false-negative penalty to the transaction amount is introduced. Evaluated on the ULB Credit Card Fraud Detection set , probability calibration cuts the dynamic-threshold cost of Logistic Regression by 92.76%. The dynamic strategy attains the lowest overall expense of €2,655.28 on XGBoost, a 27.04% drop relative to the default 0.5 threshold. The proposed framework demonstrably lowers total misclassification cost and supplies an interpretable, cost-aware foundation for threshold decisions in high-stakes applications.
References
[1] Nilson Report. (2024). Card fraud losses reach $32.34 billion worldwide. https://nilsonreport.com/
[2] Abdallah, A., Maarof, M. A., & Zainal, A. (2016). Fraud detection system: A survey. Journal of Network and Computer Applications, 68, 90–113. https://doi.org/10.1016/j.jnca.2016.04.007
[3] Bolton, R. J., & Hand, D. J. (2002). Statistical fraud detection: A review. Statistical Science, 17(3), 235–255. https://doi.org/10.1214/ss/1042727940
[4] ULB Machine Learning Group. (2016). Credit card fraud detection dataset [Data set]. Kaggle. https://www.kaggle.com/datasets/mlg-ulb/creditcardfraud
[5] Korynetskyi, D., & Stetsenko, I. (2025). Detection method of fraudulent payment transaction based on C-score metric. Information, Computing and Intelligent Systems, 6, 75–86. https://doi.org/10.20535/2786-8729.6.2025.333898
[6] Bahnsen, A. C., Stojanovic, A., Aouada, D., & Ottersten, B. (2013). Cost sensitive credit card fraud detection using Bayes minimum risk. In Proceedings of the 12th International Conference on Machine Learning and Applications (ICMLA) (pp. 333–338). Miami, FL. https://doi.org/10.1109/ICMLA.2013.134
[7] Elbadraoui, A., Mouhssine, Y., El Alaoui, A., & Ouatik El Alaoui, S. (2026). Cost-aware deep neural network for credit card fraud detection under chronological evaluation. Statistical, Optimisation and Information Computing, 15(4), 3174–3193. https://doi.org/10.19139/soic-2310-5070-3362
[8] Duda, R. O., Hart, P. E., & Stork, D. G. (2001). Pattern classification (2nd ed.). Wiley. https://www.wiley.com/en-us/Pattern+Classification%2C+2nd+Edition-p-9780471056690
[9] Sheokand, S., & Beniwal, S. (2025). Credit card fraud detection using machine learning and deep learning techniques: A review. Pertanika Journal of Science & Technology, 33(6). http://pertanika2.upm.edu.my/pjst/browse/regular-issue?article=JST-5675-2024
[10] Sang, V. N. T. (2026). Enhancing credit card fraud detection under severe class imbalance using cost-sensitive learning and threshold optimization. EAI Endorsed Transactions on Security and Machine Learning. https://doi.org/10.4108/eetismla.12078
[11] Danang, D., & Putra, T. W. A. (2025). Cost-sensitive fraud detection with reliability calibration: A practical pipeline with XGBoost and focal-proxy reweighting. International Journal of Information Technology and Business, 8(1), 13–24. https://doi.org/10.24246/ijiteb.812025.13-24
[12] IEEE. (2026). A systematic review of ML/DL for credit card fraud detection. In 2026 IEEE Conference. https://ieeexplore.ieee.org/document/11430904
[13] IEEE. (2026). Real-time credit card fraud detection using graph attention networks with cost-sensitive learning. In 2026 IEEE Conference. https://ieeexplore.ieee.org/document/11546393
[14] Bhattacharyya, S., Jha, S., Tharakunnel, K., & Westland, J. C. (2011). Data mining for credit card fraud: A comparative study. Decision Support Systems, 50(3), 602–613.
[15] Hand, D. J., & Henley, W. E. (1997). Statistical classification methods in consumer credit scoring: A review. Journal of the Royal Statistical Society: Series A, 160(4), 523–541.
[16] Bahnsen, A. C., Aouada, D., Stojanovic, A., & Ottersten, B. (2016). Feature engineering strategies for credit card fraud detection. Expert Systems with Applications, 51, 134–142.
[17] Sheokand, S., & Beniwal, S. (2025). Credit card fraud detection using machine learning and deep learning techniques: A review. Pertanika Journal of Science & Technology, 33(6).
[18] 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.
[19] Korynetskyi, D., & Stetsenko, I. (2025). Detection method of fraudulent payment transaction based on C-score metric. Information, Computing and Intelligent Systems, 6.
[20] Elkan, C. (2001). The foundations of cost-sensitive learning. In Proceedings of the 17th International Joint Conference on Artificial Intelligence (IJCAI) (pp. 973–978). Seattle, USA.
[21] Bahnsen, A. C., Aouada, D., & Ottersten, B. (2015). A cost sensitive framework for credit card fraud detection. In Proceedings of the 4th International Conference on Pattern Recognition Applications and Methods (ICPRAM) (pp. 51–58). Lisbon, Portugal.
[22] IEEE. (2026). Credit card fraud detection using DQNN and integrated gradients for adaptive decisions. In 2026 IEEE Conference. https://ieeexplore.ieee.org/document/11566609
[23] Hafez, I. Y., Hafez, A. Y., Saleh, A., & Abd Elrahman, S. (2025). A systematic review of AI-enhanced techniques in credit card fraud detection. Journal of Big Data, 12(1), 1–35.
[24] Platt, J. (1999). Probabilistic outputs for support vector machines and comparisons to regularized likelihood methods. Advances in Large Margin Classifiers, 10(3), 61–74.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Jiayi Xing

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.







