Aşırı Sınıf Dengesizliği Altında İşlem Düzeyinde Kara Para Aklama Tespiti için Makine Öğrenmesi Modellerinin Kıyaslanması

Öz

Bu çalışma, aşırı sınıf dengesizliği koşulları altında işlem düzeyinde kara para aklama tespitine yönelik denetimli makine öğrenmesi yöntemlerinin karşılaştırmalı bir değerlendirmesini sunmaktadır. Deneyler, yaklaşık yedi milyon işlemden oluşan ve kara para aklama oranı yaklaşık %0,05 olan IBM Transactions for Anti-Money Laundering veri setinin Lower Illicit–Small alt kümesi üzerinde gerçekleştirilmiştir. Lojistik Regresyon, Rastgele Orman, XGBoost ve CatBoost modelleri, modele özgü ön işleme adımları ve hiperparametre ayarlamaları kullanılarak uygulanmıştır. Model performansı, bağımsız bir test kümesi üzerinde doğruluk, kesinlik, duyarlılık, F1-skoru, ROC-AUC ve dengeli doğruluk ölçütleri kullanılarak değerlendirilmiştir. Bulgular, modeller arasında belirgin kesinlik–duyarlılık dengeleri bulunduğunu göstermektedir. Lojistik Regresyon en yüksek duyarlılık ve dengeli doğruluk değerleriyle kapsayıcı bir tespit yaklaşımını yansıtırken, CatBoost daha yüksek kesinlik ve ROC-AUC Alan performansı ile daha muhafazakâr alarm stratejilerini desteklemektedir. Sonuçlar, uygulamalı kara para aklama tarama sistemlerinde ölçüt temelli model seçiminin ve uygun karar eşiklerinin belirlenmesinin önemini ortaya koymaktadır.

Anahtar Kelimeler

• Kara Para Aklama
• Finansal Dolandırıcılık Tespiti
• Makine Öğrenmesi
• Sınıf Dengesizliği
• İşlem Sınıflandırması

Benchmarking Machine Learning Models for Transaction-Level Anti-Money Laundering under Extreme Class Imbalance

Öz

This study presents a comparative evaluation of supervised machine learning methods for transaction-level anti-money laundering detection under extreme class imbalance. Experiments are conducted on the Lower Illicit–Small subset of the IBM Transactions for anti-money laundering dataset, which includes nearly seven million transactions with a laundering prevalence of approximately 0.05%. Logistic Regression, Random Forest, XGBoost, and CatBoost models are implemented using model-specific preprocessing and hyperparameter tuning strategies. Model performance is assessed on an independent test set using accuracy, precision, recall, F1-score, ROC-AUC, and balanced accuracy metrics. The results reveal clear precision–recall trade-offs among the models. Logistic Regression achieves the highest recall and balanced accuracy, indicating a coverage-oriented detection strategy, whereas CatBoost demonstrates superior precision and ROC-AUC, supporting more conservative alerting approaches. Overall, the findings highlight the importance of metric-driven model selection and careful operating-point design in practical anti-money laundering screening systems.

Anahtar Kelimeler

• Anti-Money Laundering
• Financial Fraud Detection
• Machine Learning
• Class Imbalance
• Transaction Classification

Kaynakça

1. UNODC. "Money Laundering." United Nations Office on Drugs and Crime. https://www.unodc.org/unodc/en/money-laundering/overview.html (accessed.)
2. A. Venčkauskas, G. Š, L. Pocius, R. Brūzgienė, and A. Romanovs, "Machine Learning in Money Laundering Detection Over Blockchain Technology," IEEE Access, vol. 13, pp. 7555–7573, 2025.
3. C.R. Alexandre and J. Balsa, "Incorporating machine learning and a risk-based strategy in an anti-money laundering multiagent system," Expert Systems with Applications, vol. 217, p. 119500, 2023/05/01/ 2023.
4. S.S. Doddamani, K.G.K, and B. Bhowmik, "Money Laundering Detection in Imbalanced E-wallet Transactions with Threshold Optimization," in 2024 IEEE 9th International Conference for Convergence in Technology (I2CT), 5–7 April 2024 2024, pp. 1–6.
5. P. Sharma, A.S. Prakash, and A. Malhotra, "Application of Advanced AI Algorithms for Fintech Crime Detection," in 2024 15th International Conference on Computing Communication and Networking Technologies (ICCCNT), 24–28 June 2024 2024, pp. 1–6,.
6. G. Konstantinidis and A. Gegov, "Deep Neural Networks for Anti Money Laundering Using Explainable Artificial Intelligence," in 2024 IEEE 12th International Conference on Intelligent Systems (IS), 29–31 Aug. 2024 2024, pp. 1–6.
7. D.V. Kute, B. Pradhan, N. Shukla, and A. Alamri, "Explainable deep learning model for predicting money laundering transactions," International Journal on Smart Sensing and Intelligent Systems, vol. 17, no. 1, 2024, doi: 10.2478/ijssis-2024-0027.
8. S. Ramadhan, "Harnessing machine learning for money laundering detection: a criminological theory-centric approach," Journal of Money Laundering Control, vol. 28, no. 1, pp. 184–201, 2024, doi: 10.1108/jmlc-04-2024-0083.

9. E. Altman, J. Blanuša, L. Von Niederhäusern, B. Egressy, A. Anghel, and K. Atasu, "Realistic synthetic financial transactions for anti-money laundering models," Advances in Neural Information Processing Systems, vol. 36, pp. 29851–29874, 2023.
10. P. Salahi, R. Shady, I. A. Doush, and M. Kandil, "Performance Analysis of Machine Learning Techniques for Detecting Money Laundering in Bitcoin Transactions," in 2024 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT), 17–19 Nov. 2024 2024, pp. 156–163.
11. F. Wan and P. Li, "A Novel Money Laundering Prediction Model Based on a Dynamic Graph Convolutional Neural Network and Long Short-Term Memory," Symmetry, vol. 16, no. 3, p. 378, 2024. [Online]. Available: https://www.mdpi.com/2073-8994/16/3/378.
12. S.E and S.S.S, "Identifying illicit transactions in Bitcoin Tumbler services using Supervised Machine Learning Algorithms," in 2023 12th International Conference on Advanced Computing (ICoAC), 17–19 Aug. 2023 2023, pp. 1–8.
13. F. Irshad, T. Alkhalifah, F. Alturise, and Y. D. Khan, "GCF-MLD: Integrated Approach for Money Laundering Detection Using Machine Learning and Graph Network Analysis," IEEE Access, vol. 12, pp. 183961–183972, 2024.
14. T. H. Phyu and S. Uttama, "Improving Classification Performance of Money Laundering Transactions Using Typological Features," in 2023 7th International Conference on Information Technology (InCIT), 16–17 Nov. 2023 2023, pp. 520–525.
15. H. Gandhi, K. Tandon, S. Gite, B. Pradhan, and A. Alamri, "Navigating the Complexity of Money Laundering: Anti–money Laundering Advancements with AI/ML Insights," International Journal on Smart Sensing and Intelligent Systems, vol. 17, no. 1, 2024.
16. M.E. Lokanan, "Predicting money laundering sanctions using machine learning algorithms and artificial neural networks," Applied Economics Letters, vol. 31, no. 12, pp. 1112–1118, 2024/07/11 2024.
17. E.R. Altman. IBM Transactions for Anti Money Laundering (AML): LI-Small_Trans.csv, Kaggle. [Online]. Available: https://www.kaggle.com/datasets/ealtman2019/ibm-transactions-for-anti-money-laundering-aml
18. K.P. Fourkiotis and A. Tsadiras, "Future Internet Applications in Healthcare: Big Data-Driven Fraud Detection with Machine Learning," Future Internet, vol. 17, no. 10, p. 460, 2025.
19. A. Magklaras, C. Gogos, P. Alefragis, and A. Birbas, "Enhancing Parameters Tuning of Overlay Models with Ridge Regression: Addressing Multicollinearity in High-Dimensional Data," Mathematics, vol. 12, no. 20, p. 3179, 2024.
20. W. Sun, Q. Shen, Y. Gao, Q. Mao, T. Qi, and S. Xu, "Objective over Architecture: Fraud Detection Under Extreme Imbalance in Bank Account Opening," Computation, vol. 13, no. 12, p. 290, 2025.
21. H.-S. Jang, "Ensemble Learning Model for Industrial Policy Classification Using Automated Hyperparameter Optimization," Electronics, vol. 14, no. 20, p. 3974, 2025.
22. J. He, Z. Li, and L. Yin, "An Efficient and Accurate Random Forest Node-Splitting Algorithm Based on Dynamic Bayesian Methods," Machine Learning and Knowledge Extraction, vol. 7, no. 3, p. 70, 2025.
23. F. Imam, P. Musilek, and M. Z. Reformat, "Parametric and Nonparametric Machine Learning Techniques for Increasing Power System Reliability: A Review," Information, vol. 15, no. 1, p. 37, 2024.
24. M.B. Ayanoglu and I. Uysal, "PWFS: Probability-Weighted Feature Selection," Electronics, vol. 14, no. 11, p. 2264, 2025.
25. R. García-Carretero, R. Holgado-Cuadrado, and Ó. Barquero-Pérez, "Assessment of Classification Models and Relevant Features on Nonalcoholic Steatohepatitis Using Random Forest," Entropy, vol. 23, no. 6, p. 763, 2021.
26. M. Sasi, O.R. Adegboye, and A. Alzubi, "Explainable and Optimized Random Forest for Anomaly Detection in IoT Networks Using the RIME Metaheuristic," Electronics, vol. 14, no. 22, p. 4465, 2025.
27. A. Kundu, S.G. Kundu, S.K. Sahu, and N.D. Badgayan, "Leveraging Azure Automated Machine Learning and CatBoost Gradient Boosting Algorithm for Service Quality Prediction in Hospitality," Computers, vol. 14, no. 2, p. 32, 2025.
28. G. Marín Díaz, "A Unified Fuzzy–Explainable AI Framework (FAS-XAI) for Customer Service Value Prediction and Strategic Decision-Making," AI, vol. 7, no. 1, p. 3, 2026.
29. E. Pellegrino et al., "Extreme Gradient Boosting Tuned with Metaheuristic Algorithms for Predicting Myeloid NGS Onco-Somatic Variant Pathogenicity," Bioengineering, vol. 10, no. 7, p. 753, 2023.
30. A. Rashidi Nasab and H. Elzarka, "Optimizing Machine Learning Algorithms for Improving Prediction of Bridge Deck Deterioration: A Case Study of Ohio Bridges," Buildings, vol. 13, no. 6, p. 1517, 2023.
31. A. Andrews, N. Islam, G. Bcharah, H. Bcharah, and M. Pangasa, "Development and Internal Validation of Machine Learning Algorithms to Predict 30-Day Readmission in Patients Undergoing a C-Section: A Nation-Wide Analysis," Journal of Personalized Medicine, vol. 15, no. 10, p. 476, 2025.
32. F. Milella, L. Famiglini, G. Banfi, and F. Cabitza, "Application of Machine Learning to Improve Appropriateness of Treatment in an Orthopaedic Setting of Personalized Medicine," Journal of Personalized Medicine, vol. 12, no. 10, p. 1706, 2022.
33. Z. Chen and W. Fan, "A Freeway Travel Time Prediction Method Based on an XGBoost Model," Sustainability, vol. 13, no. 15, p. 8577, 2021.
34. S.S. Abinayaa et al., "Securing the Edge: CatBoost Classifier Optimized by the Lyrebird Algorithm to Detect Denial of Service Attacks in Internet of Things-Based Wireless Sensor Networks," Future Internet, vol. 16, no. 10, p. 381, 2024.
35. P. König et al., "Boost-Classifier-Driven Fault Prediction Across Heterogeneous Open-Source Repositories," Big Data and Cognitive Computing, vol. 9, no. 7, p. 174, 2025.
36. Q. Wang, C. Yan, Y. Zhang, Y. Xu, X. Wang, and P. Cui, "Numerical Simulation and Bayesian Optimization CatBoost Prediction Method for Characteristic Parameters of Veneer Roller Pressing and Defibering," Forests, vol. 15, no. 12, p. 2173, 2024.
37. M. Kopyt, P. Piotrowski, and D. Baczyński, "Short-Term Energy Generation Forecasts at a Wind Farm—A Multi-Variant Comparison of the Effectiveness and Performance of Various Gradient-Boosted Decision Tree Models," Energies, vol. 17, no. 23, p. 6194, 2024.
38. Y. Liu, T. Yang, L. Tian, B. Huang, J. Yang, and Z. Zeng, "Ada-XG-CatBoost: A Combined Forecasting Model for Gross Ecosystem Product (GEP) Prediction," Sustainability, vol. 16, no. 16, p. 7203, 2024.
39. Y. He, B. Yang, and C. Chu, "GA-CatBoost-Weight Algorithm for Predicting Casualties in Terrorist Attacks: Addressing Data Imbalance and Enhancing Performance," Mathematics, vol. 12, no. 6, p. 818, 2024.
40. A. Ali et al., "Towards improved fake news detection using a hybrid RoBERTa and metadata enhanced XGBoost model," Scientific Reports, 2025/12/04 2025.
41. R. Zhou, J. Yang, H. Liu, C. Xu, Y. Pu, and I. S. Babichuk, "A lightweight model and agent framework for fast and accurate surface defect detection in MoS2 films," Materials Science in Semiconductor Processing, vol. 193, p. 109494, 2025/07/01/ 2025.
42. V. Sinap, "Bankruptcy Prediction with Optuna-Enhanced Ensemble Machine Learning Methods: A Comparison of Oversampling and Undersampling Techniques," (in en), Dicle Üniversitesi Mühendislik Fakültesi Mühendislik Dergisi, vol. 16, no. 1, pp. 97–113, March 2025.
43. M.F.I. Amin, Y. Watanobe, M.M. Rahman, and A. Shirafuji, "Source Code Error Understanding Using BERT for Multi-Label Classification," IEEE Access, vol. 13, pp. 3802–3822, 2025.
44. M. Meral, F. Ozbilgin, and F. Durmus, "Fine-Tuned Machine Learning Classifiers for Diagnosing Parkinson’s Disease Using Vocal Characteristics: A Comparative Analysis," Diagnostics, vol. 15, no. 5, p. 645, 2025.
45. M. Ait Omar, I. Etebaai, M. Taher, and A. Tawfik, "Landslide susceptibility mapping in the Bokoya Massif, Northern Morocco: A geospatial and multi-factor analysis using the analytic hierarchy process (AHP)," Scientific African, vol. 30, p. e02980, 2025/12/01/ 2025.
46. M.H. Moharam, K. Ashraf, H. Alaa, M. Ahmed, and H. A. El-Hakim, "Real-time detection of Wi-Fi attacks using hybrid deep learning models on NodeMCU," Scientific Reports, vol. 15, no. 1, p. 32544, 2025/09/15 2025.
47. T.D. Waele et al., "Time-Series-Based Feature Selection and Clustering for Equine Activity Recognition Using Accelerometers," IEEE Sensors Journal, vol. 23, no. 11, pp. 11855–11868, 2023.