FİNANSAL KRİZLERDE BANKA İFLAS RİSKİNİ ÖNGÖRME: MAKİNE ÖĞRENİMİ İLE RİSK FAKTÖRLERİNİN ANALİZİ

Öz

Bu çalışma, finansal krizler sırasında bankaların iflas riskini, finansal yapılarını makine öğrenmesi yöntemleriyle analiz ederek tahmin etmeyi amaçlamaktadır. Örnek olay olarak 2001 Türkiye ekonomik krizi ele alınarak, batan bankaların finansal oranları faaliyetlerine devam eden bankalarla karşılaştırılmıştır. Kârlılık, likidite ve sermaye yeterliliği gibi finansal göstergeler, geleneksel oran analizi ve C5.0 Karar Ağacı, CART ve XGBoost gibi gelişmiş makine öğrenimi modelleri aracılığıyla değerlendirilmektedir. Bu modellerin performansı, ROC eğrisi ve AUC değeri gibi değerlendirme ölçütleri kullanılarak değerlendirilmiştir. Bulgular, Net İşletme Sermayesi/Toplam Aktifler, Takipteki Krediler Sonrası Net Faiz Geliri/Ortalama Toplam Aktifler ve Faiz Geliri/Faiz Gideri gibi finansal oranların iflas tahmininde kritik bir rol oynadığını göstermektedir. Modeller arasında CART ve XGBoost mükemmel sınıflandırma doğruluğu (AUC = 1) ile performans gösterirken, C5.0 modeli de yüksek düzeyde başarı elde etmiştir (AUC = 0.9318). Sonuçlar, bankalar için finansal sürdürülebilirliği yönetme ve erken uyarı sistemlerini geliştirme konusunda pratik bilgiler sağlamaktadır. Özellikle ekonomik istikrarsızlık dönemlerinde güçlü bir sermaye tabanının korunmasının ve faiz gelirleri ile likidite göstergelerinin izlenmesinin önemini vurgulamaktadır. Bu araştırma, yorumlanabilir makine öğrenimi yöntemlerini finansal oran analizi ile bütünleştirerek finansal risk yönetimi literatürüne katkıda bulunmaktadır.

Anahtar Kelimeler

• İflas Tahmini
• Finansal Krizler
• Bankacılık Sektörü
• Makine Öğrenimi
• Finansal Sürdürülebilirlik.

PREDICTING BANK FAILURE RISK IN FINANCIAL CRISES: ANALYSIS OF RISK FACTORS WITH MACHINE LEARNING

Abstract

This study aims to predict the bankruptcy risk of banks during financial crises by analyzing their financial structure using machine learning methods. Taking the 2001 Turkish economic crisis as a case study, the financial ratios of failed banks are compared with those that remained operational. Financial indicators such as profitability, liquidity, and capital adequacy are evaluated through traditional ratio analysis and advanced machine learning models, including C5.0 Decision Tree, CART, and XGBoost. The performance of these models is assessed using evaluation metrics such as the ROC curve and AUC value. The findings show that financial ratios like Net Working Capital/Total Assets, Net Interest Income after Non-performing Loans/Average Total Assets, and Interest Income/Interest Expense play a critical role in bankruptcy prediction. Among the models, CART and XGBoost performed with perfect classification accuracy (AUC = 1), while the C5.0 model also achieved a high level of success (AUC = 0.9318). The results provide practical insights for banks in managing financial sustainability and improving early warning systems. They emphasize the importance of maintaining a strong capital base and monitoring interest income and liquidity indicators, especially during periods of economic instability. This research contributes to financial risk management literature by integrating interpretable machine learning methods with financial ratio analysis.

Keywords

• Failure Prediction
• Financial Crises
• Banking Sector
• Machine Learning
• Financial Sustainability.

Kaynakça

1. Baesens, B., Van Gestel, T., Stepanova, M., Van den Poel, D., & Vanthienen, J. (2003). Benchmarking state-of-the-art classification algorithms for credit scoring. Journal of the Operational Research Society, 54(6), 627–635.
2. 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.
3. Ďurica, M., Frnda, J., & Švábová, L. (2019a). Decision tree based model of business failure prediction for polish companies. Oeconomia Copernicana, 10(3), 453-469. https://doi.org/10.24136/oc.2019.022
4. Ďurica, M., Podhorská, I., & Ďurana, P. (2019b). Business failure prediction using cart-based model: a case of slovak companies. Ekonomicko-Manazerske Spektrum, 13(1), 51-61. https://doi.org/10.26552/ems.2019.1.51-61
5. Ďurica, M., Frnda, J., & Švábová, L. (2021). Financial distress prediction in slovakia: an application of the cart algorithm. Journal of International Studies, 14(1), 201-215. https://doi.org/10.14254/2071-8330.2021/14-1/14
6. Ekinci, R. & Kök, R. (2020). The competition and stability relationship in the european union banking industry: an empirical analysis on the commercial banks. İzmir İktisat Dergisi, 35(4), 879-894. https://doi.org/10.24988/ije.202035414
7. Fawcett, T. (2006). An introduction to ROC analysis. Pattern Recognition Letters, 27(8), 861–874.
8. Garcia, M. T. M. & Guerreiro, J. (2016). Internal and external determinants of banks’ profitability. Journal of Economic Studies, 43(1), 90-107. https://doi.org/10.1108/jes-09-2014-0166

9. Gocheva-Ilieva, S., Kulina, H., & Ivanov, A. (2020). Assessment of students’ achievements and competencies in mathematics using cart and cart ensembles and bagging with combined model improvement by mars. Mathematics, 9(1), 62. https://doi.org/10.3390/math9010062
10. Hakimi, A., Boussaada, R., & Hamdi, H. (2020). The interactional relationships between credit risk, liquidity risk and bank profitability in mena region. Global Business Review, 23(3), 561-583. https://doi.org/10.1177/0972150919879304
11. Han, J., Kamber, M., & Pei, J. (2011). Data Mining: Concepts and Techniques. Morgan Kaufmann.
12. Hanley, J. A., & McNeil, B. J. (1982). The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology, 143(1), 29–36.
13. Hastie, T., Tibshirani, R., & Friedman, J. (2009). The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer Science & Business Media.
14. Ho, C., McCarthy, P., Yang, Y., & Ye, X. (2012). Bankruptcy in the pulp and paper industry: market’s reaction and prediction. Empirical Economics, 45(3), 1205-1232. https://doi.org/10.1007/s00181-012-0661-6
15. Hussain, M., Zhu, W., Zhang, W., & Abidi, S. M. R. (2018). Student engagement predictions in an e-learning system and their impact on student course assessment scores. Computational Intelligence and Neuroscience, 2018, 1-21. https://doi.org/10.1155/2018/6347186
16. Karki, D. & Aryal, A. (2019). Risk and resilience: examining the role of capital adequacy and credit risk in shaping the performance of nepalese commercial banks. Journal of Development and Administrative Studies, 27(1-2), 31-40. https://doi.org/10.3126/jodas.v27i1-2.60573
17. Laryea, E., Ntow‐Gyamfi, M., & Alu, A. A. (2016). Nonperforming loans and bank profitability: evidence from an emerging market. African Journal of Economic and Management Studies, 7(4), 462-481. https://doi.org/10.1108/ajems-07-2015-0088
18. Nam, Y. (2023). A study on the factors and prediction model of triple-negative breast cancer for public health promotion. Diagnostics, 13(22), 3486. https://doi.org/10.3390/diagnostics13223486
19. Olalere, O. E., Omar, W. A. W., & Kamil, S. (2017). Bank specific and macroeconomic determinants of commercial bank profitability: empirical evidence from nigeria. International Journal of Finance &Amp; Banking Studies (2147-4486), 6(1), 25. https://doi.org/10.20525/ijfbs.v6i1.627
20. Quinlan, J. R. (1996). Improved use of continuous attributes in C4.5. Journal of Artificial Intelligence Research, 4, 77–90.
21. Ramdani, N., Prasetyowati, S. S., & Sibaroni, Y. (2022). Performance analysis of bandung city traffic flow classification with machine learning and kriging interpolation. Building of Informatics, Technology and Science (BITS), 4(2), 694-704. https://doi.org/10.47065/bits.v4i2.1972
22. Rau, C., Wu, S., Chien, P., Kuo, P., Chen, Y., Hsieh, H., … & Hsieh, C. (2017). Prediction of mortality in patients with isolated traumatic subarachnoid hemorrhage using a decision tree classifier: a retrospective analysis based on a trauma registry system. International Journal of Environmental Research and Public Health, 14(11), 1420. https://doi.org/10.3390/ijerph14111420
23. Rau, C., Wu, S., Chien, P., Kuo, P., Chen, Y., Hsieh, H., … & Liu, H. (2018). Identification of pancreatic injury in patients with elevated amylase or lipase level using a decision tree classifier: a cross-sectional retrospective analysis in a level i trauma center. International Journal of Environmental Research and Public Health, 15(2), 277. https://doi.org/10.3390/ijerph15020277
24. Rokach, L., & Maimon, O. (2014). Data Mining with Decision Trees: Theory and Applications. World Scientific.
25. Sayari, M., Rahmanian Haghighi, M. R., Bagheri Lankarani, K., Ghahramani, S., & Honarvar, B. (2024). The global road traffic death rate and human development index from 2000 to 2019: a trend analysis. Archives of Iranian Medicine, 27(3), 113-121. https://doi.org/10.34172/aim.2024.18
26. Sharma, N. & Iqbal, S. J. (2023). Applying decision tree algorithm classification and regression tree (cart) algorithm to gini techniques binary splits. International Journal of Engineering and Advanced Technology, 12(5), 77-81. https://doi.org/10.35940/ijeat.e4195.0612523
27. Siddique, A., Khan, M. A., & Khan, Z. A. (2021). The effect of credit risk management and bank-specific factors on the financial performance of the south asian commercial banks. Asian Journal of Accounting Research, 7(2), 182-194. https://doi.org/10.1108/ajar-08-2020-0071
28. Tekin, B. & Gör, Y. (2022). Financial failure forecast models and an application on banking sector financial statements: altman and springate models. Adıyaman Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 40, 373-404. https://doi.org/10.14520/adyusbd.992296
29. Telmoudi, F., Ghourabi, M. E., & Limam, M. (2011). Rst-gcbr-clustering-based rga-svm model for corporate failure prediction. Intelligent Systems in Accounting, Finance and Management, 18(2-3), 105-120. https://doi.org/10.1002/isaf.323
30. The Banks Association of Turkiye, Retrieved from https://www.tbb.org.tr/istatistiki-raporlar/2000-yillik-secilmis-rasyolar. Accessed 05.08.2024.
31. Umut, E. (2024). Kazanç yönetiminde kullanılan karma modeller ve altman z modeli arasındaki i̇lişkiler: amerikan şirketleri üzerine bir analiz. Turk Turizm Arastirmalari Dergisi, 5(1), 67-95. https://doi.org/10.26677/tr1010.2024.1390
32. Zhu, Y., Tao, T., Yu, K., Qu, X., Li, S., Wickert, J., … & Semmling, M. (2020). Machine learning-aided sea ice monitoring using feature sequences extracted from spaceborne gnss-reflectometry data. Remote Sensing, 12(22), 3751. https://doi.org/10.3390/rs12223751