Building Early-Warning Models for Multiclass Credit Risk Staging: A Comparative Study of Random Forest and XGBoost with SMOTENC and Borderline-SMOTE
DOI:
https://doi.org/10.52436/1.jutif.2026.7.4.5611Keywords:
Borderline-SMOTE, Credit Staging, Early Warning System, Imbalanced Classification, Random Forest, SMOTENC, XGBoostAbstract
This paper develops and evaluates an early-warning credit risk classifier using a three-class formulation that distinguishes performing loans (L), delinquent accounts (DP), and non-performing loans (NPL). Early-warning modeling is challenging because deterioration events are relatively infrequent, yielding class imbalance and asymmetric error costs where missed NPL cases are particularly consequential. The study aims to build and compare operationally feasible models that improve NPL identification while maintaining reliable performance across all classes. Experiments use 2,253 loan records with 20 early-warning predictors and compare Random Forest and XGBoost under four imbalance-handling strategies: a no-resampling baseline, class/sample weighting, SMOTENC, and Borderline-SMOTE. Performance is estimated via stratified 5-fold cross-validation, with out-of-fold predictions aggregated for consistent fold-robust evaluation and probability-based diagnostics. Evaluation emphasizes imbalance-aware criteria, including macro F1, balanced accuracy, NPL precision/recall/F1, and ranking metrics for NPL versus non-NPL (PR-AUC and ROC-AUC), complemented by agreement statistics (Cohen’s kappa and Matthews correlation coefficient). Results show that class-sensitive metrics reveal clearer differences among strategies than global measures, with XGBoost combined with SMOTENC achieving the strongest NPL-oriented performance (recall 0.7389; F1 0.8035; PR-AUC 0.8780; ROC-AUC 0.9434). The contribution is a controlled out-of-fold benchmark of imbalance strategies for mixed-type tabular credit staging, providing a reproducible evaluation protocol for financial informatics and imbalanced multiclass learning.
Downloads
References
A. Goswami, “Does credit risk persist in the Indian banking industry? Recent evidence,” Asian Journal of Economics and Banking, vol. 6, no. 2, pp. 178–197, Aug. 2022, doi: 10.1108/ajeb-01-2021-0006.
J. Akomeah, R. Agumeh, and S. Frimpong, “Credit risk management and financial performance of listed banks in Ghana,” Research Journal of Finance and Accounting, Mar. 2020, doi: 10.7176/rjfa/11-6-05.
T. K. Bhatt, N. Ahmed, M. B. Iqbal, and M. Ullah, “Examining the determinants of credit risk management and their relationship with the performance of commercial banks in Nepal,” J. Risk Financ. Manag., vol. 16, no. 4, p. 235, Apr. 2023, doi: 10.3390/jrfm16040235.
P. Jakubík and B.-G. Moinescu, “A reversed early warning methodology for optimal bank profit retention recommendations,” Int. J. Finance Econ., vol. 30, no. 4, pp. 3457–3475, Oct. 2025, doi: 10.1002/ijfe.3073.
H. Khalfaoui and H. Guenichi, “Economic policy uncertainty effect on credit risk, lending decisions and banking performance: Evidence from Tunisian listed banks,” Journal of Economic and Administrative Sciences, vol. 38, no. 2, pp. 287–303, Apr. 2022, doi: 10.1108/jeas-09-2020-0159.
D. G. Breed et al., “A Forward-Looking IFRS 9 Methodology, Focussing on the Incorporation of Macroeconomic and Macroprudential Information into Expected Credit Loss Calculation,” Risks, vol. 11, no. 3, pp. 59–59, Mar. 2023, doi: 10.3390/risks11030059..
A. Ashofteh and J. M. Bravo, “A conservative approach for online credit scoring,” Expert Systems with Applications, vol. 176, p. 114835, Aug. 2021, doi: 10.1016/j.eswa.2021.114835..
A. A. Adenuga, O. Gaffar, A. O. Sikiru, and M. Otunba, “An IFRS 9-Compliant Impairment Automation Framework for Financial Accuracy and Regulatory Compliance,” International Journal of Multidisciplinary Research and Growth Evaluation, vol. 2, no. 2, pp. 513–523, Jan. 2021, doi: 10.54660/.ijmrge.2021.2.2.513-523.
C. CRISTIAN, C. NORA, and N. IONUT, “A Practical Approach to Development and Validation of Credit Risk Models Based on Data Analysis,” ECONOMIC COMPUTATION AND ECONOMIC CYBERNETICS STUDIES AND RESEARCH, vol. 56, no. 3/2022, pp. 51–67, Sep. 2022, doi: 10.24818/18423264/56.3.22.04.
P. Miu and B. Ozdemir, “A coherent economic framework to model correlations between PD, LGD and EaD, and its applications in EaD modelling and IFRS-9,” Journal of risk management in financial institutions, vol. 16, no. 1, pp. 52–52, Jan. 2023, doi: 10.69554/qcvi3102.
W. Lee et al., “Machine learning enhances the performance of short and long-term mortality prediction model in non-ST-segment elevation myocardial infarction,” Sci. Rep., vol. 11, no. 1, p. 12886, Jun. 2021, doi: 10.1038/s41598-021-92362-1.
P. O. de Capretz et al., “Machine learning for early prediction of acute myocardial infarction or death in acute chest pain patients using electrocardiogram and blood tests at presentation,” BMC Med. Inform. Decis. Mak., vol. 23, no. 1, Art. no. 25, Feb. 2023, doi: 10.1186/s12911-023-02119-1.
T. Xu, “Fraud detection in credit risk assessment using supervised learning algorithms,” Computer Life, vol. 12, no. 2, pp. 30–36, Aug. 2024, doi: 10.54097/qw9j1892.
G. Teles, J. J. P. Rodrigues, R. A. L. Rabêlo, and S. A. Kozlov, “Comparative study of support vector machines and random forests machine learning algorithms on credit operation,” Softw. Pract. Exp., vol. 51, no. 12, pp. 2492–2500, Dec. 2021, doi: 10.1002/spe.2842.
M. Marimo and C. Chimedza, “SME credit risk modelling in South Africa: A case study,” Acta Econ., vol. 20, no. 36, pp. 9–30, Jun. 2022, doi: 10.7251/ace2236009m.
Q. Zhou, L. Wang, J. Li, S. Zhou, and L. Li, “The study on credit risk warning of regional listed companies in China based on logistic model,” Discrete Dyn. Nat. Soc., vol. 2021, pp. 1–8, May 2021, doi: 10.1155/2021/6672146.
P. K. Roy and K. Shaw, “A multicriteria credit scoring model for SMEs using hybrid BWM and TOPSIS,” Financ. Innov., vol. 7, no. 1, Dec. 2021, doi: 10.1186/s40854-021-00295-5.
T. Hidayat, D. Masyita, S. R. Nidar, F. Ahmad, and M. A. N. Syarif, “Early warning early action for the banking solvency risk in the COVID-19 pandemic era: A case study of Indonesia,” Economies, vol. 10, no. 1, p. 6, Dec. 2021, doi: 10.3390/economies10010006.
A. Malik, H. Aziz, B. Saiti, and S. U. Din, “The impact of earnings variability and regulatory measures on income smoothing: Evidence from panel regression,” J. Cent. Bank. Theory Pract., vol. 10, no. 1, pp. 183–201, Jan. 2021, doi: 10.2478/jcbtp-2021-0009.
H. P. Kusumah and T. Faturohman, “Impact of restructuring strategies and financial instruments for loan loss provision,” European Journal of Business and Management Research, vol. 8, no. 5, pp. 198–204, Oct. 2023, doi: 10.24018/ejbmr.2023.8.5.2100.
S. Shi, R. Tse, W. Luo, S. D’Addona, and G. Pau, “Machine learning-driven credit risk: A systematic review,” Neural Comput. Appl., vol. 34, no. 17, pp. 14327–14339, Sep. 2022, doi: 10.1007/s00521-022-07472-2.
T. Le, “A comprehensive survey of imbalanced learning methods for bankruptcy prediction,” IET Commun., vol. 16, no. 5, pp. 433–441, Mar. 2022, doi: 10.1049/cmu2.12268.
Y. Lu, Y.-M. Cheung, and Y. Y. Tang, “Bayes Imbalance Impact Index: A measure of class imbalanced data set for classification problem,” IEEE Trans. Neural Netw. Learn. Syst., vol. 31, no. 9, pp. 3525–3539, Sep. 2020, doi: 10.1109/TNNLS.2019.2944962.
L. Wang, Y. Zhang, F. Li, C. Li, and H. Xu, “Mortality prediction of inpatients with NSTEMI in a premier hospital in China based on stacking model,” PLoS One, vol. 19, no. 12, Art. no. e0312448, Dec. 2024, doi: 10.1371/journal.pone.0312448.
X. Song et al., “Predicting 7-day unplanned readmission in elderly patients with coronary heart disease using machine learning,” Front. Cardiovasc. Med., vol. 10, Art. no. 1190038, Aug. 2023, doi: 10.3389/fcvm.2023.1190038.
Y. Sun et al., “Borderline SMOTE algorithm and feature selection-based network anomalies detection strategy,” Energies, vol. 15, no. 13, Art. no. 4751, Jun. 2022, doi: 10.3390/en15134751.
S. Maldonado, C. Vairetti, A. Fernández, and F. Herrera, “FW-SMOTE: A feature-weighted oversampling approach for imbalanced classification,” Pattern Recognit., vol. 124, Art. no. 108511, Apr. 2022, doi: 10.1016/j.patcog.2021.108511.
A. C. O. Bwire, “Loan asset indicators and commercial bank fragility in Kenya,” Eur. J. Manag. Issu., vol. 29, no. 1, pp. 25–36, Mar. 2021, doi: 10.15421/192103.
B. Zhang, “Application of innovative risk early warning model based on big data technology in Internet credit financial risk,” J. Inf. Technol. Res., vol. 15, no. 1, pp. 1–12, Aug. 2022, doi: 10.4018/jitr.299920.
H. Li, “Embedded microprocessor wireless communication data collection aids in early warning of default risk for Internet finance bank customers,” J. Sens., vol. 2021, Art. no. 1679907, 2021, doi: 10.1155/2021/1679907.
Additional Files
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Gupita Nurmalita Sari, Rujianto Eko Saputro, Giat Karyono
This work is licensed under a Creative Commons Attribution 4.0 International License.
</a
