Development of a conceptual integrated approach combining cluster analysis and machine learning for credit risk assessment

In the context of the rapid growth in volume and complexity of financial data, classical credit scoring models face limitations in forecast accuracy and sustainability. At the same time, clustering and machine learning methods are evolving, enabling the identification of latent borrower segments and the integration of this additional information directly into predictive models. This paper addresses the methodological gap between data segmentation and machine learning in credit risk assessment. The goal of the study is to develop a concept of a hybrid model that combines clustering analysis and gradient boosting algorithms by incorporating cluster meta-features into the feature space of a scoring model. The methodological framework includes an analysis for existing approaches to credit scoring, the design of a sequential data-processing pipeline, and the construction of a feature set that reflects the borrower’s structural position in the data space. In addition, the study evaluates the stability of clustering results with respect to the choice of metric, the number of clusters, and sample characteristics, as well as compares different ways of encoding cluster information. The proposed approach is aimed at improving model discriminatory power by accounting for latent customer segmentation. The expected deliverable is enhanced default prediction quality and expanded model interpretability, which increases their practical value for credit risk management and decision support in financial institutions.

1. New horizons: how we implement AI in working with data and the Data Quality framework at VTB // Habr. URL: https://habr.com/ru/companies/vtb/articles/732216/ (date of access: 20.01.2026).

2. Application of artificial intelligence in the financial market: a report for public consultations / Central Bank of the Russian Federation. Moscow, 2023. URL: https://www.cbr.ru/Content/Document/File/156061/Consultation_Paper_03112023.pdf (date of access: 20.01.2026).

3. Machine learning & artificial intelligence powered credit scoring models for Islamic microfinance institutions: a blockchain approach / M.M.H. Mukit, M.A. Rahman, M.A. Islam [et al.] // Risks. 2026. Vol. 14, No. 1. Art. 12. doi:10.3390/risks14010012. URL: https://www.mdpi.com/2227-9091/14/1/12 (date of access: 08.02.2026).

4. Malaya P.P., Tyunin E.B. Investigation of methods for assessing borrowers' creditworthiness and development of an information scoring subsystem for a bank // Financial markets and banks. 2024. No. 11. Pp. 216–219. URL: https://library.cbr.ru/catalog/lib/article/979343/ (date of access: 18.01.2026).

5. Improving credit scoring methods using machine learning / V.V. Dragulenko, E.V. Potekhina, A.L. Zolkin, O.S. Buryakova // Financial Management. 2024. No. 7. Pp. 241–250.

6. Ileberi E., Sun Y., Wang Z. A machine learning-based credit risk prediction engine system using a stacked classifier and a filter-based feature selection method // Journal of Big Data. 2024. Vol. 11. Art. 23. doi:10.1186/s40537-024-00882-0. URL: https://journalofbigdata.springeropen.com/articles/10.1186/s40537-024-00882-0 (date of access: 08.02.2026).

7. Bosker J.M., Görtler M., Zöllner M. Machine learning-based variable selection for clustered credit risk modeling // Journal of Business Economics. 2025. Vol. 95. Pp. 617–652. doi:10.1007/s11573-024-01213-8. URL: https://link.springer.com/article/10.1007/s11573-024-01213-8 (date of access: 08.02.2026).

8. Ayari H., Guetari R., Kraïem N. Machine learning powered financial credit scoring: a systematic literature review // Artificial Intelligence Review. 2026. Vol. 59. Art. 13. doi:10.1007/s10462-025-11416-2. URL: https://link.springer.com/article/10.1007/s10462-025-11416-2 (date of access: 08.02.2026).

9. SAS Enterprise Miner // SAS : [the official website of the software]. URL: https://www.sas.com/en_us/software/enterprise-miner.html (date of access: 05.02.2026).

10. Altair RapidMiner // Altair Engineering, Inc. : [the official website of the data analysis platform]. URL: https://altair.com/altair-rapidminer (date of access: 05.02.2026).

11. Zinovieva E.A. Modernization of banking products in the digital economy : dis. ... Candidate of Economic Sciences. Moscow, 2024. URL: https://www.fa.ru/upload/iblock/a55/4x3a5oy0ksllzlnacdrsbyauz2fupzkv/Zinoveva-107_Dissertatsiya.pdf (date of access: 18.01.2026).

12. Wang H., Mi J. Intuitive-K-prototypes: a mixed data clustering algorithm with intuitionistic distribution centroid // Pattern Recognition. 2025. Vol. 158. Art. 111062. doi:10.1016/j.patcog.2024.111062. URL: https://www.sciencedirect.com/science/article/abs/pii/S0031320324005905 (date of access: 08.02.2026).

13. Bridging accuracy and interpretability: a rescaled cluster-then-predict approach for enhanced credit scoring / H.-W. Teng, M.-H. Kang, I.-H. Lee, L.-C. Bai // International Review of Financial Analysis. 2024. Vol. 91. Art. 103005. doi:10.1016/j.irfa.2023.103005. URL: https://www.sciencedirect.com/science/article/abs/pii/S1057521923007888 (date of access: 08.02.2026).

14. Multidimensional data analysis in the task of predicting the entry of credit institutions into the risk zone / E.P. Akishina, V.V. Ivanov, A.V. Kryanev, A.S. Prikazchikova // Bulletin of the National Research Nuclear University "MEPhI". 2024. Vol. 13, No. 1. Pp. 22–29. doi:10.26583/vestnik.2024.302.

15. Lugueva A.S., Nukhova A.M., Kadyrov S.S. Machine learning models for predicting defaults in the banking sector: sustainability assessment // Economics and management: problems, solutions. 2024. Vol. 6, No. 12 (ч. 6). Pp. 156–162. URL: https://library.cbr.ru/catalog/lib/article/982761/ (date of access: 08.02.2026).

16. Zhumashev A., Kuatbayeva A.A. Comparative overview of machine learning models for credit scoring // Universum: Technical sciences. 2025. Vol. 5, Issue 134. Pp. 4–9. doi:10.32743/UniTech.2025.134.5.20167. URL: https://7universum.com/ru/tech/archive/item/20167 (date of access: 08.02.2026).

17. Saburov S.S. Methods and models of machine learning for assessing credit risks : [final qualifying thesis] / scientific supervisor A.V. Loginova. St. Petersburg : Peter the Great St. Petersburg Polytechnic University, 2025. doi:10.18720/SPBPU/3/2025/vr/vr25-80. URL: https://elib.spbstu.ru/dl/3/2025/vr/vr2580.pdf/en/info (date of access: 18.01.2026).

18. Ponomarev D.S. Library comparison for creating machine learning models based on gradient boosting methods // Modern Innovations, Systems and Technologies. 2025. Vol. 5, No. 2. Pp. 3001–3006. doi:10.47813/2782-2818-2025-5-2-3001-3006. URL: https://oajmist.com/index.php/1/article/view/182 (date of access: 08.02.2026).

19. Khuzyakhanov E.I., Sharko S.S., Polyakov P.M. Effective data preprocessing methods: how to prepare data for machine learning // Modern scientific research and innovation. 2024. No. 11. URL: https://web.snauka.ru/issues/2024/11/102801 (date of access: 18.01.2026).

20. An integrated machine learning and deep learning framework for credit card approval prediction / K. Tong, Z. Han, Y. Shen [et al.]. 2024. URL: https://arxiv.org/abs/2409.16676 (date of access: 18.01.2026).

21. Comparative analysis of machine learning technologies for credit scoring tasks / M.A. Tatarintsev, P.V. Nikitin, R.I. Gorokhova, V.I. Dolgov // Fundamental Research. 2023. No. 1. Pp. 49–53. URL: https://fundamental-research.ru/ru/article/view?id=43419 (date of access: 08.02.2026).

22. Rastorguev L.M., Mikishanina E.A. Building a scoring model to determine the creditworthiness of a potential borrower // Vector of Economics. 2022. No. 5. URL: https://www.vectoreconomy.ru/images/publications/2022/5/finance/Rastorguev_Mikishanina.pdf (date of access: 08.02.2026).

23. Zahid N. Loan Prediction Dataset 2025. Synthetic dataset for loan prediction // Kaggle. URL: https://www.kaggle.com/datasets/nabihazahid/loan-prediction-dataset-2025 (date of access: 05.02.2026).