Self-Adaptive Machine Learning Models for Financial Risk‎ Forecasting: Handling Non-Stationarity in Banking and Cryptocurrency Time Series

  • Authors

    • Proshanta Kumar Bhowmik MBA in Business Analytics, Trine University, Angola, IN, USA
    • Dil Tabassum Subha Master of Science in Business Analytics, Grand Canyon University, USA
    • Abdur Rahim Master of Science in Computer Science, University of New Haven
    • Arshad Ali Mohammed College of Graduate and Professional Studies, Trine University
    • Maksuda Begum Master of Business Administration, Trine University
    • Redoyan Chowdhury Management Information System (MIS), International American University
    • Tanveer Ehsan Chowdhury Management Information System (MIS), International American University (IAU)
    • Mohammed Wajeed Masters of Science in Information Technology Management, St.Francis College
    • Dr. Khan Mohammad Elyas International Development Consultant and Researcher, Innovation and Incubation Center for Enterprises (IICE)
    • Mahamuda Akter Shati Master's in Business Analytics, Grand Canyon University
    https://doi.org/10.14419/7x458m39

    Received date: January 17, 2026

    Accepted date: February 7, 2026

    Published date: February 15, 2026

  • Adaptive Learning, Banking Risk; Cryptocurrency; Financial Risk Forecasting; Non-Stationarity

  • Abstract

    Financial markets rarely sit still. They drift, jump, and occasionally snap into entirely new regimes, with volatility behaving differently depending on the moment. These shifts steadily erode the usefulness of static risk forecasting models. Approaches built on fixed parameters or scheduled retraining tend to lag behind reality, especially when markets are under stress and conditions change fast. This paper investigates whether self-adaptive machine learning models can generate reliable and practical volatility forecasts in live settings, without relying on repeated full retraining cycles. Using daily closing prices from the S&P 500 index and Bitcoin as examples of traditional and cryptocurrency assets, the task is set up as a strictly out-of-sample, one-step-ahead volatility forecasting problem. We compare static models, periodically retrained models, and fully self-adaptive models across calm periods and clearly defined stress regimes, including the COVID-19 crash and major cryptocurrency boom and bust cycles. The results show that self-adaptive models deliver stronger performance across both asset classes. They achieve lower forecast errors, more stable errors during stress, quicker recovery after shocks, and sharply lower computational demands. Recursive EWMA-based models stand out for their solid performance using constant memory and lightweight updates, while online gradient-based learners adapt flexibly without drifting into instability. Taken together, the evidence points to continuous adaptation as a structurally superior approach for financial risk forecasting under non-stationarity. The study demonstrates that self-adaptive models can be deployed in real-time risk systems and sets the stage for future work that connects adaptive forecasting with evolving explainability and regulatory reporting requirements.

  • References

    1. Aashish, K. C., Zamil, M. Z. H., Mridul, M. S. I., Akter, L., Sharmin, F., Ayon, E. H., … Malla, S. (2025). Towards eco-friendly cybersecurity: Machine learning-based anomaly detection with carbon and energy metrics. International Journal of Applied Mathematics, 38(9s). https://doi.org/10.12732/ijam.v38i9s.822.
    2. Andersen, T. G., Bollerslev, T., Diebold, F. X., & Labys, P. (2003). Modeling and forecasting realized volatility. Econometrica, 71(2), 579–625. https://doi.org/10.1111/1468-0262.00418.
    3. Bollerslev, T. (1986). Generalized autoregressive conditional heteroskedasticity. Journal of Econometrics, 31(3), 307–327. https://doi.org/10.1016/0304-4076(86)90063-1.
    4. Cavalcante, R. C., Oliveira, G. H. F. M., & Minku, L. L. (2016). An approach to handle concept drift in financial time series based on online se-quential extreme learning machines and explicit drift detection. In Proceedings of the International Joint Conference on Neural Networks (IJCNN). https://doi.org/10.1109/IJCNN.2015.7280721.
    5. Chouksey, A., Dola, A., Antara, U. K., Begum, S., Ahmed, T., Sultana, T., & Zabin, N. (2025). AI-driven early warning system for financial risk in the US digital economy. International Journal of Applied Mathematics, 38(9s). https://doi.org/10.12732/ijam.v38i9s.838.
    6. Danielsson, J., & de Vries, C. G. (1997). Evaluating the RiskMetrics methodology in measuring volatility. Journal of Risk, 1(4), 5–26.
    7. Das, B. C., et al. (2025). AI-driven cybersecurity threat detection: Building resilient defense systems using predictive analytics. arXiv preprint arXiv:2508.01422. https://doi.org/10.14419/hysdg957.
    8. Debnath, S., et al. (2025). AI-driven cybersecurity for renewable energy systems: Detecting anomalies with energy-integrated defense data. Inter-national Journal of Applied Mathematics, 38(5s). https://doi.org/10.12732/ijam.v38i5s.367.
    9. Dyhrberg, A. H. (2016). Bitcoin, gold and the dollar: A GARCH volatility analysis. Finance Research Letters, 16, 85–92. https://doi.org/10.1016/j.frl.2015.10.008.
    10. Fedorov, V., Panchenko, V., & Shpilkin, A. (2019). Stock price volatility prediction with long short-term memory. CS230: Deep Learning, Stan-ford University Project Report.
    11. Gama, J., Žliobaitė, I., Bifet, A., Pechenizkiy, M., & Bouchachia, A. (2014). A survey on concept drift adaptation. ACM Computing Surveys, 46(4), 44:1–44:37. https://doi.org/10.1145/2523813.
    12. Hasan, M. R., Rahman, M. A., Gomes, C. A. H., Nitu, F. N., Gomes, C. A., Islam, M. R., & Shawon, R. E. R. (2025). Building robust AI and ma-chine learning models for supplier risk management: A data-driven strategy for enhancing supply chain resilience in the USA. Advances in Con-sumer Research, 2(4).
    13. Hasan, M. S., et al. (2025). Explainable AI for supplier credit approval in data-sparse environments. International Journal of Applied Mathematics, 38(5s). https://doi.org/10.12732/ijam.v38i5s.380.
    14. Islam, M. Z., et al. (2025). Cryptocurrency price forecasting using machine learning: Building intelligent financial prediction models. arXiv preprint arXiv:2508.01419. https://doi.org/10.14419/s0pktr58.
    15. Jakir, T. (2025). Signal-to-noise analysis of crisis indicators in global finance using artificial intelligence. International Journal of Applied Mathemat-ics, 38(10s), 1815–1836. https://doi.org/10.12732/ijam.v38i10s.1075.
    16. Mocanu, D. C., Fidon, L., & Sminchisescu, C. (2015). Online learning for financial time series prediction. In Lecture Notes in Computer Science (Vol. 8955, pp. 419–428).
    17. Neri, F. (2021). Domain-specific concept drift detectors for predicting financial time series. arXiv preprint arXiv:2103.14079.
    18. Oliveira, G. H. F. M., Cavalcante, R. C., Cabral, G. G., Minku, L. L., & Lacerda, A. L. F. (2017). Time series forecasting in the presence of concept drift: A PSO-based approach. In Proceedings of the IEEE International Conference on Tools with Artificial Intelligence (ICTAI) (pp. 239–246). https://doi.org/10.1109/ICTAI.2017.00046.
    19. Proietti, T. (2014). Exponential smoothing, long memory and volatility forecasting. MPRA Paper No. 57230. https://doi.org/10.2139/ssrn.2475784.
    20. Rahman, M. S. (2025). Machine learning–enabled early warning system for detecting micro-inflation clusters in the US economy. International Journal of Applied Mathematics, 38(12s), 2743–2769. https://doi.org/10.12732/ijam.v38i12s.1585.
    21. Ray, R. K. (2025). Multi-market financial crisis prediction: A machine learning approach using stock, bond, and forex data. International Journal of Applied Mathematics, 38(8s), 706–738. https://doi.org/10.12732/ijam.v38i8s.602.
    22. Reza, S. A., et al. (2025). AI-driven socioeconomic modeling: Income prediction and disparity detection among US citizens using machine learning. Advances in Consumer Research, 2(4).
    23. Reza, S. A., Rahman, M. K., Rahman, M. D., Sharmin, S., Mithu, M. F. H., Hasnain, K. N., ... & Kabir, R. (2025). Machine learning enabled early warning system for financial distress using real-time digital signals. arXiv preprint arXiv:2510.22287.
    24. RiskMetrics Group. (1996). RiskMetrics technical document (4th ed.). J.P. Morgan/Reuters.
    25. Shalev-Shwartz, S. (2012). Online learning and online convex optimization. Foundations and Trends in Machine Learning, 4(2), 107–194. https://doi.org/10.1561/2200000018.
    26. Shawon, R. E. R., et al. (2025). Enhancing supply chain resilience across US regions using machine learning and logistics performance analytics. International Journal of Applied Mathematics, 38(4s). https://doi.org/10.12732/ijam.v38i4s.225.
    27. Shawon, R. E. R., Buiya, M. R., Pant, S., Al Jobaer, M. A., Chowdhury, M. S. R., Kawsar, M., ... & Ali, M. (2025). Detecting illicit cross-chain fund movement: Behavioral machine learning models for bridge-based laundering patterns. International Journal of Applied Mathematics, 38(12s). https://doi.org/10.12732/ijam.v38i12s.1399.
    28. Shivogo, J. (2025). Fair and explainable credit-scoring under concept drift: Adaptive explanation frameworks for evolving populations. arXiv pre-print arXiv:2511.03807.
    29. Shovon, M. S. S. (2025). Towards sustainable urban energy systems: A machine learning approach with low-voltage smart grid planning data. In-ternational Journal of Applied Mathematics, 38(8s), 1115–1155. https://doi.org/10.12732/ijam.v38i8s.631.
    30. Sizan, M. M. H., et al. (2025). Machine learning-based unsupervised ensemble approach for detecting new money laundering typologies in transac-tion graphs. International Journal of Applied Mathematics, 38(2s). https://doi.org/10.12732/ijam.v38i2s.88.
    31. Velankar, S., & Allu, R. (2021). Stock price forecasting using a time-series long short-term memory network. Finance, Accounting and Business Analysis, 3(1), 42–58.
    32. Zumbach, G. (2007). From JP Morgan’s standard-EWMA to skewed-EWMA. MPRA Paper No. 1009022.

  • Downloads

  • How to Cite

    Bhowmik, P. K., Subha, D. T., Rahim, A., Mohammed, A. A., Begum, M., Chowdhury, R., Chowdhury, T. E., Wajeed, M., Elyas, D. K. M., & Shati , M. A. (2026). Self-Adaptive Machine Learning Models for Financial Risk‎ Forecasting: Handling Non-Stationarity in Banking and Cryptocurrency Time Series. International Journal of Accounting and Economics Studies, 13(2), 123-132. https://doi.org/10.14419/7x458m39