Volatility-Sensitive Forecasting: Deep Learning Model Robustness across Calm and Crisis Periods
DOI:
https://doi.org/10.34021/ve.2026.09.01(2)Keywords:
STL-based volatility, deep learning, forecasting performance, COVID-19, Russian–Ukrainian conflictAbstract
This study investigates how the forecasting performance of deep learning models is affected by changing market conditions, with particular emphasis on periods of differing volatility. While deep learning methods are increasingly applied in financial forecasting, relatively limited evidence exists on how their predictive accuracy varies across distinct market regimes. The aim of this paper is to examine whether model performance systematically deteriorates during more turbulent periods and whether this effect differs across model architectures. The empirical analysis covers multiple asset classes, including equities, commodities, cryptocurrencies, and currency pairs, over three partially overlapping periods (2016–2018, 2018–2020, and 2020–2022), capturing both relatively stable and highly volatile market environments. Forecasting performance is evaluated using several recurrent neural network architectures, and the relationship between market volatility and prediction error is analysed using cross-sectional regression techniques. Additional sensitivity analyses are conducted to assess the robustness of the results across alternative asset samples. The results reveal a consistent positive relationship between market volatility and forecasting error, indicating that predictive performance tends to deteriorate under more turbulent market conditions. However, the magnitude of this effect varies across model architectures. In particular, GRU-based architectures appear less sensitive to volatility, whereas the use of multivariate inputs does not consistently improve forecasting accuracy. These findings suggest that greater model complexity does not necessarily translate into more robust performance under changing market conditions. The study is subject to several limitations, including the relatively small cross-sectional sample and the use of standardised model configurations. Future research could extend the analysis by considering alternative model classes, broader datasets, and more comprehensive robustness frameworks.
Downloads
References
1. Mirza, F. K., Pekcan, Ö., Hekimoğlu, M., & Baykaş, T. (2025). Stock price forecasting through symbolic dynamics and state transition graphs with a convolutional recurrent neural network architecture. Neural Computing and Applications, 37, 15855-15890. https://doi.org/10.1007/s00521-025-11325-z
2. Glova, A. M. G., & Barrios, E. B. (2025). Modelling Mixed-Frequency Time Series with Structural Change. Computational Economics, 65, 3237-3258. https://doi.org/10.1007/s10614-024-10672-8
3. Pramod, A., Naicker, H. S., & Tyagi, A. K. (2021). Machine learning and deep learning: Open issues and future research directions for the next 10 years. In A. K. Tyagi (Ed.), Computational analysis and deep learning for medical care: Principles, methods, and applications (pp. 463-490). Wiley. https://doi.org/10.1002/9781119785750.ch18
4. Sakib, M., Mustajab, S., & Alam, M. (2025). Ensemble deep learning techniques for time series analysis: a comprehensive review, applications, open issues, challenges, and future directions. Cluster Computing, 28, Article 73(2025). https://doi.org/10.1007/s10586-024-04684-0
5. Li, G., & Jung, J. J. (2023). Deep learning for anomaly detection in multivariate time series: Approaches, applications, and challenges. Information Fusion, 91, 93-102. https://doi.org/10.1016/j.inffus.2022.10.008
6. Tissaoui, K., Boubaker, S., Hkiri, B., & Azibi, N. (2024). Multi-model transfer function approach tuned by PSO for predicting stock market implied volatility explained by uncertainty indexes. Scientific Reports, 14, Article 22711. https://doi.org/10.1038/s41598-024-74456-8
7. Zhu, Y., Taasim, S. I., & Daud, A. (2025). Volatility modelling and tail risk estimation of financial assets: Evidence from gold, oil, Bitcoin, and stocks for selected markets. Risks, 13(7), Article 138. https://doi.org/10.3390/risks13070138
8. Sivakumar, G. (2025). HMM-LSTM Fusion Model for Economic Forecasting. arXiv preprint arXiv:2501.02002. https://doi.org/10.48550/arXiv.2501.02002
9. El Zaar, A., Mansouri, A., Benaya, N., Bakir, T., & El Allati, A. (2025). Hybrid Transformer-CNN architecture for multivariate time series forecasting: Integrating attention mechanisms with convolutional feature extraction. Journal of Intelligent Information Systems, 63, 1233-1264. https://doi.org/10.1007/s10844-025-00937-5
10. Bhogade, V., & Nithya, B. (2024). Time series forecasting using transformer neural network. International Journal of Computers and Applications, 46(10), 880-888. https://doi.org/10.1080/1206212X.2024.2396321
11. Kafi, A., Rahman, M. M., Arefin, S M S., Sultana, N., Rafi, M.A., Zakaria, R.M. (2025). Forecasting currency volatility in global markets using transformer models: Implications for US trade and investment strategies. Journal of Economics, Finance and Accounting Studies, 7(5), 30-43. https://doi.org/10.32996/jefas.2025.7.5.4
12. Lo, A. W. (2004). The Adaptive Markets Hypothesis: Market efficiency from an evolutionary perspective. The Journal of Portfolio Management, 30, 15-29.
13. Hastie, T., Friedman, J. H., & Tibshirani, R. (2009). The elements of statistical learning: Data mining, inference, and prediction. Springer. https://doi.org/10.1007/978-0-387-21606-5
14. Zhang, X., & Zhang, T. (2022). Barrier option pricing under a Markov Regime switching diffusion model. The Quarterly Review of Economics and Finance, 86, 273-280. https://doi.org/10.1016/j.qref.2022.08.001
15. Carpinteyro, M., Venegas-Martínez, F., & Aali-Bujari, A. (2021). Modelling precious metal returns through fractional jump-diffusion processes combined with Markov regime-switching stochastic volatility. Mathematics, 9(4), Article 407. https://doi.org/10.3390/math9040407
16. Shi, Y. (2023). Long memory and regime switching in the stochastic volatility modelling. Annals of Operations Research, 320, 999-1020. https://doi.org/10.1007/s10479-020-03841-z
17. Avinash, G., Ramasubramanian, V., Paul, R. K., Ray, M., Dahiya, S., Iquebal, M. A., Godara, S., & Manjunatha, B. (2024). Price Forecasting of TOP (Tomato, Onion and Potato) Commodities using Hidden Markov-based Deep Learning Approach. Statistics and Applications, 22(2), 1-28.
18. Aydogan-Kilic, D., & Selcuk-Kestel, A. S. (2023). Modification of hybrid RNN-HMM model in asset pricing: univariate and multivariate cases. Applied Intelligence, 53, 23812-23833. https://doi.org/10.1007/s10489-023-04762-7
19. Mari, C., & Mari, E. (2023). Deep learning based regime-switching models of energy commodity prices. Energy Systems, 14, 913-934. https://doi.org/10.1007/s12667-022-00515-6
20. Mari, C., & Mari, E. (2025). Stochastic DNN-based models meet hidden Markov models: A challenge on natural gas prices at the Henry Hub. Neural Computing and Applications, 37, 14997–15016. https://doi.org/10.1007/s00521-025-11286-3
21. Ouyang, Z. S., Yang, X. T., & Lai, Y. (2021). Systemic financial risk early warning of financial market in China using Attention-LSTM model. The North American Journal of Economics and Finance, 56, Article 101383. https://doi.org/10.1016/j.najef.2021.101383
22. Zhang, J., Ye, L., & Lai, Y. (2023). Stock price prediction using CNN-BiLSTM-Attention model. Mathematics, 11(9), Article 1985. https://doi.org/10.3390/math11091985
23. Luo, A., Zhong, L., Wang, J., Wang, Y., Li, S., & Tai, W. (2024). Short-term stock correlation forecasting based on CNN-BiLSTM enhanced by attention mechanism. IEEE Access, 12, 29617-29632. https://doi.org/10.1109/ACCESS.2024.3369419
24. Yu, H. (2025). An Adaptive Sharpe Ratio-Based Temporal Fusion Transformer for Financial Forecasting. Frontiers in Humanities and Social Research, 2(1), 54-65. https://doi.org/10.71465/fhsr23
25. Li, S., & Xu, S. (2025). Enhancing stock price prediction using GANs and transformer-based attention mechanisms. Empirical Economics, 68, 373-403. https://doi.org/10.1007/s00181-024-02644-6
26. Yahoo. (n.d.). Yahoo Finance. https://finance.yahoo.com/
27. CoinMarketCap. (n.d.). CoinMarketCap. https://coinmarketcap.com/
28. Bai, Y., Xie, J., Liu, C., Tao, Y., Zeng, B., & Li, C. (2021). Regression modelling for enterprise electricity consumption: A comparison of recurrent neural network and its variants. International Journal of Electrical Power & Energy Systems, 126, Artile 106612. https://doi.org/10.1016/j.ijepes.2020.106612
29. Xiao, H., Chen, Z., Cao, R., Cao, Y., Zhao, L., & Zhao, Y. (2022). Prediction of shield machine posture using the GRU algorithm with adaptive boosting: A case study of Chengdu Subway project. Transportation Geotechnics, 37, Article 100837. https://doi.org/10.1016/j.trgeo.2022.100837
30. Ortu, M., Uras, N., Conversano, C., Bartolucci, S., & Destefanis, G. (2022). On technical trading and social media indicators for cryptocurrency price classification through deep learning. Expert Systems with Applications, 198, Article 116804. https://doi.org/10.1016/j.eswa.2022.116804
31. Zhao, L., Li, Z., Qu, L., Zhang, J., & Teng, B. (2023). A hybrid VMD-LSTM/GRU model to predict non-stationary and irregular waves on the east coast of China. Ocean Engineering, 276, Article 114136. https://doi.org/10.1016/j.oceaneng.2023.114136
32. Islam, M. S., & Hossain, E. (2021). Foreign exchange currency rate prediction using a GRU-LSTM hybrid network. Soft Computing Letters, 3, Article 100009. https://doi.org/10.1016/j.socl.2020.100009
33. Faru, S. H., Waititu, A., & Nderu, L. (2023). A Hybrid Neural Network Model Based on Transfer Learning for Forecasting Forex Market. Journal of Data Analysis and Information Processing, 11(2), 103-120. https://doi.org/10.4236/jdaip.2023.112007
34. Cleveland, R. B., Cleveland, W. S., McRae, J. E., & Terpenning, I. (1990). STL: A seasonal-trend decomposition procedure based on loess. Journal of Official Statistics, 6(1), 3–73.
35. He, H., Gao, S., Jin, T., Sato, S., & Zhang, X. (2021). A seasonal-trend decomposition-based dendritic neuron model for financial time series prediction. Applied Soft Computing, 108, Article 107488. https://doi.org/10.1016/j.asoc.2021.107488
36. Brockwell, P. J., & Davis, R. A. (2016). Introduction to time series and forecasting. Springer.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Laszlo Vancsura, Tibor Tatay, Tibor Bareith
This work is licensed under a Creative Commons Attribution 4.0 International License.
