Application of vine copulas and gradient boosting for forecasting the price volatility of iron ore and coking coal

DOI: https://doi.org/10.30686/1609-9192-2026-3-142-150

Читать на русскоя языке E.N. Ramazanova, G.N. Kamyshova, S.N. Pozdeeva, I.V. Zaychikova
Financial University under the Government of the Russian Federation, Moscow, Russian Federation
Russian Mining Industry №3/ 2026 p. 142-150

Abstract: Iron ore and premium coking coal remain the most volatile benchmarks in the steelmaking raw materials segment: throughout 2022–2025, their quotations exhibited abrupt reversals driven by the steel production cycles in China and India, weather shocks in the Pilbara and Queensland regions, Capesize freight fluctuations, and the sanctions-driven reconfiguration of the Russian exports. Standard univariate conditional heteroscedasticity models fail to capture the asymmetric joint tail dependence between these two key burden materials, while the Gaussian copulas substantially underestimate synchronous extreme co-movements during stress episodes. The paper proposes and tests a hybrid approach, i.e. standardised residuals from the ARMA(1,1)–GJR-GARCH(1,1) marginals with skewed Student-t innovations are linked through a regular vine (R-vine) copula constructed via the Dißmann algorithm with sequential selection among the Gumbel, Clayton, BB1, BB7 and rotated pair families, enabling separate treatment of the upper and lower tails. The obtained conditional dependence parameters, realised variance from five-minute returns, and a block of exogenous factors, i.e. the Shanghai HRC futures, Baltic Capesize Index, stockpiles at the 45 largest Chinese ports, AUD/USD, DXY and the Mongolia–Australia spread, are fed into an XGBoost gradient boosting model with the Bayesian hyperparameter tuning. The empirical base covers daily Platts IODEX 62% Fe CFR China and Platts PLV HCC FOB Australia quotations from January 2022 to December 2025, extended with corresponding Russian Far East coking coal benchmarks. The combination of R-vine and gradient boosting materially reduces one-day-ahead realised volatility forecast errors relative to the DCC–GARCH, HAR–RV and LSTM baselines while maintaining the correct VaR coverage at the 1% and 5% confidence levels. The toolkit is applicable to hedging the burden basket, calibrating extreme risk limits, and to tariff and capital expenditure planning of mining companies.

Keywords: iron ore, coking coal, vine copula, gradient boosting, volatility forecasting, tail dependence, mining industry

For citation: Ramazanova E.N., Kamyshova G.N., Pozdeeva S.N., Zaychikova I.V. Application of vine copulas and gradient boosting for forecasting the price volatility of iron ore and coking coal. Russian Mining Industry. 2026;(3):142–150. https://doi.org/10.30686/1609-9192-2026-3-142-150

Article info

Received: 16.02.2026

Revised: 06.05.2026

Accepted: 06.05.2026

Information about the authors

Elvira N. Ramazanova – Cand. Sci. (Eng.), Associate Professor, Financial University under the Government of the Russian Federation, Moscow, Russian Federation; https://orcid.org/0000-0003-1802-8012; e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

Galina N. Kamyshova – Cand. Sci. (Phys.-Math.), Associate Professor, Financial University under the Government of the Russian Federation, Moscow, Russian Federation; https://orcid.org/0000-0002-8569-6259; e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

Svetlana N. Pozdeeva – Cand. Sci. (Econ.), Associate Professor, Financial University under the Government of the Russian Federation, Moscow, Russian Federation; https://orcid.org/0000-0001-8421-5858; e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

Inna V. Zaychikova – Cand. Sci. (Educ.), Associate Professor of the Department of Mathematics and Data Analysis, Financial University under the Government of the Russian Federation, Moscow, Russian Federation; https://orcid.org/0000-0002-1348-1929; e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

References

1. Chung S. Modelling and forecasting energy market volatility using GARCH and machine learning approach. Preprint. arXiv:2405.19849. 30 May 2024. https://doi.org/10.48550/arXiv.2405.19849

2. Han X., Liu Z., Wang S. An R-vine copula analysis of non-ferrous metal futures with application in Value-at-Risk forecasting. Journal of Commodity Markets. 2022;25:100188. https://doi.org/10.1016/j.jcomm.2021.100188

3. Zeng L., Huang J., Lin X. LSTM-augmented vine copula modelling for energy-finance contagion analysis. Scientific Reports. 2026;16(1):5358. https://doi.org/10.1038/s41598-026-37150-5

4. Yao Y., Chen X., Chen Z. Portfolio tail risk forecasting for international financial assets: A GARCH-MIDAS-R-Vine copula model. The North American Journal of Economics and Finance. 2025;77:102385. https://doi.org/10.1016/j.najef.2025.102385

5. Lazar E., Pan J., Wang S. On the estimation of Value-at-Risk and Expected Shortfall at extreme levels. Journal of Commodity Markets. 2024;34;100391. https://doi.org/10.1016/j.jcomm.2024.100391

6. Ewees A.A., Abd Elaziz M., Alameer Z., Ye H., Jianhua Z. Improving multilayer perceptron neural network using chaotic grasshopper optimization algorithm to forecast iron ore price volatility. Resources Policy. 2020;65:101555. https://doi.org/10.1016/j.resourpol.2019.101555

7. Souza F.M.C., Filho G.P.R., Guimarães F.G., Meneguette R.I., Pessin G. Navigating market sentiments: a novel approach to iron ore price forecasting with weighted fuzzy time series. Information. 2024;15(5):251. https://doi.org/10.3390/info15050251

8. Pan W., Liu S.Q., Kumral M., D’Ariano A., Masoud M., Khan W.A., Bakather A. Iron ore price forecast based on a multiechelon tandem learning model. Natural Resources Research. 2024;33(5):1969–1992. https://doi.org/10.1007/s11053-024-10360-2

9. Glebova A.G., Kovaleva A.A. Forecasting the volatility of the Russian stock market in the context of international economic sanctions. Finance: Theory and Practice. 2024;28(1):20–29. https://doi.org/10.26794/2587-5671-2024-28-1-20-29

10. Aksenov D.A., Toropov V.V., Mazurchuk T.M. Short-term forecasting and investment in Russian oil and gas companies taking into account benchmark markers. Finance: Theory and Practice. 2025;29(5):164–177. https://doi.org/10.26794/2587-5671-2025-29-5-164-177

11. Byvshev V.A., Yashchenko M.A. Assessment of the volatility of the main cryptocurrencies, the euro and the direct exchange rate of the ruble. Finance: Theory and Practice. 2024;28(1):133–144. https://doi.org/10.26794/2587-5671-2024-28-1-133-144

12. Li M., Yang L. Modeling the volatility of futures return in rubber and oil – A Copula-based GARCH model approach. Economic Modelling. 2013;35:576–581. https://doi.org/10.1016/j.econmod.2013.07.016

13. Kim G., Silvapulle M.J., Silvapulle P. Comparison of semiparametric and parametric methods for estimating copulas. Computational Statistics & Data Analysis. 2007;51(6):2836–2850. https://doi.org/10.1016/j.csda.2006.10.009

14. Li H., Liu Z., Wang S. Vines climbing higher: Risk management for commodity futures markets using a regular vine copula approach. International Journal of Finance & Economics. 2022;27(2):2438–2457. https://doi.org/10.1002/ijfe.2280

15. Ma Y., Wang J. Co-movement between oil, gas, coal, and iron ore prices, the Australian dollar, and the Chinese RMB exchange rates: A copula approach. Resources Policy. 2019;63:101471. https://doi.org/10.1016/j.resourpol.2019.101471

16. Babeshko L.O., Byvshev V.A. Analysis of the stability of the model for forecasting mutual volumes Russia’s trade with BRICS partners. Finance: Theory and Practice. 2025;29(4):129–145. https://doi.org/10.26794/2587-5671-2025-29-4-1902-01

17. Araya H.T., Aduda J., Berhane T. A hybrid GARCH and deep learning method for volatility prediction. Journal of Applied Mathematics. 2024;2024:6305525. https://doi.org/10.1155/2024/6305525

18. International Energy Agency. Coal 2025: Analysis and forecast to 2030. Paris: IEA; 2025. 126 p. Available at: https://www.iea.org/reports/coal-2025 (accessed: 15.04.2026).

19. Tyumentseva E.Yu., Shamis V.A., Mukhametdinova S.Kh. Factors having the most significant impact on university students’ level of financial literacy: Forecasting based on cognitive methodology. Science for Education Today. 2023;13(5):124–140. (In Russ.) Available at: http://sciforedu.ru/system/files/articles/pdf/13_tyumentseva_5-23.pdf (accessed: 22.03.2026).

20. Samarin I.V. Application of deep learning and satellite data for monitoring and forecasting forest fires in Russia: performance and perspective analysis. Voprosy Ecologii. 2024;37(1);128–155. (In Russ.) https://doi.org/10.25726/m7116-1845-7217-x

21. Kormiltseva E.A., Baygusheva I.A., Varova N.L., Starikov I.V., Shmakova A.P., Burmistrova N.A. Study of the future economists’ readiness to use artificial intelligence based on the hierarchy analysis method. Science for Education Today. 2024:14(3):113–134. (In Russ.) https://doi.org/10.15293/2658-6762.2403.06

22. Wang Y., Guo Z., Zhang Y., Hu X., Xiao J. Iron ore price prediction based on multiple linear regression model. Sustainability. 2023;15(22):15864. https://doi.org/10.3390/su152215864