Forecasting of Wheat Yields in Türkiye under Local and Global Drivers: Integrating Production and Trade Data into Machine Learning Approaches
Keywords:
ARIMAX, Exogenous Variables, Machine Learning, NARX, Policy Regime Dummies, Precision Agriculture, Wheat Production, Yield ForecastingAbstract
Accurately predicting crop yields is essential for food security and guiding sustainable agricultural planning, especially under uncertainty. This study presents a comparative framework for forecasting wheat yield in Türkiye, combining two approaches: one of them is the Autoregressive Integrated Moving Average with Exogenous Variables (ARIMAX), and the other is the Nonlinear Autoregressive Model with Exogenous Inputs (NARX). The models combine wheat domestic data with international wheat and crude oil prices, exchange rates, and policy regime indicators to capture more precise results. Drawing on a dataset that spans 1998–2023, we assess performance across four distinct policy and economic periods. ARIMAX successfully models linear dynamics, but NARX consistently delivers higher accuracy when nonlinear interactions and regime shifts are considered. The best-performing NARX specification achieved a root mean squared error (RMSE) of 0.0842 and outperformed ARIMAX in the Diebold–Mariano test. Mean absolute percentage error (MAPE) results also confirm its predictive advantage. These results suggest that machine learning methods such as NARX can outperform traditional econometric models, particularly in small sample sizes and structural breaks. In addition to its methodological contributions, the framework provides practical benefits: it supplies policymakers with dependable predictions that can guide procurement strategies, subsidy initiatives, and risk management efforts focused on enhancing food security in Türkiye.
Downloads
References
1. Food and Agriculture Organization of the United Nations, “FAOSTAT Statistical Database,” 2024. [Online]. Available: https://www.fao.org/faostat/en/#data/QCL. Accessed: Jun. 26, 2025.
2. Turkish Statistical Institute (TÜİK), “Labour Force Statistics, 2022,” 2023. [Online]. Available: https://data.tuik.gov.tr/Bulten/Index?p=Labour-Force-Statistics-2022-49390&dil=2.
3. Ö. Vanli, I. Ahmad, and B. B. Ustundag, “Area estimation and yield forecasting of wheat in southeastern Turkey using a machine learning approach,” Journal of the Indian Society of Remote Sensing, vol. 48, no. 12, pp. 1757–1766, Dec. 2020, doi: 10.1007/s12524-020-01196-3.
4. M. Fahad, I. Ahmad, M. Rehman, M. M. Waqas, and F. Gul, “Regional wheat yield estimation by integration of remotely sensed soil moisture into a crop model,” Canadian Journal of Remote Sensing, vol. 45, no. 6, pp. 770–781, Nov. 2019, doi: 10.1080/07038992.2019.1692651.
5. S. Duru, A. Gül, and S. Hayran, “Türkiye’de buğday ve buğday mamulleri dış ticaret yapısı,” Bilecik Şeyh Edebali Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, pp. 552–564, 2019, doi: 10.33905/bseusbed.563347.
6. P. Pandit et al., “Hybrid time series models with exogenous variable for improved yield forecasting of major Rabi crops in India,” Scientific Reports, vol. 13, no. 1, 2023, doi: 10.1038/s41598-023-49544-w.
7. V. Kiran Kumar, K. V. Ramesh, and V. Rakesh, “Optimizing LSTM and Bi-LSTM models for crop yield prediction and comparison with traditional ML techniques,” Applied Intelligence, vol. 53, no. 23, pp. 28291–28309, 2023, doi: 10.1007/s10489-023-05005-5.
8. D. Devyatkin and Y. Otmakhova, “Methods for mid-term forecasting of crop export and production,” Applied Sciences, vol. 11, no. 22, p. 10973, 2021, doi: 10.3390/app112210973.
9. L. Götz, T. Glauben, and B. Brümmer, “Impacts of export controls on wheat markets during the food crisis 2007/2008 in Russia and Ukraine,” 2010. [Online]. Available: http://ageconsearch.umn.edu.
10. P. McMichael, “A food regime analysis of the ‘world food crisis’,” Agriculture and Human Values, vol. 26, no. 4, pp. 281–295, 2009, doi: 10.1007/s10460-009-9218-5.
11. O. Deschênes and M. Greenstone, “The economic impacts of climate change: Evidence from agricultural output and random fluctuations in weather,” American Economic Review, vol. 97, no. 1, pp. 354–385, 2007.
12. Z. Allam, S. E. Bibri, and S. A. Sharpe, “The rising impacts of the COVID-19 pandemic and the Russia–Ukraine war,” Resources, vol. 11, no. 11, 2022, doi: 10.3390/resources11110099.
13. T. Açikgöz, “Agricultural production connectedness and networks in Türkiye,” Journal of Tekirdag Agricultural Faculty, vol. 20, no. 4, pp. 799–810, 2023, doi: 10.33462/jotaf.1166050.
14. H. M. Ay and A. Söylemez, “Grain corridor agreement and Turkey’s role in the Russia-Ukraine war,” Journal of Islamic World and Politics, vol. 7, no. 1, pp. 1–10, 2023, doi: 10.18196/jiwp.v7i1.27.
15. A. Aydın, “Türkiye’de buğday üretim sektörünün yapısı ve ARIMA modeli ile üretim tahmini,” İşletme Ekonomi ve Yönetim Araştırmaları Dergisi, vol. 5, no. 1, pp. 1–18, 2022, doi: 10.33416/baybem.982635.
16. R. K. Paul and S. Garai, “Performance comparison of wavelet-based machine learning techniques for forecasting agricultural commodity prices,” Soft Computing, vol. 25, no. 20, pp. 12857–12873, 2021, doi: 10.1007/s00500-021-06087-4.
17. K. B. Banakara et al., “Evaluation of weather parameter-based pre-harvest yield forecast models for wheat crop,” Environmental Monitoring and Assessment, vol. 195, no. 1, 2023, doi: 10.1007/s10661-022-10552-4.
18. E. R. Abraham et al., “Time series prediction with artificial neural networks: Brazilian soybean production,” Agriculture, vol. 10, no. 10, p. 475, 2020, doi: 10.3390/agriculture10100475.
19. H. Bai, D. Xiao, J. Tang, and D. L. Liu, “Evaluation of wheat yield in North China Plain under extreme climate by coupling a crop model with ML,” Computers and Electronics in Agriculture, vol. 217, 2024, doi: 10.1016/j.compag.2024.108651.
20. D. B. Lobell et al., “Greater sensitivity to drought accompanies maize yield increase in the U.S. Midwest,” Science, vol. 344, no. 6183, pp. 516–519, 2014, doi: 10.1126/science.1251423.
21. S. K. Sek, “Impact of oil price changes on domestic price inflation: Evidence from ARDL models,” Energy, vol. 130, pp. 204–217, 2017, doi: 10.1016/j.energy.2017.03.152.
22. R. P. Roy and S. Sinha Roy, “Commodity futures prices pass-through and monetary policy in India,” The Journal of Economic Asymmetries, vol. 25, p. e00229, 2022, doi: 10.1016/j.jeca.2021.e00229.
23. S. E. Said and D. A. Dickey, “Testing for unit roots in autoregressive–moving average models of unknown order,” Biometrika, vol. 71, no. 3, pp. 599–607, 1984.
24. A. Hall, “Testing for a unit root in time series with pretest data-based model selection,” Journal of Business and Economic Statistics, vol. 12, no. 4, pp. 461–470, 1994, doi: 10.1080/07350015.1994.10524568.
25. N. A. M. R. Senaviratna and T. M. J. A. Cooray, “Diagnosing multicollinearity of logistic regression models,” Asian Journal of Probability and Statistics, pp. 1–9, 2019, doi: 10.9734/ajpas/2019/v5i230132.
26. J. Y. Le Chan et al., “Mitigating the multicollinearity problem: A machine learning approach,” Mathematics, vol. 10, no. 8, 2022, doi: 10.3390/math10081283.
27. K. Enwere, E. Nduka, and U. Ogoke, “Comparative analysis of ridge, bridge and lasso regression models in the presence of multicollinearity,” IPS Intelligentsia Multidisciplinary Journal, vol. 3, no. 1, pp. 1–8, 2023, doi: 10.54117/iimj.v3i1.5.
28. A. V. Tatachar, “Comparative assessment of regression models based on model evaluation metrics,” International Research Journal of Engineering and Technology, 2021. [Online]. Available: http://www.irjet.net.
29. J. B. Ramsey, “Tests for specification errors in classical linear least-squares regression analysis,” Journal of the Royal Statistical Society: Series B, vol. 31, no. 2, pp. 350–371, 1969, doi: 10.1111/j.2517-6161.1969.tb00796.x.
30. N. Ramana et al., “Comparative assessment of ARIMAX and NNARX models for forecasting Helicoverpa armigera,” International Journal of Tropical Insect Science, vol. 42, no. 5, pp. 3571–3580, 2022, doi: 10.1007/s42690-022-00883-7.
31. M. Pena et al., “Improved rainfall prediction through nonlinear autoregressive networks with exogenous variables,” Advances in Meteorology, vol. 2020, 2020, doi: 10.1155/2020/1828319.
32. P. Pandit et al., (duplicate of #6 but kept if used twice) “Hybrid time series models with exogenous variable…,” Scientific Reports, vol. 13, 2023.
33. R. Jaiswal, G. K. Jha, R. R. Kumar, and A. Lama, “Agricultural price forecasting using NARX model for soybean oil,” Current Science, vol. 125, no. 1, pp. xx–xx, 2023.
34. S. M. Guzman, J. O. Paz, and M. L. M. Tagert, “The use of NARX neural networks to forecast daily groundwater levels,” Water Resources Management, vol. 31, no. 5, pp. 1591–1603, 2017, doi: 10.1007/s11269-017-1598-5.
35. A. A. Neath and J. E. Cavanaugh, “The Bayesian information criterion: Background, derivation, and applications,” Wiley Interdisciplinary Reviews: Computational Statistics, vol. 4, no. 2, pp. 199–203, 2012, doi: 10.1002/wics.199.
36. A. M. Karki et al., “Electric load prediction using SARIMAX,” Australian Journal of Electrical and Electronics Engineering, 2024, doi: 10.1080/1448837X.2024.2405438.
37. A. H. Dhafer et al., “Empirical analysis for stock price prediction using NARX model with exogenous technical indicators,” Computational Intelligence and Neuroscience, vol. 2022, 2022, doi: 10.1155/2022/9208640.
38. H. Bar-Gera, “The target parameter of adjusted R-squared in fixed-design experiments,” The American Statistician, vol. 71, no. 2, pp. 112–119, 2017, doi: 10.1080/00031305.2016.1200489.
39. R. K. Paul and K. Sinha, “Forecasting crop yield: A comparative assessment of ARIMAX and NARX models,” RASHI: Journal of the Society for Application of Statistics in Agriculture and Allied Sciences, vol. 1, no. 1, pp. 77–85, 2016.
40. J. Barrientos et al., “Forecasting electricity price in Colombia: A comparison between neural network, ARMA, and hybrid models,” International Journal of Energy Economics and Policy, vol. 8, no. 3, pp. 97–106, 2018. [Online]. Available: http://hdl.handle.net/11159/2106.
41. F. X. Diebold and R. S. Mariano, “Comparing predictive accuracy,” Journal of Business and Economic Statistics, vol. 20, no. 1, pp. 134–144, 2002, doi: 10.1198/073500102753410444.
42. G. Ali, “Cointegration VAR and VECM and ARIMAX econometric approaches for water quality variates,” Journal of Statistical and Econometric Methods, vol. 4, no. 1, pp. 1–38, 2015.
43. F. Islam and M. Alam, “Use of teleconnections to predict Western Australian seasonal rainfall using ARIMAX model,” Hydrology, vol. 7, no. 3, 2020, doi: 10.3390/hydrology7030052.
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Environmental Research and Technology
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
All articles in Environmental Research and Technology (ERT) are published under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0).
This license allows others to copy, distribute, and adapt the work for non-commercial purposes only, provided that proper credit is given to the original authors and to the journal.
