Natural gas consumption forecasting using machine learning methods: MLP, decision tree, and gradient boosting
DOI:
https://doi.org/10.33216/1998-7927-2026-304-6-88-101Keywords:
natural gas consumption, forecasting, machine learning methods, MLP, decision tree, gradient boosting, time seriesAbstract
The paper investigated the accuracy of natural gas consumption forecasting using machine learning methods: multilayer perceptron (MLP), decision tree, and gradient boosting. The forecast was formed for consecutive three-hour intervals, and the accuracy of the selected models was evaluated over the full annual test interval of 2019. The main focus was not only on the choice of algorithm, but also on the method of presenting information about the previous dynamics of gas consumption. The same set of 47 input features was used for all models. It included 16 meteorological, six cyclic time, 11 lag, nine moving, and five difference features. This made it possible to compare the forecasting capabilities of algorithms with the same composition of input features. Data from previous years were used to build and tune the models, and 2019 was left for independent verification. The results showed that the decision tree forms the highest forecasting accuracy over the annual test interval. Gradient boosting and MLP also reproduce the general dynamics of gas consumption quite accurately, but are more likely to make mistakes in areas with rapid regime changes. The largest discrepancies between the models are observed in the fall, when gas consumption goes from a low summer level to a level typical of the heating season. Analysis of the importance of input features showed that the forecast is determined to the greatest extent by the previous values of gas consumption, while meteorological and calendar characteristics make it possible to take into account the influence of weather conditions, seasonality and time patterns of the process. The results obtained confirm that the forecast accuracy depends not only on the choice of model, but also on how fully the time structure of the process is reflected in the input data. Statistical testing showed that the decision tree has a statistically confirmed advantage over MLP and gradient boosting, i.e. the detected differences in forecast accuracy cannot be explained only by the random variability of errors in the test sample. The practical significance of the work lies in the possibility of using the obtained results to build systems for operational forecasting of gas consumption in conditions of seasonal variability of demand and weather changes.
References
1. Liu J., Wang S., Wei N., Chen X., Xie H., Wang J. Natural gas consumption forecasting: A discussion on forecasting history and future challenges // Journal of Natural Gas Science and Engineering. 2021. Vol. 90. Article 103930. DOI: https://doi.org/10.1016/j.jngse.2021.103930
2. Šebalj D., Mesarić J., Dujak D. Analysis of methods and techniques for prediction of natural gas consumption: A literature review // Journal of Information and Organizational Sciences. 2019. Vol. 43, No. 1. P. 99–117. DOI: https://doi.org/10.31341/jios.43.1.6
3. Szoplik J. Forecasting of natural gas consumption with artificial neural networks // Energy. 2015. Vol. 85. P. 208–220. DOI: https://doi.org/10.1016/j.energy.2015.03.084
4. Merkel G. D., Povinelli R. J., Brown R. H. Short-Term Load Forecasting of Natural Gas with Deep Neural Network Regression // Energies. 2018. Vol. 11, No. 8. Article 2008. DOI: https://doi.org/10.3390/en11082008
5. Laib O., Khadir M. T., Mihaylova L. Toward efficient energy systems based on natural gas consumption prediction with LSTM Recurrent Neural Networks // Energy. 2019. Vol. 177. P. 530–542. DOI: https://doi.org/10.1016/j.energy.2019.04.075
6. Sharma V., Cali Ü., Sardana B., Kuzlu M., Banga D., Pipattanasomporn M. Data-driven short-term natural gas demand forecasting with machine learning techniques // Journal of Petroleum Science and Engineering. 2021. Vol. 206. Article 108979. DOI: https://doi.org/10.1016/j.petrol.2021.108979
7. Tian N., Shao B., Bian G., Zeng H., Li X., Zhao W. Application of forecasting strategies and techniques to natural gas consumption: A comprehensive review and comparative study // Engineering Applications of Artificial Intelligence. 2024. Vol. 129. Article 107644. DOI: https://doi.org/10.1016/j.engappai.2023.107644
8. Zhao M., Guo G., Fan L., Han L., Yu Q., Wang Z. Short-term natural gas load forecasting based on EL-VMD-Transformer-ResLSTM // Scientific Reports. 2024. Vol. 14. Article 20343. DOI: https://doi.org/10.1038/s41598-024-70384-9
9. Lin Z., Xie L., Zhang S. A compound framework for short-term gas load forecasting combining time-enhanced perception transformer and two-stage feature extraction // Energy. 2024. Vol. 298. Article 131365. DOI: https://doi.org/10.1016/j.energy.2024.131365
10. Wang H., Gao X., Zhang Y., Yang Y. Natural Gas Consumption Forecasting Model Based on Feature Optimization and Incremental Long Short-Term Memory // Sensors. 2025. Vol. 25, No. 10. Article 3079. DOI: https://doi.org/10.3390/s25103079
11. Xu X., Xu H., Mei K., Tong L., Liu Z., Wang T., Fang K. Integrating multi-modal data into transformer model for short-term gas consumption forecasting // Alexandria Engineering Journal. 2025. Vol. 122. P. 655–664. DOI: https://doi.org/10.1016/j.aej.2025.03.029
12. Niu P., Du J., Xu N., Wang B., Liao Q., Qiu R., Cai S., Liang Y. P-KTFNet: A prior knowledge enhanced time-frequency forecasting model for natural gas consumption // Energy. 2025. Vol. 328. Article 136559. DOI: https://doi.org/10.1016/j.energy.2025.136559
13. Qiao W., Ma Q., Yang Y., Xi H., Huang N., Yang X. et al. Natural gas consumption forecasting using a novel two-stage model based on improved sparrow search algorithm // Journal of Pipeline Science and Engineering. 2025. Vol. 5, No. 1. Article 100220. DOI: https://doi.org/10.1016/j.jpse.2024.100220
14. Lytvynenko I., Lupenko S., Nazarevych O., Shymchuk H., Hotovych V. Additive mathematical model of gas consumption process // Scientific Journal of TNTU. 2021. Vol. 104, No. 4. P. 87–97. DOI: https://doi.org/10.33108/visnyk_tntu2021.04.087
15. Lytvynenko I., Lupenko S., Kunanets N., Nazarevych O., Shymchuk G., Hotovych V. Simulation of gas consumption process based on the mathematical model in the form of cyclic random process considering the scale factors // CEUR Workshop Proceedings. 2021. Vol. 3039. P. 97–106. URL: https://ceur-ws.org/Vol-3039/paper23.pdf
16. Shymchuk G., Lytvynenko I., Hromyak R., Lytvynenko S., Hotovych V. Gas Consumption Forecasting Using Machine Learning Methods and Taking Into Account Climatic Indicators // CEUR Workshop Proceedings. 2023. Vol. 3468. P. 156–163. URL: https://ceur-ws.org/Vol-3468/short8.pdf
17. Bergmeir C., Benítez J. M. On the use of cross-validation for time series predictor evaluation // Information Sciences. 2012. Vol. 191. P. 192–213. DOI: https://doi.org/10.1016/j.ins.2011.12.028
18. Tashman L. J. Out-of-sample tests of forecasting accuracy: an analysis and review // International Journal of Forecasting. 2000. Vol. 16, No. 4. P. 437–450. DOI: https://doi.org/10.1016/S0169-2070(00)00065-0
19. Nair V., Hinton G. E. Rectified Linear Units Improve Restricted Boltzmann Machines // Proceedings of the 27th International Conference on Machine Learning. 2010. P. 807–814. URL: https://proceedings.mlr.press/v9/nair10a.html
20. Kingma D. P., Ba J. Adam: A Method for Stochastic Optimization // Proceedings of the 3rd International Conference on Learning Representations (ICLR 2015). 2015. URL: https://arxiv.org/abs/1412.6980
21. Breiman L., Friedman J. H., Olshen R. A., Stone C. J. Classification and Regression Trees. Belmont, CA : Wadsworth, 1984.
22. Pedregosa F., Varoquaux G., Gramfort A. et al. Scikit-learn: Machine Learning in Python // Journal of Machine Learning Research. 2011. Vol. 12. P. 2825–2830. URL: https://jmlr.org/papers/v12/pedregosa11a.html
23. Friedman J. H. Greedy Function Approximation: A Gradient Boosting Machine // The Annals of Statistics. 2001. Vol. 29, No. 5. P. 1189–1232. DOI: https://doi.org/10.1214/aos/1013203451
24. Hyndman R. J., Koehler A. B. Another look at measures of forecast accuracy // International Journal of Forecasting. 2006. Vol. 22, No. 4. P. 679–688. DOI: https://doi.org/10.1016/j.ijforecast.2006.03.001
25. Diebold F. X., Mariano R. S. Comparing predictive accuracy // Journal of Business & Economic Statistics. 1995. Vol. 13, No. 3. P. 253–263. DOI: https://doi.org/10.1080/07350015.1995.10524599
26. Altmann A., Toloşi L., Sander O., Lengauer T. Permutation importance: a corrected feature importance measure // Bioinformatics. 2010. Vol. 26, No. 10. P. 1340–1347. DOI: https://doi.org/10.1093/bioinformatics/btq134
