Published May 2026, Pg. 45-49

Section: Oil and gas preparing and transportation

UOT: 622.27:622.69

DOI: 10.37474/0365-8554/2026-05-45-49

Forecasting natural gas demand using support vector machine (SVM) method

S.E. Imanov - “Oil-Gas Scientific Research Project” Institute

Accurate forecasting of natural gas demand is crucial for energy security, infrastructure optimization, and efficient policy planning. This study investigates the application of the Support Vector Regression (SVR) method for predicting monthly natural gas consumption in Azerbaijan. Data obtained from industrial metering devices covered the period of 2019–2022 for training and 2023 for testing. 
By employing logarithmic transformation, feature engineering, and STL decomposition in the modeling process, precise capture of both seasonal cyclicity and non-linear variations was ensured. Evaluation results confirmed high accuracy. The forecast indicators showed close alignment with actual values, successfully replicating winter peaks and summer minimums. 
The research results demonstrate that the SVR method is a reliable approach for short-term natural gas demand forecasting and offers practical opportunities for energy system management.

References:

1. Mohandes M. Support vector machines for short-term electrical load forecasting // International Journal of Energy Research, 2002, vol. 26(4), pp. 335-345.

2. Ceperic E., Ceperic V., Baric A. A strategy for short-term load forecasting by Support Vector Regression machi-
nes // IEEE Transactions on Power Systems, 2013, vol. 28(4), pp. 4356-4364.

3. Ma X., Deng Y., Yuan H. Forecasting the natural gas supply and consumption in China using a novel grey wavelet support vector regressor // Systems, 2023, vol. 11(8), p. 428.

4. Gao F., Shao X. Forecasting annual natural gas consumption via the application of a novel hybrid model // Environmental Science and Pollution Research, 2021, vol. 28, pp. 21411-21424.

5. Rao C., Zhang Y., Wen J., Xiao X., Goh M. Energy demand forecasting in China: A support vector regression–compositional data second exponential smoothing model // Energy, 2023, vol. 263, Art. 125955.

6. Bai Y. Daily natural gas consumption forecasting based on a structure-calibrated support vector regression
(SC-SVR) approach // Energy Conversion and Management, 2016, vol. 122, pp. 49-59.

7. Dung N.T., Phuong T. Short-term electric load forecasting using standardized load profile (SLP) and SVR // Engineering, Technology & Applied Science Research, 2019, vol. 9(2), pp. 3925-3929.

8. Hong W.C., Dong Y., Lai C.Y., Chen L.Y. Electric load forecasting by support vector regression with immune algorithm // Energy, 2009, vol. 34(1), pp. 78-90.