Natural gas spot price prediction research under the background of Russia-Ukraine conflict - based on FS-GA-SVR hybrid model

Abstract

The ongoing Russia-Ukraine conflict has led to significant upheaval in the worldwide natural gas sector. Accurate natural gas price forecasting, as an essential tool for mitigating market uncertainty, plays a crucial role in commodity trading and regulatory decision-making. This study aims to develop a hybrid forecasting model, the FS-GA-SVR model, which integrates feature selection (FS), genetic algorithm (GA), and support vector regression (SVR) to investigate Henry Hub natural gas price prediction amidst the Russia-Ukraine conflict. The results show that: (1) The feature selection automates model input variable selection, decreasing the time required while improving the model's accuracy. (2) The use of genetic algorithm for selecting support vector regression hyperparameters significantly improves the accuracy of natural gas price predictions. The algorithm leads to a decrease of approximately 70% in measurement indicators. (3) During the Russia-Ukraine conflict, the FS-GA-SVR hybrid model demonstrates more consistent and accurate predictions for natural gas spot prices than the base SVR model. This study serves as a valuable theoretical reference for energy policymakers and natural gas market investors worldwide, supporting their ability to anticipate fluctuations in natural gas prices.

Keywords: Feature selection; Genetic algorithm; Natural gas prices; Russia-Ukraine conflict; Support vector regression.