Natural Gas Spot Price Prediction Using a Machine Learning Datacentric Approach

Abstract

Abstract The ability to accurately predict natural gas prices asides being beneficial to stakeholders of the natural gas market also have positive economic impacts on energy management and environmental sustainability. This paper explores the application of machine learning algorithms for the purpose of accurately predicting monthly natural gas spot prices. Henry Hub natural gas spot price data from January 2001 to November 2021 were utilized alongside four machine learning algorithms namely; Artificial Neural Networks (ANN), Support Vector Regression (SVR), Random Forest Regressor and Gradient Boosting Machine (GBM). The models were trained with 11 variables with 80% of the dataset utilized for training and 20% for testing purposes. A 10-fold cross validation technique was implemented for model validation purposes. The accuracy of each model was evaluated using the Root Mean Square error metric. After model evaluation, all four models generated distinct results, with the Artificial Neural Network model having the most accurate prediction of all four models.