Oil Demand Forecasting in Importing and Exporting Countries: AI-Based Analysis of Endogenous and Exogenous Factors
Hui Zhu
DOI: https://doi.org/10.3390/su151813592
IF: 3.9
2023-09-13
Sustainability
Abstract:Given the prevalence of the digital world, artificial intelligence (AI) stands out as one of the most prominent technologies for demand prediction. Although numerous studies have explored energy demand forecasting using machine learning models, previous research has been limited to incorporating either a country's macroeconomic characteristics or exogenous elements as input variables. The simultaneous consideration of both endogenous and exogenous economic elements in demand forecasting has been disregarded. Furthermore, the stability of machine learning models for energy exporters and importers facing varying uncertainties has not been adequately examined. Therefore, this study aims to address these gaps by investigating these issues comprehensively. To accomplish this objective, data from 30 countries spanning the period from 2000 to 2020 was selected. In predicting oil demand, endogenous economic variables, such as carbon emissions, income level, energy price, gross domestic product (GDP), population growth, urbanization, trade liberalization, inflation, foreign direct investment (FDI), and financial development, were considered alongside exogenous factors, including energy sanctions and the COVID-19 pandemic. The findings indicate that among the input variables examined in demand forecasting, oil sanctions and the COVID-19 pandemic have had the most significant impact on reducing oil demand, while trade liberalization has proven to be the most influential factor in increasing oil demand. Furthermore, the support vector regression (SVR) model outperforms other models in terms of lower prediction error, as revealed by the error assessment of statistical models and AI in forecasting oil demand. Additionally, when comparing the stability of models in oil exporting and importing countries facing different levels of demand uncertainty, the SVR model demonstrates higher stability compared to other models.
environmental sciences,environmental studies,green & sustainable science & technology