Advanced machine learning approaches for predicting energy and fossil fuel consumption for green growth

A large increase in energy usage has occurred due to rapid industrial growth and urbanization, which has resulted in environmental degradation and heavy reliance on the consumption of fossil fuels. The purpose of this research study is to address the pressing problem of accurately projecting energy usage in China by deploying ML algorithms. For sustainable development and long-term industry competitiveness, green growth and innovation must be introduced. Using a database comprising a set of different input parameters, the study optimizes the ML schemes. Not only are four hybrid schemes (CatBoost-ARO, LightGBM-ARO, XGBR-ARO, and HGBR-ARO) evaluated using statistical evaluation indices, but they are also compared in terms of their projection of overall energy usage, fossil fuel consumption, and electricity consumption. The results show that the hybrid schemes CatBoost-ARO, XGBR-ARO, and LightGBM-ARO provide a much better approximation of the energy usage parameter than the others, based on a set of objects that are first converted into points. Moreover, CatBoost-ARO demonstrated superior overall energy usage and electricity consumption projection accuracy, whereas XGBR-ARO provided the best performance in fossil fuel consumption prediction. The expected and actual values had a strong correlation for all four hybrid schemes, which successfully captured the underlying patterns and variations in the database. • Rapid industrialization and urbanization in China have caused increased energy use and environmental degradation. • The study aims to accurately predict energy usage using machine learning (ML) algorithms. • Emphasis is placed on green growth, sustainable development, and long-term industrial competitiveness. • Four hybrid ML models were tested: CatBoost-ARO, LightGBM-ARO, XGBR-ARO, and HGBR-ARO. • Models were evaluated based on their projections of overall energy use, fossil fuel consumption, and electricity consumption. • CatBoost-ARO showed superior performance in overall energy and electricity consumption forecasting.