Advanced Neural Network Models in Forecasting Carbon Emissions

Abstract

The detrimental trend of the surging carbon emissions (CEs) and the subsequent issues of environmental problems have attracted global attention. Despite control measures that have been taken by governments worldwide to achieve carbon neutrality, further policymaking remains challenging. After browsing various research, forecasting CEs in advance is considered a proper measure to mitigate the trend by making the carbon capture data-oriented. Judging from influencing factors of concentrations, it is seen that a great number of researchers have focused on machine learning methods, especially for neural network models. Then neural network models implemented in three perspectives, including worldwide, local, and transportation, are analyzed in six metrics: Coefficient of determination (R2), root mean square error (RMSE), mean square error (MSE), mean absolute error (MAE), mean absolute percentage error (MAPE), Relative Error (RE). The results following these metrics suggest that the future of utilizing neural network models in forecasting CEs is promising in guiding future carbon neutrality works.