Construction and analysis of China’s carbon emission model based on machine learning

Abstract

To cope with the severe challenge of China’s huge carbon emissions, this study introduced a comprehensive research paradigm, “modelling + SHAP analysis + scenario prediction,” from the machine learning perspective. Firstly, by calculating the Spearman correlation coefficient, it is found that China’s carbon emissions are significantly correlated with nine explanatory variables, such as the proportion of coal in total energy consumption and urbanization rate. Then, the contribution value of each explanatory variable in the optimal model is quantified by the SHAP method, and it is revealed that energy intensity and urbanization rate are the key factors affecting China’s carbon emissions and have negative and positive effects, respectively. Finally, through the policy scenario simulation, it is found that China’s carbon emissions will level off from 2022 to 2028 and peak in 2028, which is expected to be about 9.72 billion tons of carbon emissions by 2030.