Industrial Exports and Global Carbon Emissions: Assessing the Impact of the U.S. and China Using a Decision Tree Algorithm

This research employs a decision tree approach to investigate the impact of industrial exports from China and the United States on global carbon emissions in the industrial sector. Drawing on panel data from 1992 to 2023 provided by the World Bank, the results reveal a substantial positive relationship between China's manufacturing exports- accounting for 81% of the feature importance- and global industrial CO2 emissions (Pearson coefficient: 0.837). The feature importance of U.S. exports, on the other hand, is lower at 19%, suggesting a negative association (Pearson coefficient: ?0.901). With an R-squared value of 0.96 and a mean squared error (MSE) of 0.002, the decision tree model outperformed the other two models, random forests and support vector machines, among the five machine learning models evaluated. These findings demonstrate the disparate environmental effects of export-driven industrialization, with a focus on China's significant emissions from energy-intensive manufacturing, and the urgent need for focused measures, like carbon pricing, the incorporation of green technologies, and sustainable trade agreements, to bring industrial growth into line with climate goals, especially the Sustainable Development Goals (SDGs 9 and 13) of the United Nations. This study advances the application of machine learning in environmental economics, providing valuable insights for balancing economic and ecological priorities in international trade systems.