Evaluating the Impact of Green Credit on Sustainable Innovation in Manufacturing Enterprises Using Random Forest

Against the backdrop of China's “dual carbon” strategy, green finance has emerged as a key policy tool to promote sustainable innovation in enterprises. As a core instrument of green finance, the actual impact of green credit on innovation performance in the manufacturing sector remains to be empirically explored. This study utilizes panel data from 657 A-share listed manufacturing firms in China from 2018 to 2023 to examine the influence of green credit on firms’ sustainable innovation capabilities. Methodologically, the study adopts a Random Forest Regression model, a widely used ensemble machine learning algorithm, to model and predict the relationship between green credit, environmental responsibility, financial indicators, corporate governance, and sustainable innovation. The model also evaluates the relative importance of each feature. The results show that social responsibility (SP), environmental performance (EP), and operational efficiency (ROA) are the most influential variables in explaining sustainable innovation performance. Although green credit (CG) is a focal variable, its direct explanatory power is relatively limited, implying a possible indirect effect through mediating variables. The model achieves high predictive accuracy (R² = 0.940, RMSE = 0.062), demonstrating the effectiveness of machine learning in capturing complex economic relationships. This research contributes to the empirical understanding of the link between green finance and corporate innovation and provides valuable data-driven insights for policy development and strategic enterprise decision-making.