Research on Enterprise Carbon Credit Rating Methods under the Dual-Carbon Goal: Multi-Source Data Fusion and Quantitative Modeling

Against the backdrop of China’s “dual carbon” strategy targeting carbon peaking and carbon neutrality, the scientific evaluation of enterprises’ carbon credit rating has become a critical element in building a green financial system and advancing low-carbon corporate transformation. Particularly in the power sector, which is characterized by high carbon intensity, the development of an effective carbon credit rating model can help regulatory bodies accurately identify carbon-risk enterprises and provide a solid basis for financial institutions’ green credit and investment decisions. This study constructs a multi-source dataset integrating structured data (e.g., carbon emissions, energy consumption, financial indicators) and unstructured data (e.g., policy responses, Environmental, Social, and Governance (ESG) performance, and public sentiment). It applies a combined Analytic Hierarchy Process (AHP)-entropy weighting method to determine indicator weights and builds a classification model using logistic regression and a BP neural network. An empirical analysis of listed power enterprises on China’s A-share market shows that the proposed approach outperforms traditional singlemethod models in terms of accuracy and robustness, offering strong applicability and practical significance.