Developing a Financial Ratio-Based Model for ESG Sustainability Prediction in Private Firms: A Machine Learning Approach

This thesis explores the use of financial ratios to predict the Environmental, Social, and Governance (ESG) performance of private firms through a machine learning approach. Given the limited ESG data for private companies, a Random Forest classifier is employed, trained on public companies with established ESG ratings, to assess private firms' sustainability status. The study highlights the potential and limitations of using financial indicators as proxies for ESG performance, offering an alternative method for evaluating corporate sustainability in the absence of formal disclosures.