Predicting ESG Ratings by Machine Learning and Analyzing Influencing Factors by XAI

This study examines the role of Environmental, Social, and Governance (ESG) management in corporate strategy, particularly focusing on predicting ESG ratings with machine learning. Given the diverse ESG evaluation criteria employed by global rating agencies, there's a need for clear guidelines to facilitate effective ESG management. The research aims to develop an ESG rating prediction model utilizing a triennial compendium of Korean corporate financial data. This process involves a comparative analysis of linear models, tree-based models, and neural network-based models. Additionally, this study explains the importance of various variables by applying SHAP, one of the XAI techniques. The results indicate that XGB is the most effective, achieving an 85.1% F1 score in ESG rating predictions. By understanding how financial factors impact ESG ratings, companies can develop more effective ESG strategies, forming an essential foundation for sustainable growth.