Publication Type

Conference Paper

Version

submittedVersion

Publication Date

12-2024

Abstract

Environmental, Social, and Governance (ESG) factors are increasingly essential in evaluating corporate performance, driving demand for accurate ESG risk assessments. However, smaller companies often face challenges in obtaining validated ESG scores due to resource constraints. This study explores the use of Machine Learning (ML) to predict ESG risk scores for U.S. companies, leveraging data from Wharton Research Data Services (WRDS), Yahoo Finance, and Sustainalytics. The XGBoost model demonstrated the best performance, significantly improving the accuracy of ESG risk predictions. These findings suggest that ML can enhance ESG risk assessments, offering valuable insights for investors, regulatory bodies, and corporate management.

Keywords

ESG, machine learning, sustainability

Discipline

Artificial Intelligence and Robotics

Research Areas

Data Science and Engineering

Publication

Annual International Open Innovation Conference 2024

Publisher

Taylor and Francis Group

City or Country

Vietnam

Share

COinS