Machine learning for sustainable investing: Current applications and overcoming obstacles in ESG analysis

Abstract

The intersection of Environmental, Social, and Governance (ESG) issues and Machine Learning (ML) has garnered significant attention in recent years as companies and investors increasingly recognize the paramount importance of sustainable and responsible business practices. ML techniques have been actively explored to tackle various ESG-related challenges, including enhancing ESG data quality and availability, developing comprehensive and dynamic ESG risk models, and optimizing ESG portfolios. The overall process of applying ML models in ESG analysis involves data collection, preprocessing, model training and evaluation, and model interpretation. Commonly used ML models in ESG analysis include logistic regression, decision trees, random forests, and support vector machines. However, there are notable obstacles to overcome, such as the lack of standardization and transparency in ESG data, as well as the potential for bias and ethical concerns in ML-based approaches. Further research and collaborative efforts among researchers and practitioners are crucial to fully realize the potential of ML in enhancing ESG analysis while ensuring transparency, ethical use, and alignment with sustainable and responsible investing principles.