Machine learning applications in climate finance: An overview

The application of machine learning is increasingly driving transformative advances in climate finance research. This paper presents a critical review of the literature published between 2015 and 2024, examining the use of machine learning techniques in key domains of climate finance, including carbon markets, climate financing, corporate carbon footprints, climate disclosures, climate risk assessment, and climate-related investments. The existing body of literature demonstrates the application of a broad range of machine learning techniques, such as time series forecasting, regression, classification, natural language processing (NLP), unsupervised learning, causal inference, reinforcement learning, federated learning, and explainable machine learning, to address core challenges in climate finance. This review evaluates both the achievements and limitations of existing studies, identifies key research gaps, and outlines promising directions for future research.