摘要
With the confluence of rapidly continuing economic development and increasing greenhouse gas emissions, the frequency and severity of natural disasters caused by global warming have increased in recent years. The international Financial Stability Board established the Task Force on Climate-Related Financial Disclosures (TCFD) to provide recommendations for collecting and disclosing climate-related information that is helpful for decision-making such that investors, decision-makers and supervision units can understand and more accurately assess climate-related risks as well as opportunities. However, the TCFD recommendations only provide a general framework of risks and opportunities from which companies cannot derive specific disaster-related data for regional climate disclosures. For stakeholders in different continents and regions, there is a need to develop a disaster data-driven model that can precisely disclose regional risks. Our study develops a regional climate framework for TCFD based on natural language processing that can process and analyze data from historical local news. The paper aims to offer suitable climate recommendations for different regions and suggest possible preventive measures and responses to emergencies caused by various disasters. The contributions of this research contain: (1) using unsupervised learning to present climate change in a quantitative way; (2) helping enterprises adjust sustainable strategies according to conditions in different regions; (3) extending involvement from a small area to a large area, quantifying climate data so that investors can make decisions based on factual results; (4) combining artificial intelligence with climate change in the field of TCFD.