报告时间：2024.05.24，15:30——17:30
报告地点：11-302会议室
报告人： 张涤天博士（东南大学）
题目：Exploring the drivers of global energy risk spillover: an interpretable machine learning approach
主要内容：A comprehensive understanding of global energy risk spillovers and their underlying factors is crucial for ensuring national energy security and financial stability. This research investigates risk spillovers across energy markets in 28 countries, revealing that each country's role within the global energy risk spillover system fluctuates over time. Utilizing innovative interpretable machine learning models, specifically the Explainable Boosting Machine (EBM) and the Tree SHAP method, we analyze the macroeconomic factors determining whether a country acts as a transmitter or receiver of energy risk. The EBM and Tree SHAP models excel in revealing the dynamics between these factors and energy risk spillovers, including their nonlinear interactions, surpassing traditional linear regression models which struggle with variable collinearity and nonlinear relationships. Key drivers identified, such as GDP, manufacturing value added as a percentage of GDP, and carbon intensity, significantly influence a country's position in the energy risk spillover framework. This investigation extends the use of machine learning in risk analysis and offers valuable insights for policymakers worldwide.
邀请人：陈雪平
欢迎全校感兴趣师生参加