主题：Causal Machine Learning and Early Warning Indicator for Crisis Forecasting
主讲人：杜在超 复旦大学
时间：2025年6月27日（周五）上午10:00
地点：暨南大学石牌校区经济学院大楼（中惠楼）202室
摘要
This paper proposes a causal machine learning method and an early warning indicator (EWI) for crisis forecasting. We first construct a crisis-related multi-source dataset for 17 OECD countries. We then apply not only commonly used machine learning methods, but also some complex models whose number of parameters exceeds the sample size, as well as the Nyström method to forecasting crisis. A double machine learning method enables us to identify causal factors for crises, and an EWI is further constructed. Empirical results show that our EWI produces more accurate and timely crisis predictions.
主讲人简介
杜在超，复旦大学经济学院教授，博士生导师，《世界经济文汇》编辑，入选国家级青年人才计划。2010年获得美国印第安纳大学经济学博士。论文发表于Management Science，Journal of Econometrics，China Quarterly和《经济研究》等国内外著名杂志。主持国家自然科学基金2项，教育部人文社科项目1项，上海市教育委员会科研创新计划重大项目1项。
欢迎感兴趣的师生参加！
校对| 区昕然
责编| 彭毅
初审| 陈创练
终审发布| 何凌云
（来源：暨南大学经济学院微信公众号）
