题目：Navigating the Factor Streams

主讲人：北京大学光华管理学院 涂云东教授

主持人：统计学院 常晋源教授

时间：2025年01月08日（周三）上午10:00-11:00

地点：西南财经大学光华校区光华楼1003会议室

报告摘要：

This article introduces a new procedure to identify the group membership of high-dimensional panel data with a latent multilevel factor structure. The time series are grouped using an agglomerative hierarchical clustering algorithm, based on a correlation measure among the common components obtained by principal component analysis. A new information criterion is further proposed to determine the number of latent groups. Under some regularity conditions, the consistency of the membership estimators as well as that of the group number estimator are established. Monte Carlo simulations demonstrate the improved finite sample performance of our proposed method compared to existing alternatives in terms of the clustering accuracy. The application to an international macroeconomic dataset reveals that our identified groups differ in important ways from those identified by geography alone, and further illustrates the advantages of our clustering approach in the out-of-sample forecasting.

主讲人简介：

涂云东，北京大学光华管理学院教授。入选“日出东方”北大光华青年人才，北京大学优秀博士学位论文指导教师，北京大学优秀研究生导师，国家级人才项目入选者，国家杰出青年科学基金取得者。2004年和2006年先后获武汉大学理学学士学位和经济学硕士学位，2012年获美国加州大学河滨分校经济学博士学位。亚太青年计量经济学者会议发起人和主要组织者。40余篇学术论文发表在多个国际国内知名专业杂志。著作教材《时间序列分析》由人民邮电出版社于2022年9月出版。研究领域涵盖时间序列分析、非参数计量方法、大数据分析、金融计量和预测等。