Modeling complex time series with tensor techniques

李国栋

<p>The tensor, as an extension of matrices, has recently gained significant attention from econometricians and statisticians, and the literature has witnessed two types of its applications. Firstly, the involvement of tensors can facilitate the development of new inference tools, which are impossible under the framework of matrices alone. Secondly, many big data come in the form of tensors. This talk will introduce several of my research studies on high-dimensional time series modeling using tensors, and I will also present two inference tools for tensor-valued time series.</p>

2023-06-13 (Tuesday) 16:30-18:00

中科院数学与系统科学研究院南楼N204、厦大经济楼N302(线下分会场), 腾讯会议：92921396317

中文

中国科学院大学经济与管理学院、中国科学院预测科学研究中心、厦门大学邹至庄经济研究院、NSFC“计量建模与经济政策研究”基础科学中心

系列讲座

洪永淼

<p>李国栋，现任香港大学统计精算系教授。本科和硕士毕业于北京大学数学学院，2007年于香港大学统计精算系获得统计学博士，随后在南洋理工大学任助理教授。主要研究方向包括计量经济、时间序列分析、分位数回归、高维统计数据分析和机器学习。已发表学术论文60余篇，其中20余篇发表在Journal of Econometrics、 Econometric Theory和Journal of Business and Economic Statistics等计量经济学顶级期刊，以及统计学4大顶级期刊和机器学习的3大顶级会议上。</p>

“邹至庄讲座”杰出学者论坛（第23期）