题目：An Autocovariance-based Framework for Vast Curve Time Series

主讲人：英国伦敦政治经济学院统计系 乔兴昊助理教授

主持人：西南财经大学统计学院 常晋源教授

时间：2019年4月24日（星期三）15:00-16:00

地点：西南财经大学光华校区光华楼1007会议室

报告摘要：

It is commonly assumed in functional data analysis (FDA) that samples of each functional variable are independent realizations of an underlying stochastic process, and are observed over a grid of points contaminated by i.i.d. measurement errors. In practice, however, the temporal dependence across curve observations may exist and the diagonal assumption on the error covariance structure could be unrealistic. We consider the model setting for serially dependent curve observations, when the contamination by errors is genuinely functional with a fully nonparametric covariance structure. The classical covariance-based methods in FDA are not applicable here due to the contamination that can result in substantial estimation bias. We propose an autocovariance-based framework to address large-scale error-contaminated curve time series problems. Under the proposed framework, we discuss several important problems including dimension reduction, regularized generalized method-of-moments estimators for functional additive linear regression and vector functional autoregression under "large p, small n" scenarios.

主讲人简介：

Dr. Xinghao Qiao is an assistant professor of Statistics at London School of Economics. His research areas include functional data analysis, high dimensional statistics, time series analysis and Bayesian nonparametrics. Prior to joining LSE, Dr. Qiao earned his PhD in Business Statistics from Marshall School of Business at the University of Southern California.