High-Dimensional Low-Rank Linear Time Series Modeling

Prof. Guodong LI
University of Hong Kong

Date: 19 October 2023 (Thursday)
Time: 3:00 pm – 4:30 pm
Venue: E22-G015
Host: Prof. Yan LIN, Assistant Professor in Business Intelligence and Analytics
Online registration: https://umac.au1.qualtrics.com/jfe/form/SV_8vlB2IEUrgUyFz8


Motivated by Tucker tensor decomposition, this paper imposes low-rank structures to the column and row spaces of coefficient matrices in a multivariate infinite-order vector autoregression (VAR), which leads to a newly proposed concept of supervised factor models, where two factor modelings are conducted to responses and predictors simultaneously. Interestingly, the stationarity condition implies an intrinsic weak group sparsity mechanism of infinite-order VAR, and hence a rank-constrained group Lasso estimation is considered to make inference on high-dimensional time series. Its non-asymptotic properties are also discussed thoughtfully by balancing the estimation, approximation and truncation errors. Moreover, an alternating gradient descent algorithm with thresholding is designed to search for the high-dimensional estimate, and its theoretical justifications, including statistical and convergence analysis, are also provided. Theoretical and computational properties of the proposed methodology are verified by simulation experiments, and the advantages over existing methods are demonstrated by two empirical examples. This is a joint work with my two PhD students, Feiqing Huang and Kexin Lu.


Prof. Guodong joined the Department of Statistics & Actuarial Science, University of Hong Kong, in 2009 as an Assistant Professor, and is currently a Professor. He received his PhD in Statistics in The University of Hong Kong. In 2022, he published three papers in Journal of Business and Economic Statistics, Journal of Econometrics, Econometric Theory (ABS4) respectively.

All are welcome!