A0493
Title: Tensor methods for high-dimensional time series modeling
Authors: Yao Zheng - University of Connecticut (United States) [presenting]
Abstract: Tensor decomposition is a powerful dimensionality reduction tool that has gained much interest in modern machine learning applications. However, its development in the areas of time series analysis and econometrics is still in its infancy. We will present some recent work on high-dimensional time series modeling via tensor methods. Specifically, we will discuss the use of Tucker decomposition in high-dimensional vector autoregressive modeling, tensor-valued autoregressive time series modeling, and high-dimensional mixed-data sampling (MIDAS) regression. The focus will be more on motivations, model formulations, interpretations and empirical examples. Estimation methods and theoretical properties will be mentioned briefly.
