Helping Effects Against the Curse of Dimensionality in Threshold Factor Models for High-Dimensional Matrix Time Series

• Activity Number: 241 - Estimation Challenges and New Approaches
• Type: Contributed
• Date/Time: Monday, July 29, 2019 : 2:00 PM to 3:50 PM
• Sponsor: Business and Economic Statistics Section
• Author(s): Xialu Liu* and YI CHEN
• Companies: San Diego State University and Princeton University
• Keywords: Factor model; Thresholding effect; High-dimensional time series; Matrix-valued time series

Abstract

In this paper, we introduce a threshold factor models to analyze matrix-valued high- dimensional time series data. The estimation methods for loading spaces, threshold value, and the number of factors are proposed. The impacts of thresholding and interaction between regimes on the estimators when dimensions and sample size go to infinity are investigated. We find that comparing with one-regime vector factor model, estimators for loadings and threshold value in our model experience ’helping’ effects against the curse of dimensionality, when two states have different level of strength and two directions of dimension reduction have different levels of thresholding. The simulated and real data are analyzed.