Modeling large dimensional matrix time series with partially known and latent factors

TL;DR

This paper extends matrix factor models for large-dimensional matrix time series by incorporating known factors and covariates, providing theoretical convergence results and demonstrating improved performance on stock data.

Contribution

It introduces a regression-enhanced matrix factor model with proven convergence rates, advancing modeling of high-dimensional matrix time series with known and latent factors.

Findings

Theoretical convergence rates match existing literature.

Numerical studies confirm estimation accuracy in finite samples.

Model outperforms traditional approaches on stock return data.

Abstract

This article considers to model large-dimensional matrix time series by introducing a regression term to the matrix factor model. This is an extension of classic matrix factor model to incorporate the information of known factors or useful covariates. We establish the convergence rates of coefficient matrix, loading matrices and the signal part. The theoretical results coincide with the rates in Wang et al. (2019). We conduct numerical studies to verify the performance of our estimation procedure in finite samples. Finally, we demonstrate the superiority of our proposed model using the daily returns of stocks data.