English

Bivariate Matrix-valued Linear Regression (BMLR): Finite-sample performance under Identifiability and Sparsity Assumptions

Statistics Theory 2025-12-08 v3 Machine Learning Statistics Theory

Abstract

This study explores the estimation of parameters in a matrix-valued linear regression model, where the TT responses (Yt)t=1TRn×p(Y_t)_{t=1}^T \in \mathbb{R}^{n \times p} and predictors (Xt)t=1TRm×q(X_t)_{t=1}^T \in \mathbb{R}^{m \times q} satisfy the relationship Yt=AXtB+EtY_t = A^* X_t B^* + E_t for all t=1,,Tt = 1, \ldots, T. In this model, AR+n×mA^* \in \mathbb{R}_+^{n \times m} has L1L_1-normalized rows, BRq×pB^* \in \mathbb{R}^{q \times p}, and (Et)t=1T(E_t)_{t=1}^T are independent noise matrices following a matrix Gaussian distribution. The primary objective is to estimate the unknown parameters AA^* and BB^* efficiently. We propose explicit optimization-free estimators and establish non-asymptotic convergence rates to quantify their performance. Additionally, we extend our analysis to scenarios where AA^* and BB^* exhibit sparse structures. To support our theoretical findings, we conduct numerical simulations that confirm the behavior of the estimators, particularly with respect to the impact of the dimensions n,m,p,qn, m, p, q, and the sample size TT on finite-sample performances. We complete the simulations by investigating the denoising performances of our estimators on noisy real-world images.

Keywords

Cite

@article{arxiv.2412.17749,
  title  = {Bivariate Matrix-valued Linear Regression (BMLR): Finite-sample performance under Identifiability and Sparsity Assumptions},
  author = {Nayel Bettache},
  journal= {arXiv preprint arXiv:2412.17749},
  year   = {2025}
}

Comments

36 pages, 8 figures

R2 v1 2026-06-28T20:47:05.080Z