Robust regularized covariance matrix estimation: well-posedness and convergent algorithm
Methodology
2026-03-31 v1 Computation
Abstract
In this paper, we study properties of penalized and structured M-estimators of multivariate scatter, based on geodesically convex but not necessarily smooth penalty functions. Existence and uniqueness conditions for these penalized and structured estimators are given. However, we show that the standard fixed-point algorithm which is usually applied to an M-estimation problem does not necessarily converge for penalized M-estimation problems. Hence, we develop a new but simple re-weighting algorithm and prove that it has monotone convergence for a broad class of penalized and structured M-estimators of multivariate scatter.
Cite
@article{arxiv.2603.27487,
title = {Robust regularized covariance matrix estimation: well-posedness and convergent algorithm},
author = {Mengxi Yi and David Tyler},
journal= {arXiv preprint arXiv:2603.27487},
year = {2026}
}