Regularized $M$-estimators of scatter matrix

Abstract

In this paper, a general class of regularized $M$-estimators of scatter matrix are proposed which are suitable also for low or insufficient sample support (small $n$ and large $p$) problems. The considered class constitutes a natural generalization of $M$-estimators of scatter matrix (Maronna, 1976) and are defined as a solution to a penalized $M$-estimation cost function that depend on a pair $(\alpha ,\beta )$ of regularization parameters. We derive general conditions for uniqueness of the solution using concept of geodesic convexity. Since these conditions do not include Tyler's $M$-estimator, necessary and sufficient conditions for uniqueness of the penalized Tyler's cost function are established separately. For the regularized Tyler's $M$-estimator, we also derive a simple, closed form and data dependent solution for choosing the regularization parameter based on shape matrix matching in the mean squared sense. An iterative algorithm that converges to the solution of the regularized $M$-estimating equation is also provided. Finally, some simulations studies illustrate the improved accuracy of the proposed regularized $M$-estimators of scatter compared to their non-regularized counterparts in low sample support problems. An example of radar detection using normalized matched filter (NMF) illustrate that an adaptive NMF detector based on regularized $M$-estimators are able to maintain accurately the preset CFAR level and at at the same time provide similar probability of detection as the (theoretical) NMF detector.

Submitted to IEEE Transactions on Signal Processing (contains a corrected proof of convergence of the proposed iterative algorithm)

References in corpus (3)

• Robust Shrinkage Estimation of High-dimensional Covariance Matrices
• Generalized robust shrinkage estimator and its application to STAP detection problem
• Multivariate Generalized Gaussian Distribution: Convexity and Graphical Models

Cited by in corpus (24)

• An Overview of Robust Subspace Recovery
• Radar Adaptive Detection Architectures for Heterogeneous Environments
• Optimal shrinkage covariance matrix estimation under random sampling from elliptical distributions
• Interpolation and Extrapolation of Toeplitz Matrices via Optimal Mass Transport
• Adaptive Radar Detection in Heterogeneous Clutter-dominated Environments
• Robust Sparse Covariance Estimation by Thresholding Tyler's M-Estimator
• Challenges of cellwise outliers
• Simultaneous penalized M-estimation of covariance matrices using geodesically convex optimization
• Robust M-Estimation Based Bayesian Cluster Enumeration for Real Elliptically Symmetric Distributions
• Regularized Covariance Estimation for Polarization Radar Detection in Compound Gaussian Sea Clutter
• New insights into the statistical properties of $M$-estimators
• On the asymptotics of Maronna's robust PCA
• Large-dimensional behavior of regularized Maronna's M-estimators of covariance matrices
• Convergence and Fluctuations of Regularized Tyler Estimators
• Real Elliptically Skewed Distributions and Their Application to Robust Cluster Analysis
• A universal framework for learning the elliptical mixture model
• R-NL: Covariance Matrix Estimation for Elliptical Distributions based on Nonlinear Shrinkage
• Riemannian optimization for non-centered mixture of scaled Gaussian distributions
• Robust self-tuning semiparametric PCA for contaminated elliptical distribution
• Cross-Validated Tuning of Shrinkage Factors for MVDR Beamforming Based on Regularized Covariance Matrix Estimation
• Tyler's and Maronna's M-estimators: Non-Asymptotic Concentration Results
• Robust high-dimensional precision matrix estimation
• Three Operator Splitting with a Nonconvex Loss Function
• Breakdown points of penalized and hybrid M-estimators of covariance