Performance Analysis of Tyler's Covariance Estimator
arXiv:1401.6926 · doi:10.1109/TSP.2014.2376911
Abstract
This paper analyzes the performance of Tyler's M-estimator of the scatter matrix in elliptical populations. We focus on the non-asymptotic setting and derive the estimation error bounds depending on the number of samples n and the dimension p. We show that under quite mild conditions the squared Frobenius norm of the error of the inverse estimator decays like p^2/n with high probability.
References in corpus (6)
- Regularized estimation of large covariance matrices
- Sparse permutation invariant covariance estimation
- On the Fisher information matrix for multivariate elliptically contoured distributions
- Large Dimensional Analysis and Optimization of Robust Shrinkage Covariance Matrix Estimators
- The Random Matrix Regime of Maronna's M-estimator with elliptically distributed samples
- Marchenko-Pastur Law for Tyler's M-estimator
Cited by in corpus (6)
- On the asymptotics of Maronna's robust PCA
- Rigorous Guarantees for Tyler's M-estimator via quantum expansion
- Ledoit-Wolf linear shrinkage with unknown mean
- Group Symmetric Robust Covariance Estimation
- Tyler's and Maronna's M-estimators: Non-Asymptotic Concentration Results
- Gaussian and Robust Kronecker Product Covariance Estimation: Existence and Uniqueness