Covariance estimation for distributions with 2+ moments
From MaRDI portal
Publication:378788
Abstract: We study the minimal sample size N=N(n) that suffices to estimate the covariance matrix of an n-dimensional distribution by the sample covariance matrix in the operator norm, with an arbitrary fixed accuracy. We establish the optimal bound N=O(n) for every distribution whose k-dimensional marginals have uniformly bounded moments outside the sphere of radius . In the specific case of log-concave distributions, this result provides an alternative approach to the Kannan-Lovasz-Simonovits problem, which was recently solved by Adamczak et al. [J. Amer. Math. Soc. 23 (2010) 535-561]. Moreover, a lower estimate on the covariance matrix holds under a weaker assumption - uniformly bounded moments of one-dimensional marginals. Our argument consists of randomizing the spectral sparsifier, a deterministic tool developed recently by Batson, Spielman and Srivastava [SIAM J. Comput. 41 (2012) 1704-1721]. The new randomized method allows one to control the spectral edges of the sample covariance matrix via the Stieltjes transform evaluated at carefully chosen random points.
Recommendations
- How close is the sample covariance matrix to the actual covariance matrix?
- Estimating the covariance of random matrices
- Robust covariance estimation under \(L_4\)-\(L_2\) norm equivalence
- Partial estimation of covariance matrices
- Quantitative estimates of the convergence of the empirical covariance matrix in log-concave ensembles
Cites work
- A note on the largest eigenvalue of a large dimensional sample covariance matrix
- Concentration of mass on convex bodies
- Extremal properties of Rademacher functions with applications to the Khintchine and Rosenthal inequalities
- How close is the sample covariance matrix to the actual covariance matrix?
- scientific article; zbMATH DE number 1254560 (Why is no real title available?)
- Limit of the smallest eigenvalue of a large dimensional sample covariance matrix
- On the subspaces of \(L^p\) \((p > 2)\) spanned by sequences of independent random variables
- Quantitative estimates of the convergence of the empirical covariance matrix in log-concave ensembles
- Random vectors in the isotropic position
- Random walks and anO*(n5) volume algorithm for convex bodies
- Sharp bounds on the rate of convergence of the empirical covariance matrix
- Some estimates of norms of random matrices
- The eigenvalues and eigenvectors of finite, low rank perturbations of large random matrices
- Twice-Ramanujan sparsifiers
Cited in
(53)- Coverings of random ellipsoids, and invertibility of matrices with i.i.d. heavy-tailed entries
- Almost sure convergence of the largest and smallest eigenvalues of high-dimensional sample correlation matrices
- The spectral norm of random inner-product kernel matrices
- The smallest singular value of a shifted d-regular random square matrix
- Sub-Gaussian estimators of the mean of a random matrix with heavy-tailed entries
- Robust high-dimensional factor models with applications to statistical machine learning
- On Monte-Carlo methods in convex stochastic optimization
- Exact minimax risk for linear least squares, and the lower tail of sample covariance matrices
- Covariance estimation under one-bit quantization
- Sampling discretization and related problems
- Covariances with OWA operators and Bonferroni means
- Estimating covariance and precision matrices along subspaces
- Generalized canonical correlation analysis for classification
- On the convergence of the extremal eigenvalues of empirical covariance matrices with dependence
- On the interval of fluctuation of the singular values of random matrices
- The smallest singular value of random rectangular matrices with no moment assumptions on entries
- On the behaviour of the smallest eigenvalue of a high-dimensional sample covariance matrix
- Controlling the least eigenvalue of a random Gram matrix
- The masked sample covariance estimator: an analysis using matrix concentration inequalities
- The relative effects of dimensionality and multiplicity of hypotheses on the F-test in linear regression
- The lower tail of random quadratic forms with applications to ordinary least squares
- Concentration phenomena in high dimensional geometry
- Quantitative estimates of the convergence of the empirical covariance matrix in log-concave ensembles
- Estimating the covariance of random matrices
- Restricted invertibility revisited
- On higher order isotropy conditions and lower bounds for sparse quadratic forms
- The limit of the smallest singular value of random matrices with i.i.d. entries
- Approximating the covariance ellipsoid
- On the discrepancy of random matrices with many columns
- Stochastic analysis of covariance when the error distribution is long-tailed symmetric
- Sparse recovery under weak moment assumptions
- scientific article; zbMATH DE number 7164767 (Why is no real title available?)
- Quantitative Version of a Silverstein’s Result
- On Sufficient Conditions in the Marchenko--Pastur Theorem
- Non-asymptotic bounds for the \(\ell_{\infty}\) estimator in linear regression with uniform noise
- Dimension-free bounds for sums of independent matrices and simple tensors via the variational principle
- Dimension-free bounds for sums of dependent matrices and operators with heavy-tailed distributions
- Minimizing convex functions with rational minimizers
- Large-dimensional central limit theorem with fourth-moment error bounds on convex sets and balls
- Robust estimation of covariance matrices: adversarial contamination and beyond
- Covariance estimation under missing observations and L₄ - L₂ moment equivalence
- Asymptotic geometric analysis: achievements and perspective
- Covariance estimation: optimal dimension-free guarantees for adversarial corruption and heavy tails
- Matrix Freedman inequality for sub-Weibull martingales
- Eldan's stochastic localization and the KLS conjecture: isoperimetry, concentration and mixing
- Sharper dimension-free bounds on the Frobenius distance between sample covariance and its expectation
- Randomized low-rank approximations beyond Gaussian random matrices
- Partial estimation of covariance matrices
- Sharp bounds for max-sliced Wasserstein distances
- Concentration and moment inequalities for sums of independent heavy-tailed random matrices
- Reducing isotropy and volume to KLS: faster rounding and volume algorithms
- Comparison theorems for the minimum eigenvalue of a random positive-semidefinite matrix
- Moment bounds for large autocovariance matrices under dependence
This page was built for publication: Covariance estimation for distributions with \({2+\varepsilon}\) moments
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q378788)
