Multidimensional linear functional estimation in sparse Gaussian models and robust estimation of the mean

DOI10.1214/19-EJS1590zbMATH Open1432.62203arXiv1712.05495OpenAlexW2970121140MaRDI QIDQ2323942FDOQ2323942

Authors: Olivier Collier, Arnak S. Dalalyan 

Publication date: 13 September 2019

Published in: Electronic Journal of Statistics (Search for Journal in Brave)

Abstract: We consider two problems of estimation in high-dimensional Gaussian models. The first problem is that of estimating a linear functional of the means of $n$ independent $p$-dimensional Gaussian vectors, under the assumption that most of these means are equal to zero. We show that, up to a logarithmic factor, the minimax rate of estimation in squared Euclidean norm is between $(s^{2}wedgen)+sp$ and $(s^{2}wedgenp)+sp$. The estimator that attains the upper bound being computationally demanding, we investigate suitable versions of group thresholding estimators that are efficiently computable even when the dimension and the sample size are very large. An interesting new phenomenon revealed by this investigation is that the group thresholding leads to a substantial improvement in the rate as compared to the element-wise thresholding. Thus, the rate of the group thresholding is $s^{2}sqrtp+sp$, while the element-wise thresholding has an error of order $s^{2}p+sp$. To the best of our knowledge, this is the first known setting in which leveraging the group structure leads to a polynomial improvement in the rate. The second problem studied in this work is the estimation of the common $p$-dimensional mean of the inliers among $n$ independent Gaussian vectors. We show that there is a strong analogy between this problem and the first one. Exploiting it, we propose new strategies of robust estimation that are computationally tractable and have better rates of convergence than the other computationally tractable robust (with respect to the presence of the outliers in the data) estimators studied in the literature. However, this tractability comes with a loss of the minimax-rate-optimality in some regimes.

Full work available at URL: https://arxiv.org/abs/1712.05495

Recommendations

• Excess-risk consistency of group-hard thresholding estimator in robust estimation of Gaussian mean
• All-in-one robust estimator of the Gaussian mean
• Optimal adaptive estimation of linear functionals under sparsity
• Minimax estimation of linear and quadratic functionals on sparsity classes
• On estimation of nonsmooth functionals of sparse normal means

zbMATH Keywords

high-dimensional inferencerobust estimationlinear transformationminimax estimationgroup-sparsitycolumn-sparsity

Mathematics Subject Classification ID

Linear regression; mixed models (62J05) Estimation in multivariate analysis (62H12) Robustness and adaptive procedures (parametric inference) (62F35)

Cites Work

• High-dimensional additive modeling
• Adaptive estimation of high-dimensional signal-to-noise ratios
• Robust Lasso With Missing and Grossly Corrupted Observations
• Model Selection and Estimation in Regression with Grouped Variables
• Robust Estimation of a Location Parameter
• On a Non-Parametric Analogue of the Information Matrix
• Geometric median and robust estimation in Banach spaces
• Component selection and smoothing in multivariate nonparametric regression
• Adaptive estimation of a quadratic functional by model selection.
• Minimax estimation of linear and quadratic functionals on sparsity classes
• The Group Square-Root Lasso: Theoretical Properties and Fast Algorithms
• Oracle inequalities and optimal inference under group sparsity
• Robust matrix completion
• Title not available (Why is that?)
• Adaptive estimation of linear functionals in the convolution model and applications
• Some theoretical results on the grouped variables Lasso
• High dimensional robust M-estimation: asymptotic variance via approximate message passing
• Minimax estimation of linear functionals over nonconvex parameter spaces.
• Adaptive estimation of linear functionals by model selection
• On a Problem of Adaptive Estimation in Gaussian White Noise
• Testing composite hypotheses, Hermite polynomials and optimal estimation of a nonsmooth functional
• On estimation of the \(L_r\) norm of a regression function
• On spatially adaptive estimation of nonparametric regression
• An oracle approach to adaptive estimation of linear functionals in a Gaussian model
• Curve registration by nonparametric goodness-of-fit testing
• Nonparametric estimation by convex programming
• Robust estimation for an inverse problem arising in multiview geometry
• A general decision theory for Huber's \(\epsilon\)-contamination model
• Sub-Gaussian mean estimators
• Minimax quadratic estimation of a quadratic functional
• Optimal adaptive estimation of a quadratic functional
• Minimax testing of a composite null hypothesis defined via a quadratic functional in the model of regression
• Adaptive estimates of linear functionals
• On adaptive estimation of linear functionals
• Sharp adaptive estimation of linear functionals.
• Robust estimators in high-dimensions without the computational intractability
• Tight conditions for consistency of variable selection in the context of high dimensionality
• Robust covariance and scatter matrix estimation under Huber's contamination model
• Learning from MOM's principles: Le Cam's approach
• Optimal adaptive estimation of linear functionals under sparsity
• Estimating linear functionals of a sparse family of Poisson means
• Estimating minimum effect with outlier selection

Cited In (7)

• Excess-risk consistency of group-hard thresholding estimator in robust estimation of Gaussian mean
• Robust shrinkage estimation and selection for functional multiple linear model through LAD loss
• All-in-one robust estimator of the Gaussian mean
• Confidence regions and minimax rates in outlier-robust estimation on the probability simplex
• Adaptive and robust multi-task learning
• Finite sample properties of parametric MMD estimation: robustness to misspecification and dependence
• On sampling from a log-concave density using kinetic Langevin diffusions

This page was built for publication: Multidimensional linear functional estimation in sparse Gaussian models and robust estimation of the mean

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2323942)