The two-to-infinity norm and singular subspace geometry with applications to high-dimensional statistics

DOI10.1214/18-AOS1752zbMATH Open1470.62065arXiv1705.10735OpenAlexW2966291568WikidataQ114060556 ScholiaQ114060556MaRDI QIDQ2328047FDOQ2328047

Authors: Joshua Cape, Minh Tang, Carey E. Priebe 

Publication date: 9 October 2019

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: The singular value matrix decomposition plays a ubiquitous role throughout statistics and related fields. Myriad applications including clustering, classification, and dimensionality reduction involve studying and exploiting the geometric structure of singular values and singular vectors. This paper provides a novel collection of technical and theoretical tools for studying the geometry of singular subspaces using the two-to-infinity norm. Motivated by preliminary deterministic Procrustes analysis, we consider a general matrix perturbation setting in which we derive a new Procrustean matrix decomposition. Together with flexible machinery developed for the two-to-infinity norm, this allows us to conduct a refined analysis of the induced perturbation geometry with respect to the underlying singular vectors even in the presence of singular value multiplicity. Our analysis yields singular vector entrywise perturbation bounds for a range of popular matrix noise models, each of which has a meaningful associated statistical inference task. In addition, we demonstrate how the two-to-infinity norm is the preferred norm in certain statistical settings. Specific applications discussed in this paper include covariance estimation, singular subspace recovery, and multiple graph inference. Both our Procrustean matrix decomposition and the technical machinery developed for the two-to-infinity norm may be of independent interest.

Full work available at URL: https://arxiv.org/abs/1705.10735

Recommendations

• Rate-optimal perturbation bounds for singular subspaces with applications to high-dimensional statistics
• High-dimensional \(p\)-norms
• Estimation in high dimensions: a geometric perspective
• scientific article; zbMATH DE number 1740315
• scientific article; zbMATH DE number 1302143
• Asymptotic normality of interpoint distances for high-dimensional data with applications to the two-sample problem
• A high-dimensional two-sample test for the mean using random subspaces
• An introduction to recent advances in high/infinite dimensional statistics
• The geometry of statistical efficiency and matrix statistics
• Perspective functions: proximal calculus and applications in high-dimensional statistics

zbMATH Keywords

principal component analysishigh-dimensional statisticssingular value decompositionspectral methodsProcrustes analysiseigenvector perturbation

Mathematics Subject Classification ID

Factor analysis and principal components; correspondence analysis (62H25) Classification and discrimination; cluster analysis (statistical aspects) (62H30) Estimation in multivariate analysis (62H12) Inference from stochastic processes and spectral analysis (62M15)

Cites Work

• Principal component analysis.
• Spectral clustering and the high-dimensional stochastic blockmodel
• Statistics on special manifolds
• The Rotation of Eigenvectors by a Perturbation. III
• Large Covariance Estimation by Thresholding Principal Orthogonal Complements
• Spectral analysis of large dimensional random matrices
• On the distribution of the largest eigenvalue in principal components analysis
• Consistency of spectral clustering in stochastic block models
• Statistical shape analysis. With applications in R
• Title not available (Why is that?)
• The Geometry of Algorithms with Orthogonality Constraints
• High-dimensional probability. An introduction with applications in data science
• A useful variant of the Davis-Kahan theorem for statisticians
• Random Dot Product Graph Models for Social Networks
• Title not available (Why is that?)
• Finite sample approximation results for principal component analysis: A matrix perturbation approach
• Exact matrix completion via convex optimization
• A nonparametric two-sample hypothesis testing problem for random graphs
• Rate-optimal perturbation bounds for singular subspaces with applications to high-dimensional statistics
• Entrywise eigenvector analysis of random matrices with low expected rank
• Sparse PCA: optimal rates and adaptive estimation
• Spectra of edge-independent random graphs
• Consistent adjacency-spectral partitioning for the stochastic block model when the model parameters are unknown
• Perturbation bounds in connection with singular value decomposition
• Random matrix theory in statistics: a review
• A Consistent Adjacency Spectral Embedding for Stochastic Blockmodel Graphs
• Perfect clustering for stochastic blockmodel graphs via adjacency spectral embedding
• Role of normalization in spectral clustering for stochastic blockmodels
• The eigenvalues and eigenvectors of finite, low rank perturbations of large random matrices
• Large sample covariance matrices and high-dimensional data analysis
• Delocalization of eigenvectors of random matrices with independent entries
• Manifold matching: joint optimization of fidelity and commensurability
• The Procrustes Problem for Orthogonal Stiefel Matrices
• Non-Euclidean statistics for covariance matrices, with applications to diffusion tensor imaging
• New asymptotic results in principal component analysis
• Concentration inequalities and moment bounds for sample covariance operators
• Eigenvectors of random matrices: A survey
• Norms of random matrices: local and global problems
• Signal-plus-noise matrix models: eigenvector deviations and fluctuations
• Estimation of functionals of sparse covariance matrices
• Limit theorems for eigenvectors of the normalized Laplacian for random graphs
• Optimization via low-rank approximation for community detection in networks
• Random perturbation of low rank matrices: improving classical bounds
• An \(\ell_{\infty}\) eigenvector perturbation bound and its application
• Unperturbed: spectral analysis beyond Davis-Kahan
• Estimating Mixed Memberships With Sharp Eigenvector Deviations
• Asymptotically efficient estimators for stochastic blockmodels: the naive MLE, the rank-constrained MLE, and the spectral estimator
• Information Recovery in Shuffled Graphs via Graph Matching

Cited In (32)

• A Schatten-\(q\) low-rank matrix perturbation analysis via perturbation projection error bound
• Compressed spectral screening for large-scale differential correlation analysis with application in selecting glioblastoma gene modules
• Crowdsourcing Utilizing Subgroup Structure of Latent Factor Modeling
• A note on the orthogonal Procrustes problem and norm-dependent optimality
• The Sup-norm Perturbation of HOSVD and Low Rank Tensor Denoising
• Rate-optimal perturbation bounds for singular subspaces with applications to high-dimensional statistics
• Entrywise eigenvector analysis of random matrices with low expected rank
• Strong consistency guarantees for clustering high-dimensional bipartite graphs with the spectral method
• Normal approximation and confidence region of singular subspaces
• Asymptotically efficient estimators for stochastic blockmodels: the naive MLE, the rank-constrained MLE, and the spectral estimator
• Uniform Bounds for Invariant Subspace Perturbations
• Entrywise limit theorems for eigenvectors of signal-plus-noise matrix models with weak signals
• A performance guarantee for spectral clustering
• Spectral Clustering via Adaptive Layer Aggregation for Multi-Layer Networks
• Robust high-dimensional factor models with applications to statistical machine learning
• Subspace estimation from unbalanced and incomplete data matrices: \({\ell_{2,\infty}}\) statistical guarantees
• On estimation and inference in latent structure random graphs
• Title not available (Why is that?)
• Spectral Embedding of Weighted Graphs
• Estimating Mixed Memberships With Sharp Eigenvector Deviations
• Title not available (Why is that?)
• Euclidean Representation of Low-Rank Matrices and Its Geometric Properties
• Exact minimax optimality of spectral methods in phase synchronization and orthogonal group synchronization
• Leave-one-out singular subspace perturbation analysis for spectral clustering
• Spectral clustering revisited: information hidden in the Fiedler vector
• Applications of dual regularized Laplacian matrix for community detection
• Vertex nomination, consistent estimation, and adversarial modification
• Bayesian sparse spiked covariance model with a continuous matrix shrinkage prior
• Singular vector and singular subspace distribution for the matrix denoising model
• Inference for heteroskedastic PCA with missing data
• Analysis of spectral clustering algorithms for community detection: the general bipartite setting
• An \({\ell_p}\) theory of PCA and spectral clustering

This page was built for publication: The two-to-infinity norm and singular subspace geometry with applications to high-dimensional statistics

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2328047)