Optimal sparse singular value decomposition for high-dimensional high-order data

DOI10.1080/01621459.2018.1527227zbMATH Open1428.62262arXiv1809.01796OpenAlexW2891722575MaRDI QIDQ5208075FDOQ5208075

Authors: Anru Zhang, Rungang Han 

Publication date: 15 January 2020

Published in: Journal of the American Statistical Association (Search for Journal in Brave)

Abstract: In this article, we consider the sparse tensor singular value decomposition, which aims for dimension reduction on high-dimensional high-order data with certain sparsity structure. A method named Sparse Tensor Alternating Thresholding for Singular Value Decomposition (STAT-SVD) is proposed. The proposed procedure features a novel double projection & thresholding scheme, which provides a sharp criterion for thresholding in each iteration. Compared with regular tensor SVD model, STAT-SVD permits more robust estimation under weaker assumptions. Both the upper and lower bounds for estimation accuracy are developed. The proposed procedure is shown to be minimax rate-optimal in a general class of situations. Simulation studies show that STAT-SVD performs well under a variety of configurations. We also illustrate the merits of the proposed procedure on a longitudinal tensor dataset on European country mortality rates.

Full work available at URL: https://arxiv.org/abs/1809.01796

Recommendations

• Rank-one approximation to high order tensors
• A new tensor decomposition
• Provable Sparse Tensor Decomposition
• On the Tensor SVD and the Optimal Low Rank Orthogonal Approximation of Tensors
• A Multilinear Singular Value Decomposition

zbMATH Keywords

singular value decompositionsparsityhigh-dimensional high-order dataprojection and thresholdingTucker low-rank tensor

Mathematics Subject Classification ID

Nonparametric estimation (62G05) Factor analysis and principal components; correspondence analysis (62H25)

Cites Work

• Principal component analysis.
• On consistency and sparsity for principal components analysis in high dimensions
• Tensor Decompositions and Applications
• Selecting the number of principal components: estimation of the true rank of a noisy matrix
• Sparse principal component analysis via regularized low rank matrix approximation
• Rate-optimal perturbation bounds for singular subspaces with applications to high-dimensional statistics
• Sparse PCA: optimal rates and adaptive estimation
• A Multilinear Singular Value Decomposition
• On the Best Rank-1 and Rank-(R₁ ,R₂ ,. . .,R_N) Approximation of Higher-Order Tensors
• Influential features PCA for high dimensional clustering
• Tensor Rank and the Ill-Posedness of the Best Low-Rank Approximation Problem
• Provable Sparse Tensor Decomposition
• Most tensor problems are NP-hard
• Generalized L-, M-, and R-statistics
• Tensor decompositions for learning latent variable models
• Optimal estimation of low rank density matrices
• Biclustering via sparse singular value decomposition
• An alternative point of view on Lepski's method
• Rate Optimal Denoising of Simultaneously Sparse and Low Rank Matrices
• Dynamic tensor clustering

Cited In (23)

• Model-Based Tensor Low-Rank Clustering
• Adaptive singular value shrinkage estimate for low rank tensor denoising
• Tucker-3 decomposition with sparse core array using a penalty function based on Gini-index
• Guaranteed Functional Tensor Singular Value Decomposition
• Multivariate Temporal Point Process Regression
• Smoothing composite proximal gradient algorithm for sparse group Lasso problems with nonsmooth loss functions
• Covariate-Assisted Sparse Tensor Completion
• ISLET: fast and optimal low-rank tensor regression via importance sketching
• Title not available (Why is that?)
• Tensor clustering with planted structures: statistical optimality and computational limits
• An optimal statistical and computational framework for generalized tensor estimation
• Inference for low-rank tensors -- no need to debias
• A Doubly Enhanced EM Algorithm for Model-Based Tensor Clustering
• A general framework for tensor screening through smoothing
• Title not available (Why is that?)
• Title not available (Why is that?)
• Generalized Liquid Association Analysis for Multimodal Data Integration
• Several approximation algorithms for sparse best rank-1 approximation to higher-order tensors
• Statistical inference on the significance of rows and columns for matrix-valued data in an additive model
• Community detection on mixture multilayer networks via regularized tensor decomposition
• Adaptive higher-order spectral estimators
• Correlation Tensor Decomposition and Its Application in Spatial Imaging Data
• High order singular value decomposition for plant diversity estimation

Uses Software

• MultiCluster

This page was built for publication: Optimal sparse singular value decomposition for high-dimensional high-order data

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5208075)