Nonparametric estimation of low rank matrix valued function

DOI10.1214/19-EJS1582MaRDI QIDQ2326073zbMATH OpenFDO

Authors Yanyan Li

Publication date 4 October 2019

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

Full work available at URL https://arxiv.org/abs/1802.06292, https://projecteuclid.org/euclid.ejs/1570068045

zbMATH Keywords

local polynomial estimatorlow rankmatrix completionminimax lower boundmodel selectionnonparametric estimationnuclear norm penalization

Mathematics Subject Classification ID

Nonparametric estimation (62G05) Nonparametric regression and quantile regression (62G08) Estimation in multivariate analysis (62H12)

Abstract: Let $A:[0,1]ightarrowmathbb{H}_m$ (the space of Hermitian matrices) be a matrix valued function which is low rank with entries in H"{o}lder class $$. The goal of this paper is to study statistical estimation of $A$ based on the regression model $mathbbE(Y_j|a{u}_j,X_j)=langleA(a{u}_j),X_{j}angle,$ where $a{u}_j$ are i.i.d. uniformly distributed in $[0,1]$, $X_j$ are i.i.d. matrix completion sampling matrices, $Y_j$ are independent bounded responses. We propose an innovative nuclear norm penalized local polynomial estimator and establish an upper bound on its point-wise risk measured by Frobenius norm. Then we extend this estimator globally and prove an upper bound on its integrated risk measured by $L_2$-norm. We also propose another new estimator based on bias-reducing kernels to study the case when $A$ is not necessarily low rank and establish an upper bound on its risk measured by $L_{infty}$-norm. We show that the obtained rates are all optimal up to some logarithmic factor in minimax sense. Finally, we propose an adaptive estimation procedure based on Lepskii's method and model selection with data splitting which is computationally efficient and can be easily implemented and parallelized.

Recommendations

• Nuclear-norm penalization and optimal rates for noisy low-rank matrix completion
• Estimation of high-dimensional low-rank matrices
• Robust low-rank matrix estimation
• Robust matrix completion
• Rank penalized estimators for high-dimensional matrices

Cites work

• A Bennett concentration inequality and its application to suprema of empirical processes
• A Singular Value Thresholding Algorithm for Matrix Completion
• Convex trace functions and the Wigner-Yanase-Dyson conjecture
• Distance shrinkage and Euclidean embedding via regularized kernel estimation
• Distributed optimization and statistical learning via the alternating direction method of multipliers
• Estimation of high-dimensional low-rank matrices
• Exact matrix completion via convex optimization
• Guaranteed minimum-rank solutions of linear matrix equations via nuclear norm minimization
• scientific article; zbMATH DE number 991833 (Why is no real title available?)
• scientific article; zbMATH DE number 3133919 (Why is no real title available?)
• scientific article; zbMATH DE number 17222 (Why is no real title available?)
• scientific article; zbMATH DE number 1943013 (Why is no real title available?)
• Local Rademacher complexities and oracle inequalities in risk minimization. (2004 IMS Medallion Lecture). (With discussions and rejoinder)
• Matrix completion via an alternating direction method
• Matrix estimation by universal singular value thresholding
• Model selection in nonparametric regression
• New concentration inequalities in product spaces
• Nuclear-norm penalization and optimal rates for noisy low-rank matrix completion
• On a Problem of Adaptive Estimation in Gaussian White Noise
• On the Tchebychef Inequality of Bernstein
• Optimal pointwise adaptive methods in nonparametric estimation
• Optimal spatial adaptation to inhomogeneous smoothness: An approach based on kernel estimates with variable bandwidth selectors
• Oracle inequalities in empirical risk minimization and sparse recovery problems. École d'Été de Probabilités de Saint-Flour XXXVIII-2008.
• Recovering Low-Rank Matrices From Few Coefficients in Any Basis
• Restricted strong convexity and weighted matrix completion: optimal bounds with noise
• Sharp oracle inequalities in low rank estimation
• Strong converse for identification via quantum channels
• The Power of Convex Relaxation: Near-Optimal Matrix Completion
• Uniqueness of low-rank matrix completion by rigidity theory
• User-friendly tail bounds for sums of random matrices
• Von Neumann entropy penalization and low-rank matrix estimation

Cited in

(2)

• Estimation of high-dimensional low-rank matrices
• The statistical rate for support matrix machines under low rankness and row (column) sparsity

This page was built for publication: Nonparametric estimation of low rank matrix valued function

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2326073)