Intelligent Initialization and Adaptive Thresholding for Iterative Matrix Completion: Some Statistical and Algorithmic Theory forAdaptive-Impute

DOI10.1080/10618600.2018.1518238OpenAlexW2963834768MaRDI QIDQ3391239

Juhee Cho, Donggyu Kim, Karl Rohe

Publication date: 28 March 2022

Published in: Journal of Computational and Graphical Statistics (Search for Journal in Brave)

Full work available at URL: https://doi.org/10.1080/10618600.2018.1518238

zbMATH Keywords

nonconvex optimizationthresholded singular value decomposition\textit{softImpute}generalized-\textit{softImpute}

Mathematics Subject Classification ID

Statistics (62-XX)

Uses Software

• R
• ggplot2
• softImpute
• dplyr
• rARPACK

Cites Work

• Unnamed Item
• Unnamed Item
• Matrix Completion and Low-Rank SVD via Fast Alternating Least Squares
• The Adaptive Lasso and Its Oracle Properties
• Minimax risk of matrix denoising by singular value thresholding
• Estimation of (near) low-rank matrices with noise and high-dimensional scaling
• Nuclear-norm penalization and optimal rates for noisy low-rank matrix completion
• Matrix estimation by universal singular value thresholding
• The Optimal Hard Threshold for Singular Values is <inline-formula> <tex-math notation="TeX">\(4/\sqrt {3}\) </tex-math></inline-formula>
• A Singular Value Thresholding Algorithm for Matrix Completion
• Adaptive Thresholding for Sparse Covariance Matrix Estimation
• Asymptotic theory for estimating the singular vectors and values of a partially-observed low rank matrix with noise
• Large Covariance Estimation by Thresholding Principal Orthogonal Complements