SuperPCA: subspace analysis and an efficient algorithm for high-dimensional PCA

DOI10.48550/arXiv.2609.26406MaRDI QIDQ7353516arXivFDO

Authors Irina-Beatrice Haas, Maike Meier, Yuji Nakatsukasa, Taejun Park

Publication date 22 September 2026

Published in arXiv

Tldr We present SuperPCA (SUbsPace subsamplER PCA), an algorithm that efficiently estimates leading signals in high-dimensional data by leveraging an approximate eigenspace of the sample covariance matrix obtained from only a small number of subsampled coordinates, achieving up to 10x accuracy gain over classical PCA.

Full work available at URL https://doi.org/10.48550/arXiv.2609.26406

arXiv classification

cs.LGmath.NAstat.COstat.ML

Cites work

• A limit theorem for the norm of random matrices
• A new selection operator for the discrete empirical interpolation method -- improved a priori error bound and extensions
• A rationale and test for the number of factors in factor analysis
• Accuracy and stability of CUR decompositions with oversampling
• Accuracy and Stability of CUR decompositions with Oversampling
• Asymptotic Theory for Principal Component Analysis
• Asymptotics of sample eigenstructure for a large dimensional spiked covariance model
• Determinantal point processes for machine learning
• DISTRIBUTION OF EIGENVALUES FOR SOME SETS OF RANDOM MATRICES
• Eigenvalues of large sample covariance matrices of spiked population models
• Fast approximation of matrix coherence and statistical leverage
• Finite sample approximation results for principal component analysis: A matrix perturbation approach
• Finite Sample Guarantees for PCA in Non-Isotropic and Data-Dependent Noise
• Generalized power method for sparse principal component analysis
• Gradient-based learning applied to document recognition
• Large sample covariance matrices and high-dimensional data analysis
• Limit of the smallest eigenvalue of a large dimensional sample covariance matrix
• Matrix Computations
• Non-Parametric Detection of the Number of Signals: Hypothesis Testing and Random Matrix Theory
• Nonasymptotic upper bounds for the reconstruction error of PCA
• Nonlinear model reduction via discrete empirical interpolation
• Normal approximation and concentration of spectral projectors of sample covariance
• Numerical methods for large eigenvalue problems
• On consistency and sparsity for principal components analysis in high dimensions
• On the distribution of the largest eigenvalue in principal components analysis
• Perturbation bounds in connection with singular value decomposition
• Phase transition of the largest eigenvalue for nonnull complex sample covariance matrices
• Randomly pivoted Cholesky: practical approximation of a kernel matrix with few entry evaluations
• Sample Eigenvalue Based Detection of High-Dimensional Signals in White Noise Using Relatively Few Samples
• Sampling algorithms for l₂ regression and applications
• Sharp error bounds for Ritz vectors and approximate singular vectors
• Simpler is better: A comparative study of randomized algorithms for computing the CUR decomposition
• Sparse Principal Component Analysis
• SubApSnap: Solving parameter-dependent linear systems with a snapshot and subsampling
• The Rotation of Eigenvectors by a Perturbation. III
• The smallest eigenvalue of a large dimensional Wishart matrix
• The Symmetric Eigenvalue Problem
• The Use of the L-Curve in the Regularization of Discrete Ill-Posed Problems
• Truncated power method for sparse eigenvalue problems

This page was built for publication: SuperPCA: subspace analysis and an efficient algorithm for high-dimensional PCA

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7353516)