A Higher-Order Generalized Singular Value Decomposition for Rank-Deficient Matrices

DOI10.1137/21M1443881MaRDI QIDQ6116662zbMATH OpenOpenAlexFDO

Authors Paul J. Goulart, Stephen R. Duncan

Publication date 16 August 2023

Published in SIAM Journal on Matrix Analysis and Applications (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2102.09822

zbMATH Keywords

diagonalizationhigher-order cosine-sine decompositionhigher-order generalized singular value decompositionmultimodal data fusion

Mathematics Subject Classification ID

Factorization of matrices (15A23) Numerical computation of eigenvalues and eigenvectors of matrices (65F15) Numerical methods for low-rank matrix approximation; matrix compression (65F55)

Abstract: The higher-order generalized singular value decomposition (HO-GSVD) is a matrix factorization technique that extends the GSVD to $Nge2$ data matrices, and can be used to identify shared subspaces in multiple large-scale datasets with different row dimensions. The standard HO-GSVD factors $N$ matrices $A_{i}inmathbb{R}^{m_{i}imesn}$ as $A_i=U_{i}Sigm{a}_i{V}^{e}xtT$, but requires that each of the matrices $A_i$ has full column rank. We propose a modification of the HO-GSVD that extends its applicability to rank-deficient data matrices $A_i$. If the matrix of stacked $A_i$ has full rank, we show that the properties of the original HO-GSVD extend to our approach. We extend the notion of common subspaces to isolated subspaces, which identify features that are unique to one $A_i$. We also extend our results to the higher-order cosine-sine decomposition (HO-CSD), which is closely related to the HO-GSVD. Our extension of the standard HO-GSVD allows its application to datasets with with $m_i<n$ or $extrank(A_i)<n$, such as are encountered in bioinformatics, neuroscience, control theory or classification problems.

Recommendations

• A new matrix maximization model for computing ratios of generalized singular values from high-order GSVD
• A Multilinear Singular Value Decomposition
• Randomized generalized singular value decomposition
• The GSVD: where are the ellipses?, matrix trigonometry, and more
• Dimensionality reduction in higher-order signal processing and rank-\((R_1,R_2,\ldots,R_N)\) reduction in multilinear algebra

Cites work

• A Multilinear Singular Value Decomposition
• A Practical Randomized CP Tensor Decomposition
• A survey of multilinear subspace learning for tensor data
• Computing the CS and the generalized singular value decompositions
• Generalizing the Singular Value Decomposition
• scientific article; zbMATH DE number 3678541 (Why is no real title available?)
• scientific article; zbMATH DE number 3014822 (Why is no real title available?)
• scientific article; zbMATH DE number 6159604 (Why is no real title available?)
• Linear algebra and its applications
• On a Variational Formulation of the Generalized Singular Value Decomposition
• On the Sensitivity of the Eigenvalue Problem Ax = \lambda Bx
• Practical leverage-based sampling for low-rank tensor decomposition
• Randomized algorithms for generalized singular value decomposition with application to sensitivity analysis
• The generalized singular value decomposition and the general (A- B)-problem
• Towards a Generalized Singular Value Decomposition

Cited in

(2)

• A new matrix maximization model for computing ratios of generalized singular values from high-order GSVD
• Control of multi-array cross-directional systems using the generalised singular value decomposition

This page was built for publication: A Higher-Order Generalized Singular Value Decomposition for Rank-Deficient Matrices

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6116662)