English

A Quadratically Convergent Algorithm for Structured Low-Rank Approximation

Numerical Analysis 2014-10-28 v2 Symbolic Computation Numerical Analysis

Abstract

Structured Low-Rank Approximation is a problem arising in a wide range of applications in Numerical Analysis and Engineering Sciences. Given an input matrix MM, the goal is to compute a matrix MM' of given rank rr in a linear or affine subspace EE of matrices (usually encoding a specific structure) such that the Frobenius distance MM\lVert M-M'\rVert is small. We propose a Newton-like iteration for solving this problem, whose main feature is that it converges locally quadratically to such a matrix under mild transversality assumptions between the manifold of matrices of rank rr and the linear/affine subspace EE. We also show that the distance between the limit of the iteration and the optimal solution of the problem is quadratic in the distance between the input matrix and the manifold of rank rr matrices in EE. To illustrate the applicability of this algorithm, we propose a Maple implementation and give experimental results for several applicative problems that can be modeled by Structured Low-Rank Approximation: univariate approximate GCDs (Sylvester matrices), low-rank Matrix completion (coordinate spaces) and denoising procedures (Hankel matrices). Experimental results give evidence that this all-purpose algorithm is competitive with state-of-the-art numerical methods dedicated to these problems.

Keywords

Cite

@article{arxiv.1312.7279,
  title  = {A Quadratically Convergent Algorithm for Structured Low-Rank Approximation},
  author = {Éric Schost and Pierre-Jean Spaenlehauer},
  journal= {arXiv preprint arXiv:1312.7279},
  year   = {2014}
}

Comments

37 pages, Maple package available at http://www.pjspaenlehauer.net/data/software/NewtonSLRA_notes.html

R2 v1 2026-06-22T02:35:45.423Z