English

On the approximation of a matrix

Numerical Analysis 2021-08-31 v1 Machine Learning Numerical Analysis Machine Learning

Abstract

Let FF^{*} be an approximation of a given (a×b)(a \times b) matrix FF derived by methods that are not randomized. We prove that for a given FF and FF^{*}, HH and TT can be computed by randomized algorithm such that (HT)(HT) is an approximation of FF better than FF^{*}.

Keywords

Cite

@article{arxiv.2108.13195,
  title  = {On the approximation of a matrix},
  author = {Samriddha Sanyal},
  journal= {arXiv preprint arXiv:2108.13195},
  year   = {2021}
}
R2 v1 2026-06-24T05:31:38.116Z