We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math.RA

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Rings and Algebras

Title: A measure concentration effect for matrices of high, higher, and even higher dimension

Abstract: Let $n>m$, and let $A$ be an $(m\times n)$-matrix of full rank. Then obviously the estimate $\|Ax\|\leq\|A\|\|x\|$ holds for the euclidean norm of $x$ and $Ax$ and the spectral norm as the assigned matrix norm. We study the sets of all $x$ for which, for fixed $\delta<1$, conversely $\|Ax\|\geq\delta\,\|A\|\|x\|$ holds. It turns out that these sets fill, in the high-dimensional case, almost the complete space once $\delta$ falls below a bound that depends on the extremal singular values of $A$ and on the ratio of the dimensions. This effect has much to do with the random projection theorem, which plays an important role in the data sciences. As a byproduct, we calculate the probabilities this theorem deals with exactly.
Comments: to appear in SIAM Journal on Matrix Analysis and Applications
Subjects: Rings and Algebras (math.RA); Numerical Analysis (math.NA)
MSC classes: 15A99, 15B52, 60B20
Journal reference: SIAM J. Matrix Anal. Appl. 43 (2022), pp.464-478
DOI: 10.1137/20M1376029
Cite as: arXiv:2010.13505 [math.RA]
  (or arXiv:2010.13505v4 [math.RA] for this version)

Submission history

From: Harry Yserentant [view email]
[v1] Mon, 26 Oct 2020 12:03:00 GMT (120kb,D)
[v2] Wed, 11 Nov 2020 09:24:57 GMT (255kb,D)
[v3] Mon, 1 Mar 2021 13:59:15 GMT (131kb,D)
[v4] Sun, 20 Feb 2022 07:53:27 GMT (171kb,D)

Link back to: arXiv, form interface, contact.