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

Download:

Current browse context:

stat.ML
< prev | next >
new | recent | 1709

Change to browse by:

cs
cs.CG
cs.LG
stat

References & Citations

Bookmark

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

Statistics > Machine Learning

Title: A Quasi-isometric Embedding Algorithm

Authors: David W. Dreisigmeyer
(Submitted on 6 Sep 2017 (v1), last revised 3 Nov 2017 (this version, v3))
Abstract: The Whitney embedding theorem gives an upper bound on the smallest embedding dimension of a manifold. If a data set lies on a manifold, a random projection into this reduced dimension will retain the manifold structure. Here we present an algorithm to find a projection that distorts the data as little as possible.
Subjects: Machine Learning (stat.ML); Computational Geometry (cs.CG); Machine Learning (cs.LG)
Cite as: arXiv:1709.01972 [stat.ML]
  (or arXiv:1709.01972v3 [stat.ML] for this version)

Submission history

From: David W. Dreisigmeyer [view email]
[v1] Wed, 6 Sep 2017 19:35:02 GMT (7kb)
[v2] Thu, 14 Sep 2017 17:35:59 GMT (7kb)
[v3] Fri, 3 Nov 2017 13:40:50 GMT (20kb)
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Link back to: arXiv, form interface, contact.