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


Current browse context:


Change to browse by:

References & Citations


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

Mathematics > Numerical Analysis

Title: Sampling-based Nyström Approximation and Kernel Quadrature

Abstract: We analyze the Nystr\"om approximation of a positive definite kernel associated with a probability measure. We first prove an improved error bound for the conventional Nystr\"om approximation with i.i.d. sampling and singular-value decomposition in the continuous regime; the proof techniques are borrowed from statistical learning theory. We further introduce a refined selection of subspaces in Nystr\"om approximation with theoretical guarantees that is applicable to non-i.i.d. landmark points. Finally, we discuss their application to convex kernel quadrature and give novel theoretical guarantees as well as numerical observations.
Comments: 27 pages
Subjects: Numerical Analysis (math.NA); Machine Learning (cs.LG); Machine Learning (stat.ML)
Cite as: arXiv:2301.09517 [math.NA]
  (or arXiv:2301.09517v1 [math.NA] for this version)

Submission history

From: Satoshi Hayakawa [view email]
[v1] Mon, 23 Jan 2023 16:05:56 GMT (88kb,D)

Link back to: arXiv, form interface, contact.