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

Download:

Current browse context:

math.ST

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 > Statistics Theory

Title: Variational limits of k-NN graph based functionals on data clouds

Abstract: We consider i.i.d. samples $x_1, \dots, x_n$ from a measure $\nu$ with density supported on a bounded Euclidean domain $D \subseteq R^d $ where $d\geq3$. A graph on the point cloud is obtained by connecting two points if one of them is among the $k$-nearest neighbors of the other. Our goal is to study consistency of graph based procedures to clustering, classification and dimensionality reduction by studying the variational convergence of the graph total variation associated to such $k$-NN graph. We prove that provided $k:=k_n$ scales like $n \gg k_n \gg \log(n)$, then the $\Gamma$-convergence of the graph total variation towards an appropriate weighted total variation is guaranteed.
Subjects: Statistics Theory (math.ST); Analysis of PDEs (math.AP); Optimization and Control (math.OC); Machine Learning (stat.ML)
Cite as: arXiv:1607.00696 [math.ST]
  (or arXiv:1607.00696v1 [math.ST] for this version)

Submission history

From: Nicolas Garcia Trillos [view email]
[v1] Sun, 3 Jul 2016 22:51:48 GMT (22kb)
[v2] Mon, 20 Feb 2017 16:18:44 GMT (22kb)
[v3] Mon, 21 May 2018 20:45:05 GMT (554kb,D)

Link back to: arXiv, form interface, contact.