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

Statistics > Methodology

Title: Finding Singular Features

Abstract: We present a method for finding high density, low-dimensional structures in noisy point clouds. These structures are sets with zero Lebesgue measure with respect to the $D$-dimensional ambient space and belong to a $d<D$ dimensional space. We call them "singular features." Hunting for singular features corresponds to finding unexpected or unknown structures hidden in point clouds belonging to $\R^D$. Our method outputs well defined sets of dimensions $d<D$. Unlike spectral clustering, the method works well in the presence of noise. We show how to find singular features by first finding ridges in the estimated density, followed by a filtering step based on the eigenvalues of the Hessian of the density.
Subjects: Methodology (stat.ME); Machine Learning (stat.ML)
Cite as: arXiv:1606.00265 [stat.ME]
  (or arXiv:1606.00265v1 [stat.ME] for this version)

Submission history

From: Larry Wasserman [view email]
[v1] Wed, 1 Jun 2016 12:50:12 GMT (1967kb,D)

Link back to: arXiv, form interface, contact.