Bandwidth selection in clustering with errors in variables

Abstract : We consider the problem of clustering when we observe a corrupted sample. We test two bandwidth selection methods. It allows to deal with the isotropic and the anisotropic problem of selecting the bandwidth of a deconvolution kernel. These methods are based on Lepski's type procedure. The first method compares empirical risks associated with different bandwidths by using ICI (Intersection of Confidence Intervals) rule whereas the second one computes the gradient of the empirical risk and allows us to construct an anisotropic data-driven bandwidth. Numerical experiments are proposed to illustrate the efficiency of the methods.
Liste complète des métadonnées

Cited literature [21 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01060476
Contributor : Sébastien Loustau <>
Submitted on : Thursday, November 6, 2014 - 4:50:48 PM
Last modification on : Wednesday, December 19, 2018 - 2:08:04 PM
Document(s) archivé(s) le : Saturday, February 7, 2015 - 10:05:19 AM

File

bandwidthselection.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01060476, version 1

Collections

Citation

Sébastien Loustau, Simon Souchet. Bandwidth selection in clustering with errors in variables. 2014. 〈hal-01060476〉

Share

Metrics

Record views

201

Files downloads

106