Régression ridge a noyau dans un graphe : Application au réseau de distribution d'eau potable

Abstract : The aim of this paper is to present the methodology of statistical inference and prediction for processes defined on network graphs when those processes are only observed on a small number of nodes. The kernel regression approach presented here already raise the question of eigenvectors computation of a very large adjacency matrix.
Document type :
Preprints, Working Papers, ...
Liste complète des métadonnées

Cited literature [8 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01515313
Contributor : Christophe Dumora <>
Submitted on : Tuesday, May 9, 2017 - 11:43:54 AM
Last modification on : Thursday, January 18, 2018 - 3:26:01 PM
Document(s) archivé(s) le : Thursday, August 10, 2017 - 12:16:54 PM

File

JDS_2017_Regression_ridge_a_no...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01515313, version 1

Collections

Citation

Christophe Dumora, Jérémie Bigot, Vincent Couallier, David Auber, Cyril Leclerc. Régression ridge a noyau dans un graphe : Application au réseau de distribution d'eau potable. 2017. ⟨hal-01515313⟩

Share

Metrics

Record views

223

Files downloads

116