Data-driven kernel representations for sampling with an unknown block dependence structure under correlation constraints

Abstract : The multidimensional Gaussian kernel-density estimation (G-KDE) is a powerful tool to identify the distribution of random vectors when the maximal information is a set of independent realizations. For these methods, a key issue is the choice of the kernel and the optimization of the bandwidth matrix. To optimize these kernel representations, two adaptations of the classical G-KDE are presented. First, it is proposed to add constraints on the mean and the covariance matrix in the G-KDE formalism. Secondly, it is suggested to separate in different groups the components of the random vector of interest that could reasonably be considered as independent. This block by block decomposition is carried out by looking for the maximum of a cross-validation likelihood quantity that is associated with the block formation. This leads to a tensorized version of the classical G-KDE. Finally, it is shown on a series of examples how these two adaptations can improve the nonparametric representations of the densities of random vectors, especially when the number of available realizations is relatively low compared to their dimensions.
Type de document :
Article dans une revue
Computational Statistics and Data Analysis, Elsevier, 2018
Liste complète des métadonnées

https://hal-upec-upem.archives-ouvertes.fr/hal-01794809
Contributeur : Guillaume Perrin <>
Soumis le : jeudi 17 mai 2018 - 20:33:03
Dernière modification le : samedi 2 juin 2018 - 01:08:22
Document(s) archivé(s) le : mardi 25 septembre 2018 - 17:27:53

Fichier

Data_driven_PERRIN_JMS.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

  • HAL Id : hal-01794809, version 1

Collections

Citation

Guillaume Perrin, C Soize, N. Ouhbi. Data-driven kernel representations for sampling with an unknown block dependence structure under correlation constraints. Computational Statistics and Data Analysis, Elsevier, 2018. 〈hal-01794809〉

Partager

Métriques

Consultations de la notice

25

Téléchargements de fichiers

27