Skip to Main content Skip to Navigation
Conference papers

Analyse en composantes principales partielle de données séquentielles d'espérance et de matrice de covariance variables dans le temps

Romain Bar 1, 2 Jean-Marie Monnez 1, 2
1 BIGS - Biology, genetics and statistics
IECN - Institut Élie Cartan de Nancy, INRIA Lorraine
Abstract : High dimensional batch data are supposed to be independent observations of a random vector Z, expectation and covariance matrix of which vary with time n. A recursive method of on-line estimation of direction vectors of the r first principal axes of a partial principal components analysis (PCA) of Z is defined. This is applied next to the particular case of a partial generalized canonical correlation analysis (gCCA) after defining a stochastic approximation process of the Robbins-Monro type to estimate recursively the inverse of a covariance matrix.
Complete list of metadata

Cited literature [5 references]  Display  Hide  Download

https://hal.inria.fr/hal-00841181
Contributor : Romain Bar <>
Submitted on : Thursday, July 4, 2013 - 9:54:47 AM
Last modification on : Tuesday, March 2, 2021 - 5:12:05 PM
Long-term archiving on: : Saturday, October 5, 2013 - 4:16:48 AM

File

resume_sfds_2013.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00841181, version 1

Collections

Citation

Romain Bar, Jean-Marie Monnez. Analyse en composantes principales partielle de données séquentielles d'espérance et de matrice de covariance variables dans le temps. 45èmes Journées de Statistiques - 2013, May 2013, Toulouse, France. ⟨hal-00841181⟩

Share

Metrics

Record views

602

Files downloads

531