Skip to Main content Skip to Navigation
Conference papers

Rank-based multiple change-point detection in multivariate time series

Flore Harlé 1, 2 Florent Chatelain 3 Cedric Gouy-Pailler 2 Sophie Achard 4
1 GIPSA-VIBS - GIPSA - Vision and Brain Signal Processing
GIPSA-DIS - Département Images et Signal
2 LADIS - Laboratoire d'analyse des données et d'intelligence des systèmes
DM2I - Département Métrologie Instrumentation & Information : DRT/LIST/DM2I
3 GIPSA-SAIGA - GIPSA - Signal et Automatique pour la surveillance, le diagnostic et la biomécanique
GIPSA-DA - Département Automatique, GIPSA-DIS - Département Images et Signal
Abstract : In this paper, we propose a Bayesian approach for multivariate time series segmentation. A robust non-parametric test, based on rank statistics, is derived in a Bayesian framework to yield robust distribution-independent segmentations of piecewise constant multivariate time series for which mutual dependencies are unknown. By modelling rank-test p-values, a pseudo-likelihood is proposed to favour change-points detection for significant p-values. A vague prior is chosen for dependency structure between time series, and a MCMC method is applied to the resulting posterior distribution. The Gibbs sampling strategy makes the method computationally efficient. The algorithm is illustrated on simulated and real signals in two practical settings. It is demonstrated that change-points are robustly detected and localized, through implicit dependency structure learning or explicit structural prior introduction.
Complete list of metadatas

https://hal-cea.archives-ouvertes.fr/cea-01844442
Contributor : Léna Le Roy <>
Submitted on : Thursday, July 19, 2018 - 1:31:08 PM
Last modification on : Thursday, November 19, 2020 - 1:01:18 PM

Identifiers

  • HAL Id : cea-01844442, version 1

Citation

Flore Harlé, Florent Chatelain, Cedric Gouy-Pailler, Sophie Achard. Rank-based multiple change-point detection in multivariate time series. 22nd European Signal Processing Conference, EUSIPCO 2014, Sep 2014, Lisbonne, Portugal. pp.1337-1341. ⟨cea-01844442⟩

Share

Metrics

Record views

244