Données fonctionnelles multivariées issues d'objets connectés : une méthode pour classer les individus

Abstract : The emergence of numerical sensors in many aspects of everyday life leads to an increasing need of methods to analyze multivariate functional data. This work presents a clustering technique (Schmutz et al, 2017) in order to ease the modeling and understanding of those multivariate functional data. This method is based on a functional latent mixture model which fits the data in group-specific functional subspaces through a multivariate functional principal component analysis. An EM-like algorithm is proposed for model inference and the choice of hyper-parameters is carried out through model selection. Model eficiency will be tested on an original example of speed prediction, for classical example see Schmutz et al (2017).
Document type :
Conference papers
Complete list of metadatas

Cited literature [13 references]  Display  Hide  Download

https://hal.inria.fr/hal-01784279
Contributor : Amandine Schmutz <>
Submitted on : Thursday, May 3, 2018 - 11:36:57 AM
Last modification on : Thursday, April 11, 2019 - 4:02:09 PM
Long-term archiving on : Tuesday, September 25, 2018 - 9:23:35 AM

File

gabarit_text_long_v3.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01784279, version 1

Citation

Amandine Schmutz, Julien Jacques, Charles Bouveyron, Laurence Cheze, Pauline Martin. Données fonctionnelles multivariées issues d'objets connectés : une méthode pour classer les individus. Journées des Statistiques, May 2018, Saclay, France. ⟨hal-01784279⟩

Share

Metrics

Record views

250

Files downloads

181