On selecting the hyperparameters of the DPM models for the density estimation of observation errors

Asma Rabaoui 1 Emmanuel Duflos 2, 3, * Juliette Marais 4 Nicolas Viandier 4
* Corresponding author
2 SEQUEL - Sequential Learning
LIFL - Laboratoire d'Informatique Fondamentale de Lille, LAGIS - Laboratoire d'Automatique, Génie Informatique et Signal, Inria Lille - Nord Europe
3 LAGIS-SI
LAGIS - Laboratoire d'Automatique, Génie Informatique et Signal
Abstract : The Dirichlet Process Mixture (DPM) models represent an attractive approach to modeling latent distributions parametrically. In DPM models the Dirichlet process (DP) is applied especially when the distribution of latent parameters is to be considered as multimodal. DPMs allow for uncertainty in the choice of parametric forms and in the number of mixing components (clusters). The parameters of a DP include the precision a and the base probability measure G0(μ, Σ). In most applications, the choice of priors and posteriors computation for the hyperparameters (α, μ, Σ) clearly influences inferences about the level of clustering in the mixture. This is the main focus of this paper. We consider the problem of density estimation of an observation noise distribution in a dynamic nonlinear model from a Bayesian nonparametric view-point. Our approach is illustrated in a real-world data analysis task dealing with the estimation of pseudorange errors in a GNSS based localization context.
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-00712723
Contributor : Emmanuel Duflos <>
Submitted on : Wednesday, June 27, 2012 - 10:17:27 PM
Last modification on : Thursday, February 21, 2019 - 11:02:54 AM

Identifiers

Citation

Asma Rabaoui, Emmanuel Duflos, Juliette Marais, Nicolas Viandier. On selecting the hyperparameters of the DPM models for the density estimation of observation errors. International Conference on Acoustic, Speech and Signal Processing (ICASSP°, May 2011, Prague, Czech Republic. pp.4092-4095, ⟨10.1109/ICASSP.2011.5947252⟩. ⟨hal-00712723⟩

Share

Metrics

Record views

767