hal-00476076, version 2
Component-level aggregation of probabilistic PCA mixtures using variational-Bayes
Technical Report. This report of an extended version of our ICPR'2010 paper. (2011)
Résumé : This paper proposes a technique for aggregating mixtures of probabilistic principal component analyzers, which are a powerful probabilistic generative model for coping with a high-dimensional, non linear, data set. Aggregation is carried out through Bayesian estimation with a specific prior and an original variational scheme. We demonstrate how such models may be aggregated by accessing model parameters only, rather than original data, which can be advantageous for learning from distributed data sets. Experimental results illustrate the effectiveness of the proposal.
- 1 :
- INRIA – Université de Nantes
- 2 :
- CNRS : UMR6241 – Université de Nantes – École Nationale Supérieure des Mines - Nantes
- Domaine : Informatique/Recherche d'information
- Mots-clés : variational-Bayes – Bayesian – Clustering – PCA – aggregation
- Versions disponibles : v1 (14-06-2010) v2 (21-02-2011)
- hal-00476076, version 2
- http://hal.archives-ouvertes.fr/hal-00476076
- oai:hal.archives-ouvertes.fr:hal-00476076
- Contributeur :
- Soumis le : Dimanche 20 Février 2011, 12:23:56
- Dernière modification le : Lundi 21 Février 2011, 13:39:11




Documents associés
Exporter