Title :

Parameter Setting for Evolutionary Latent Class Clustering

Author(s) :

Tessier, Damien [Auteur]
Machine Learning and Optimisation [TAO]
INRIA Futurs
Schoenauer, Marc [Auteur]
Machine Learning and Optimisation [TAO]
INRIA Futurs
Biernacki, Christophe [Auteur]
Laboratoire Paul Painlevé - UMR 8524 [LPP]
Celeux, Gilles [Auteur]
Model selection in statistical learning [SELECT]
Govaert, Gérard [Auteur]
Heuristique et Diagnostic des Systèmes Complexes [Compiègne] [Heudiasyc]

Scientific editor(s) :

Lishan Kang and Yong Liu and Sanyou Y. Zeng

Conference title :

Second International Symposium, ISICA 2007

City :

Wuhan

Country :

Chine

Start date of the conference :

2007-09-21

Journal title :

LNCS

Publisher :

Springer Verlag

Publication date :

2007

HAL domain(s) :

Informatique [cs]/Intelligence artificielle [cs.AI]
Mathématiques [math]/Statistiques [math.ST]

English abstract : [en]

The latent class model or multivariate multinomial mixture is a powerful model for clustering discrete data. This model is expected to be useful to represent non-homogeneous populations. It uses a conditional independence ...
Show more >The latent class model or multivariate multinomial mixture is a powerful model for clustering discrete data. This model is expected to be useful to represent non-homogeneous populations. It uses a conditional independence assumption given the latent class to which a statistical unit is belonging. However, it leads to a criterion that proves difficult to optimise by the standard approach based on the EM algorithm. An Evolutionary Algorithms is designed to tackle this discrete optimisation problem, and an extensive parameter study on a large artificial dataset allows to derive stable parameters. Those parameters are then validated on other artificial datasets, as well as on some well-known real data: the Evolutionary Algorithm performs repeatedly better than other standard clustering techniques on the same data.Show less >

Language :

Anglais

Peer reviewed article :

Oui

Audience :

Internationale

Popular science :

Non

Collections :

• Laboratoire Paul Painlevé - UMR 8524

Source :

Harvested from HAL