Model-Based Co-clustering for Ordinal Data
Type de document :
Article dans une revue scientifique
URL permanente :
Titre :
Model-Based Co-clustering for Ordinal Data
Auteur(s) :
Titre de la revue :
Computational Statistics and Data Analysis
Numéro :
123
Pagination :
101-115
Éditeur :
Elsevier
Date de publication :
2018-07
ISSN :
0167-9473
Mot(s)-clé(s) :
Ordinal data
Co-clustering
Gibbs sampler
EM algorithm
Latent block model
SEM-Gibbs algorithm
Co-clustering
Gibbs sampler
EM algorithm
Latent block model
SEM-Gibbs algorithm
Discipline(s) HAL :
Mathématiques [math]/Statistiques [math.ST]
Résumé en anglais : [en]
A model-based co-clustering algorithm for ordinal data is presented. This algorithm relies on the latent block model embedding a probability distribution specific to ordinal data (the so-called BOS or Binary Ordinal Search ...
Lire la suite >A model-based co-clustering algorithm for ordinal data is presented. This algorithm relies on the latent block model embedding a probability distribution specific to ordinal data (the so-called BOS or Binary Ordinal Search distribution). Model inference relies on a Stochastic EM algorithm coupled with a Gibbs sampler, and the ICL-BIC criterion is used for selecting the number of co-clusters (or blocks). The main advantage of this ordinal dedicated co-clustering model is its parsimony, the interpretability of the co-cluster parameters (mode, precision) and the possibility to take into account missing data. Numerical experiments on simulated data show the efficiency of the inference strategy, and real data analyses illustrate the interest of the proposed procedure.Lire moins >
Lire la suite >A model-based co-clustering algorithm for ordinal data is presented. This algorithm relies on the latent block model embedding a probability distribution specific to ordinal data (the so-called BOS or Binary Ordinal Search distribution). Model inference relies on a Stochastic EM algorithm coupled with a Gibbs sampler, and the ICL-BIC criterion is used for selecting the number of co-clusters (or blocks). The main advantage of this ordinal dedicated co-clustering model is its parsimony, the interpretability of the co-cluster parameters (mode, precision) and the possibility to take into account missing data. Numerical experiments on simulated data show the efficiency of the inference strategy, and real data analyses illustrate the interest of the proposed procedure.Lire moins >
Langue :
Anglais
Audience :
Internationale
Vulgarisation :
Non
Date de dépôt :
2020-06-08T14:11:38Z
2020-06-09T09:29:59Z
2020-06-09T09:29:59Z
Fichiers
- documen
- Accès libre
- Accéder au document