Parameter-Wise Co-Clustering for ...
Document type :
Compte-rendu et recension critique d'ouvrage
Title :
Parameter-Wise Co-Clustering for High-Dimensional Data
Author(s) :
Gallaugher, Michael [Auteur]
Baylor University
Biernacki, Christophe [Auteur]
MOdel for Data Analysis and Learning [MODAL]
Mcnicholas, Paul [Auteur]
McMaster University [Hamilton, Ontario]
Baylor University
Biernacki, Christophe [Auteur]
MOdel for Data Analysis and Learning [MODAL]
Mcnicholas, Paul [Auteur]
McMaster University [Hamilton, Ontario]
Journal title :
Computational Statistics
Publisher :
Springer Verlag
Publication date :
2022-10
ISSN :
0943-4062
HAL domain(s) :
Statistiques [stat]
Statistiques [stat]/Méthodologie [stat.ME]
Statistiques [stat]/Méthodologie [stat.ME]
English abstract : [en]
In recent years, data dimensionality has increasingly become a concern, leading to many parameter and dimension reduction techniques being proposed in the literature. A parameter-wise co-clustering model, for (possibly ...
Show more >In recent years, data dimensionality has increasingly become a concern, leading to many parameter and dimension reduction techniques being proposed in the literature. A parameter-wise co-clustering model, for (possibly high-dimensional) data modelled via continuous random variables, is presented. The proposed model, although allowing more flexibility, still maintains the very high degree of parsimony and interpretability achieved by traditional co-clustering. More precisely, the keystone consists of dramatically increasing the number of column-clusters while expressing each as a combination of a limited number of mean-dependent and variance-dependent column-clusters. A stochastic expectation-maximization (SEM) algorithm along with a Gibbs sampler is used for parameter estimation and an integrated complete log-likelihood criterion is used for model selection. Simulated and real datasets are used for illustration and comparison with traditional co-clustering.Show less >
Show more >In recent years, data dimensionality has increasingly become a concern, leading to many parameter and dimension reduction techniques being proposed in the literature. A parameter-wise co-clustering model, for (possibly high-dimensional) data modelled via continuous random variables, is presented. The proposed model, although allowing more flexibility, still maintains the very high degree of parsimony and interpretability achieved by traditional co-clustering. More precisely, the keystone consists of dramatically increasing the number of column-clusters while expressing each as a combination of a limited number of mean-dependent and variance-dependent column-clusters. A stochastic expectation-maximization (SEM) algorithm along with a Gibbs sampler is used for parameter estimation and an integrated complete log-likelihood criterion is used for model selection. Simulated and real datasets are used for illustration and comparison with traditional co-clustering.Show less >
Language :
Anglais
Popular science :
Non
Collections :
Source :
Files
- document
- Open access
- Access the document
- PWCoClust_Revised2.pdf
- Open access
- Access the document