 |
Clustering-based models in R, with application on univariate Gaussian mixtures: review, evaluation and extensions
1 : Laboratoire de Probabilités, Statistiques et Modélisations
Sorbonne Université : UMR_8001, Centre National de la Recherche Scientifique : UMR_8001, Université de Paris : UMR_8001
2 : Mathématiques appliquées Paris 5
(MAP5)
-
Site web
CNRS : UMR8145, Université Paris Descartes
UFR de Maths et informatique 45 rue des Saints Pères 75270 PARIS CEDEX 06 -
France
3 : Laboratoire dÍnformatique de Paris 6
(LIP6)
-
Site web
Université Pierre et Marie Curie - Paris 6, Centre National de la Recherche Scientifique : UMR7606
4 Place JUSSIEU 75252 PARIS CEDEX 05 -
France
4 : Institut de Recherches Internationales Servier
* : Auteur correspondant
Institut de Recherches Internationales Servier [Suresnes]
Suresnes -
France
Finite mixture models are increasingly used for modeling and dealing with stochastic problems such as
clustering, classification and regression, with many applications in biological fields. Unsurprisingly, many
packages have been developed to fit mixture models, arising the natural question which is the best suited one,
depending on the use case.
However, to our knowledge, no review describing the main features offered by these packages and comparing
their computational and statistical performances has been performed. In this talk, we focus on packages
implementing the EM algorithm on univariate Gaussian mixture distributions, being the most common use case.
- Poster
