[Forum SIS] Short Course "Model-based Clustering" Prof. Cristina Tortora

[Nai Ruscone Marta]

Apologies for cross-posting

********************************************************************************
AVVISO DI SEMINARIO
********************************************************************************
La Scuola di Economia e Management della LIUC - Università Carlo Cattaneo di Castellanza ha organizzato il mini-corso:

Model-based Clustering
Cristina Tortora
Department of Mathematics and Statistics
San Jose State University, USA

Giovedì 10 e venerdì 11 Gennaio 2018, dalle ore 9.00 alle ore 12.00
Aula CRI3
Corso Matteotti 22, Castellanza (Varese).

Abstract:
Cluster analysis is a multivariate statistical technique that identifies homogeneous groups of units within the data. One of the most commonly used technique is model based clustering.  Model based clustering assumes that a population is a mixture of sub-population, each component is modeled through a probability density function and a component can be considered a cluster. The scope of this short course is to introduce model based clustering on continuous data. The course will start with the introduction of the most traditional model, Gaussian mixture models, with some details on the algorithm. Some flexible techniques, based on non-Gaussian distributions, will then be introduced. The last topic covered will be the recently proposed mixture of contaminated normal distributions, that has the advantage of simultaneously clustering the data and detecting outliers. The course will also contains a brief tutorial on some R packages for the presented models.


  *   Introduction



  *   Cluster analysis
  *   Model based clustering



  *   Gaussian mixture models



  *   Maximum likelihood estimates
  *   Expectation-Maximization (EM-algorithm)
  *   R package mclust



  *   Non-Gaussian mixture models



  *   Mixture of Student-t distributions
  *   R package teigen
  *   Mixture of generalized hyperbolic distributions
  *   R package MixGHD



  *   Cluster analysis and outlier detection



  *   Mixture of contaminated normal distributions
  *   Mixture of multiple scaled contaminated normal distributions
  *   R package ContaminatedMixt



I colleghi interessati sono cordialmente invitati a partecipare.
La partecipazione al mini-corso è libera, ma i posti sono limitati: gli interessati sono pregati di inviare una e-mail a ricerca a liuc.it<mailto:ricerca a liuc.it>.


*** --- ***
Marta Nai Ruscone
School of Economics and Management
Corso Matteotti 22
21053 Castellanza (VA)
Italia

Home page: http://per.liuc.it/marta.nairuscone



[http://Static.unicatt.it/layout/img/layout/5x1000.gif]
Destina il tuo 5 per mille all’Università Cattolica
CF 02133120150
www.unicatt.it/5permille<http://www.unicatt.it/5permille/>

-------------- parte successiva --------------
Un allegato HTML è stato rimosso...
URL: <http://www.stat.unipg.it/pipermail/sis/attachments/20190104/cf607485/attachment-0001.html>