CoModes-package: CoModes a package for clustering categorical data

In CoModes: This package performs clustering for categorical dataset by using a conditional modes mixture model

Description

CoModes is a tool for clustering categorical data. The clustering goal is achieved by a mixture model which considers the intra-class dependencies since the observed variables are grouped into conditionally independent blocks. The block distribution is parsimonious since each block follows a multinomial distribution per modes. Under this distribution, the free parameters correspond to the probabilities of the most probable levels (the modes) while the other levels (the non modes) are assumed to be equiprobable.

Details

Package:	CoModes
Type:	Package
Version:	1.0.0
Date:	2016-01-20
License:	GPL-2
LazyLoad:	yes

Author(s)

Author: Marbac Matthieu, Biernacki Christophe and Vandewalle Vincent

References

Marbac Matthieu, Biernacki Christophe and Vandewalle Vincent (2015). Latent class model with conditional dependency per modes to cluster categorical data. arXiv:1402.5103

Examples

## Not run: 
# Data Loading
data(alzheimer)
# Model selection and Parameter estimation for CMM with 2 components (8 MCMC chains are used for model selection)
results <- CoModescluster(alzheimer, 2, 8)
# Summary of the results
summary(results)
# Display the probabilities of the modes
barplot(results)

## End(Not run)

CoModes documentation built on May 2, 2019, 5:49 p.m.