clustercat: clustercat function

In Clustericat: This package performs clustering for categorical dataset.

Description

This function performs clustering for categorical data using the conditional correlated mixture model.

Usage

  clustercat(data, nb_cluster,modal=0,
    strategy = strategycat(data))

Arguments

data	Input data as matrix of non-zero integers.
nb_cluster	Integer vector specifying the number of classes.
modal	Vector of modalities. If modal=0, then the modalities number of a variable is equal to the number of the different observed levels.
strategy	An instance of the strategycat class which contains the adjustments parameters.

Value

Return an instance of the clustcat class. Those two attributes will contains all outputs:

References

Marbac M., Biernacki C., Vandewalle V., 2014. "Model-based clustering for conditionally correlated categorical data". Rapport de recherche INRIA RR-8232.

Examples

# Simple example with binary data
data("dentist")

# to define the parameters of the algorithm performing the estimation
st=strategycat(dentist,nb_init=35,stop_criterion=200)


# estimation of the model for a classes number equal to 1,2.
res <- clustercat(dentist, 1:2,modal=rep(2,5), strategy=st)

# presentation of the best model
summary(res)

# presentation of the parameters of the conditional dependencies for the best model
summary_dependencies(res)

# a plot summarizing the best best model
plot(res)

Clustericat documentation built on May 2, 2019, 5:45 p.m.