kml (version 1.1.3)

plotAllCriterion: ~ Function: plotAllCriterion ~

Description

This function graphically displays three quality criterions for each possible cluster number.ll the Clusterization of a ClusterizLongData object.

Usage

plotAllCriterion(x,method="linearInterpolation")

Arguments

x

[ClusterizLongData]: object whose quality criterion should be displayed.

method

[character]: imputation methode that should be used in case of missing values. See imputation for detail.

Value

• No value are return. A graph is printed.

Author(s)

Christophe Genolini PSIGIAM: Paris Sud Innovation Group in Adolescent Mental Health INSERM U669 / Maison de Solenn / Paris Contact author : [e-mail]

English translation

Rapha�l Ricaud Laboratoire "Sport & Culture" / "Sports & Culture" Laboratory University of Paris 10 / Nanterre

Details

This function display graphically three quality criterion (probably to decide the best clusters' number). The three criterions are Calinski & Harabatz, Ray & Turi and Davies & Bouldin. It displays only the best result for each clusters number : this helps to find the local maximum, which is classically used to chose the "correct" clusters' number.

References

Article "KmL: K-means for Longitudinal Data", in Computational Statistics (accepted on 11-11-2009) Web site: http://christophe.genolini.free.fr/kml

Examples

#################
### Data generation
dn <- as.cld(gald())

### Trying several clusters number and several starting condition
kml(dn)

### Display the quality criterion, both way :
plotAllCriterion(dn)