kml (version 2.1.2)

plotCriterion: ~ Function: plotCriterion ~

Description

This function graphicaly display the quality criterion of all the Clusterization of a ClusterizLongData object.

Usage

plotCriterion(x, nbCriterion = 100, allCrit = FALSE)

Arguments

x

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

nbCriterion

[numeric]: if there is a big number of Clusterization, the graphical display of all of them can be slow. nbCriterion let the user limit the number of criterion that will be take in account.

allCrit

[logical]: if TRUE, up to nbCriterion for each possible clusters number are considered. If FALSE, only the best criterion for each possible clusters number are displayed.

Value

• No value are return. A graph is printed.

Details

This function display graphicaly the quality criterion (probably to decide the best clusters' number). It can either display all the criterion ; this is usefull to see the consistancy of the result : is the best clusterization obtain several time or only one ? It can also display only the best result for each clusters number : this help to find the local maximum, which is classicaly used to chose the "correct" clusters' number.

Examples

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

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

### Display the quality criterion, both way :
par(mfrow=c(1,2))
plotCriterion(dn)
plotCriterion(dn,allCrit=TRUE)