partition {longitudinalData}

~ Function: partition ~

Description

partition is the constructor of the class Partition. It can be build either alone or relatively to a object LongData.

Usage

partition(clusters, traj, details=character())

Arguments

clusters

[vector(factor)]: cluters to which each individual belongs. Each clusters is represented by an upper letters.

traj

[matrix] or [array]: if an object LongData is provide, it will be used to compute the quality criterion of the clustering. array are simply turn into matrix by "sticking" all the variables one behind the other.

details

[vector(character)]: the slot details is used to store various informations. If the Partition has been find using an algorithm, it can store the name of the algorithm, the time before convergence, the number of iteration and any other informations. The syntaxe is details=c(algoritm="kmeans",convergenceTime="6",otherInfo="WhateverIWant").

Details

partition construct a object of class Partition. It does not provide any default values. yLongData and details are optional.

Value

An object of class Partition.

Author

Christophe Genolini
1. UMR U1027, INSERM, Université Paul Sabatier / Toulouse III / France
2. CeRSM, EA 2931, UFR STAPS, Université de Paris Ouest-Nanterre-La Défense / Nanterre / France

References

[1] Christophe M. Genolini and Bruno Falissard
"KmL: k-means for longitudinal data"
Computational Statistics, vol 25(2), pp 317-328, 2010

[2] Christophe M. Genolini and Bruno Falissard
"KmL: A package to cluster longitudinal data"
Computer Methods and Programs in Biomedicine, 104, pp e112-121, 2011

See Also

Partition,ordered

Examples

### Empty partition
partition()

### Small partition
partition(clusters=c("A","B","A","C","C"))

### Random partition
partition(clusters=LETTERS[floor(runif(100,1,5))])

### Partition that clusters correctly some data
###   Quality criterion are high
data(artificialLongData)
traj <- as.matrix(artificialLongData[,-1])
partition(clusters=rep(1:4,each=50),traj)

### Partition that does not cluster correctly the data
###   Quality criterion are low
partition(clusters=rep(1:4,50),traj)