kml3d (version 0.7)

calculTrajMean: ~ Function: calculTrajMean ~

Description

Given some longitudinal data and a Partition, calculTrajMean computes the mean trajectories of each cluster.

Usage

calculTrajMean(traj, part,centerMethod=function(x){mean(x,na.rm=TRUE)})

Arguments

traj

[array]: longitudinal data.

part

[vector(character)]: affectation of each individual.

centerMethod

[numeric <- function(vector)]: function used to compute the clusters' centers.

Value

• An array of dimention (k,t,v) with k number of groups, t number of time mesurement and v number of variables.

Author(s)

Christophe Genolini INSERM U669 / PSIGIAM: Paris Sud Innovation Group in Adolescent Mental Health Modal'X / Universite Paris Ouest-Nanterre- La Defense Contact author : [e-mail]

Details

EM algorithm (like k-means) alternates between two phases : Esperance and Maximisation. During Esperance, the center of each cluster is evaluated. This is what calculTrajMean does. Note that calculTrajMean does not work with ClusterLongData object but with an array of trajectories. affectIndiv used with calculTrajMean simulates one k-means step.

References

Article "KmL: K-means for Longitudinal Data", in Computational Statistics, Volume 25, Issue 2 (2010), Page 317. Web site: http://christophe.genolini.free.fr/kml

Examples

#######################
### calculTrajMean

### Some LongitudinalData
traj <- gald()["traj"]

### A partition
part <- partition(floor(runif(150,1,5)),3)
plot(as.longData(traj),part)

### Clusters center
(center <- calculTrajMean(traj,part["clusters"]))

### Unusual center
calculTrajMean(traj,part["clusters"],centerMethod=function(x)median(x,na.rm=TRUE))


#################
### K-means simulation (4 steps)
plot(as.longData(traj),part)
for (i in 1:4){
    part <- affectIndiv(traj,center)
    center <- calculTrajMean(traj,part["clusters"])
    plot(as.longData(traj),part)
}