kml3d (version 0.7)

calculCriterion: ~ Function: calculCriterion ~

Description

Given a LongData and a Partition, the fonction calculCriterion calcul some qualities criterion.

Usage

calculCriterion(traj, part, imputationMethod = "LI-Bissectrice", criterionNames = c("calinski","ray","davies","random"))

Arguments

traj

[array]: object containning the trajectories.

part

[vector(character)]: partition of the trajectories.

imputationMethod

[character]: if some value are missing in the LongData, it is necessary to impute them. The function calculCriterion call the function imputation using

criterionNames

[vector(character)]: names of the criterions that should be calculate.

Value

• A list:
  • calinski
  {[numeric]: Calinski and Harabatz criterion c(k)=Trace(B)/Trace(W)*(n-k)/(k-1)}
• davies
  [numeric]: Davies and Bouldin criterion (lowest is the best)
• ray
  [numeric]: Ray and Turing (lowest is the best)
• random
  code{[numeric]}: random value random=rnorm()

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

Given a LongData and a Partition, the fonction calculCriterion calculate some qualities criterion. Available criterions include : "calinski","ray","davies" and "random".

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

See Also

LongData, Partition, imputation.

Examples

##################
### Preparation of some artificial data
myCld <- gald()

### Correct partition
part1 <- partition(rep(1:3,each=50),3)
(cr1 <- calculCriterion(myCld,part1))
plot(myCld,part1)


### Random partition
part2 <- partition(floor(runif(150,1,4)),3)
(cr2 <- calculCriterion(myCld,part2))
plot(myCld,part2)