Clusterization-class {kml} | R Documentation |
An object of class Clusterization
is a partition of
trajectories in subgroups.
The object also contains some information like the
percentage of trajectories that each group contains or the Calinski criterion.
Objects are not intend to be created by users. Clusterization
are created
by kml
and directly added to a "ClusterizLongData"
object.
id
nbClusters
clusters
A
to LETTERS[nbClusters]
. Please note that a group might
be empty. The groups are sort by size: (A
is the
biggest group, B
is the next,...)percentEachCluster
calinski
trace(B)/trace(W)*(g-1)/(n-g)
with
B=matrice of variance Between, W=matrice of variance Within, g=number of groups and n=number of individual.)convergenceTime
:kml
to find the Clusterization
A class Clusterization
object must follow some rules to be valid:
nbClusters
has to be lower or equal to 25 (twenty five clusters maximum).
clusters
and id
have to have the same length.
clusters
has to be a factor in LETTERS[1:nbCluster]
(with
nbClusters
lower than 25).
table(clusters)
has to be in the reverse order
(clusters are sorted form the biggest to the smallest).
id
cannot be duplicated.
id
cannot be missing.
Class Clusterization
objects are constructed through the
kml
procedure and are directy add to a
ClusterizLongData
object. They are not intend to be construct
by the users.
id
)nbClusters
)clusters
)value
.
"percentEachClusters"
, "calinski"
and
"convergenceTime"
are calculate by the creation function
according to the clusterization. They are not
suppose to be change by te user.
Christophe Genolini
PSIGIAM: Paris Sud Innovation Group in Adolescent Mental Health
INSERM U669 / Maison de Solenn / Paris
Contact author : <genolini@u-paris10.fr>
Raphaël Ricaud
Laboratoire "Sport & Culture" / "Sports & Culture" Laboratory
University of Paris 10 / Nanterre
Article submited
Web site: http://christophe.genolini.free.fr/kml
Overview: kml-package
Classes : ClusterizLongData
, Clusterization
, ArtificialLongData
Methods : kml
showClass("Clusterization")