Uses the
tree$merge component of the output from
mclust to determine the
classification corresponding to a given number of clusters.
USAGE:
mclass(m, n.clust, aux = NULL)
REQUIRED ARGUMENTS:
m
the output of
mclust, or else an integer vector corresponding to the
tree$merge component of the output of
mclust.
n.clust
number of clusters.
OPTIONAL ARGUMENTS:
aux
output from a previous call to
mclass for a larger number of clusters.
VALUE:
classification
an integer vector in which the kth element identifies the classification
of the kth object. Each cluster is identified by its lowest-numbered object.
Singletons are numbered 0.
n.clust
number of clusters in the final classification.
mark
an integer vector accumulated as the merge tree is traversed for the purpose
of classification.
SEE ALSO:
,
.
EXAMPLES:
years <- c("1960", "1964", "1968", "1972", "1976")
votes.clust <- mclust(votes.repub[,years], method = "S", noise = T)
# plot the awe on the current graphics device to decide
# how many groups there are
plot(x = 1:length(votes.clust$awe), y = votes.clust$awe)
votes.9 <- mclass(votes.clust, 9)
votes.3 <- mclass(votes.clust, 3, votes.9)