1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
|
\name{medoids}
\alias{medoids}
\title{Compute \code{pam}-consistent Medoids from Clustering}
\description{
Given a data matrix or dissimilarity \code{x} for say \eqn{n}
observational units and a clustering,
compute the \code{\link{pam}()}-consistent medoids.
}
\usage{
medoids(x, clustering, diss = inherits(x, "dist"), USE.NAMES = FALSE, ...)
}
\arguments{
\item{x}{Either a data matrix or data frame, or dissimilarity matrix or
object, see also \code{\link{pam}}.}
\item{clustering}{an integer vector of length \eqn{n}, the number of
observations, giving for each observation the number ('id') of the
cluster to which it belongs. In other words, \code{clustering} has
values from \code{1:k} where \code{k} is the number of clusters, see
also \code{\link{partition.object}} and \code{\link{cutree}()}, for
examples where such clustering vectors are computed.}
\item{diss}{see also \code{\link{pam}}.}
\item{USE.NAMES}{a logical, typical false, passed to the
\code{\link{vapply}()} call computing the medoids.}
\item{\dots}{optional further argument passed to \code{\link{pam}(xj, k=1, \dots)},
notably \code{metric}, or \code{variant="f_5"} to use a faster algorithm, or
\code{trace.lev = k}.}
}
%% \details{
%% }
\value{
a numeric vector of length
}
%% \references{
%% }
\author{Martin Maechler, after being asked how \code{\link{pam}()} could be used
instead of \code{\link{kmeans}()}, starting from a previous clustering.
}
%% \note{
%% }
\seealso{
\code{\link{pam}}, \code{\link{kmeans}}.
Further, \code{\link{cutree}()} and \code{\link{agnes}} (or \code{\link{hclust}}).
}
\examples{
## From example(agnes):
data(votes.repub)
agn1 <- agnes(votes.repub, metric = "manhattan", stand = TRUE)
agn2 <- agnes(daisy(votes.repub), diss = TRUE, method = "complete")
agnS <- agnes(votes.repub, method = "flexible", par.method = 0.625)
for(k in 2:11) {
print(table(cl.k <- cutree(agnS, k=k)))
stopifnot(length(cl.k) == nrow(votes.repub), 1 <= cl.k, cl.k <= k, table(cl.k) >= 2)
m.k <- medoids(votes.repub, cl.k)
cat("k =", k,"; sort(medoids) = "); dput(sort(m.k), control={})
}
}
\keyword{cluster}
|
