index.S: Calculates Rousseeuw's Silhouette internal cluster quality...

View source: R/index.S.r

index.SR Documentation

Calculates Rousseeuw's Silhouette internal cluster quality index

Description

Calculates Rousseeuw's Silhouette internal cluster quality index

Usage

index.S(d,cl,singleObject=0)

Arguments

d

'dist' object

cl

A vector of integers indicating the cluster to which each object is allocated

singleObject

0 - s(i)=0 or 1 - s(i)=1. When cluster contains a single object, it is unclear how a(i) of Silhouette index should be defined (see Kaufman & Rousseeuw (1990), p. 85).

Details

See file $R_HOME\library\clusterSim\pdf\indexS_details.pdf for further details

Value

calculated Silhouette index

Author(s)

Marek Walesiak marek.walesiak@ue.wroc.pl, Andrzej Dudek andrzej.dudek@ue.wroc.pl

Department of Econometrics and Computer Science, University of Economics, Wroclaw, Poland

References

Gatnar, E., Walesiak, M. (Eds.) (2004), Metody statystycznej analizy wielowymiarowej w badaniach marketingowych [Multivariate statistical analysis methods in marketing research], Wydawnictwo AE, Wroclaw, 342-343, erratum.

Kaufman, L., Rousseeuw, P.J. (1990), Finding groups in data: an introduction to cluster analysis, Wiley, New York, pp. 83-88. ISBN: 978-0-471-73578-6.

See Also

index.G1, index.G2, index.G3, index.C, index.KL, index.H, index.Gap, index.DB

Examples

# Example 1
library(clusterSim)
data(data_ratio)
d <- dist.GDM(data_ratio)
c <- pam(d, 5, diss = TRUE)
icq <- index.S(d,c$clustering)
print(icq)

# Example 2
library(clusterSim)
data(data_ratio)
md <- dist(data_ratio, method="manhattan")
# nc - number_of_clusters
min_nc=2
max_nc=20
res <- array(0, c(max_nc-min_nc+1, 2))
res[,1] <- min_nc:max_nc
clusters <- NULL
for (nc in min_nc:max_nc)
{
cl2 <- pam(md, nc, diss=TRUE)
res[nc-min_nc+1, 2] <- S <- index.S(md,cl2$cluster)
clusters <- rbind(clusters, cl2$cluster)
}
print(paste("max S for",(min_nc:max_nc)[which.max(res[,2])],"clusters=",max(res[,2])))
print("clustering for max S")
print(clusters[which.max(res[,2]),])
#write.table(res,file="S_res.csv",sep=";",dec=",",row.names=TRUE,col.names=FALSE)
plot(res,type="p",pch=0,xlab="Number of clusters",ylab="S",xaxt="n")
axis(1, c(min_nc:max_nc))

clusterSim documentation built on Sept. 30, 2024, 9:15 a.m.