Total separation between clusters - Internal Measure

Share:

Description

Function computes total separation between clusters.

Usage

1
clv.Dis(cluster.center)

Arguments

cluster.center

numeric matrix or data.frame where columns correspond to variables and rows cluster centers.

Details

The definition of total separation between clusters is given by equation:

Dis = (Dmax/Dmin) * sum{forall i in 1:|C|} 1 /( sum{forall j in 1:|C|} ||vi - vj|| )

where:

|C| - number of clusters,
vi, vj - centers of clusters i and j,
Dmax - defined as: max{||vi - vj||: vi,vj - centers of clusters },
Dmin - defined as: min{||vi - vj||: vi,vj - centers of clusters },
||x|| - means: sqrt(x*x').

This value is a part of clv.SD and clv.SDbw.

Value

As result Dis value is returned.

Author(s)

Lukasz Nieweglowski

References

M. Haldiki, Y. Batistakis, M. Vazirgiannis On Clustering Validation Techniques, http://citeseer.ist.psu.edu/513619.html

See Also

clv.SD and clv.SDbw

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
# load and prepare data
library(clv)
data(iris)
iris.data <- iris[,1:4]

# cluster data
agnes.mod <- agnes(iris.data) # create cluster tree 
v.pred <- as.integer(cutree(agnes.mod,5)) # "cut" the tree 

# compute Dis index
scatt <- clv.Scatt(iris.data, v.pred)
dis <- clv.Dis(scatt$cluster.center)