apcluster (version 1.4.11)

labels-methods: Generate label vector from clustering result

Description

Generate a label vector from an clustering result

Usage

# S4 method for ExClust
labels(object, type="names")

Value

returns a label vector as long as the number of samples in the original data set

Arguments

object

object of class APResult or ExClust

type

specifies which kind of label vector should be created, see details below

Author

Ulrich Bodenhofer & Andreas Kothmeier

Details

The function labels creates a label vector from a clustering result. Which kind of labels are produced is controlled by the argument type:

“names”

(default) returns the name of the exemplar to which each data sample belongs to; if no names are available, the function stops with an error;

“enum”

returns the index of the cluster to which each data sample belongs to, where clusters are enumerated consecutively from 1 to the number of clusters (analogous to other clustering methods like kmeans);

“exemplars”

returns the index of the exemplar to which each data sample belongs to, where indices of exemplars are within the original data, which is nothing else but the slot object@idx with attributes removed.

References

http://www.bioinf.jku.at/software/apcluster/

Bodenhofer, U., Kothmeier, A., and Hochreiter, S. (2011) APCluster: an R package for affinity propagation clustering. Bioinformatics 27, 2463-2464. DOI: tools:::Rd_expr_doi("10.1093/bioinformatics/btr406").

See Also

APResult, ExClust, cutree

Examples

## create two simple clusters
x <- c(1, 2, 3, 7, 8, 9)
names(x) <- c("a", "b", "c", "d", "e", "f")

## compute similarity matrix (negative squared distance)
sim <- negDistMat(x, r=2)

## run affinity propagation
apres <- apcluster(sim)

## show details of clustering results
show(apres)

## label vector (names of exemplars)
labels(apres)

## label vector (consecutive index of exemplars)
labels(apres, type="enum")

## label vector (index of exemplars within original data set)
labels(apres, type="exemplars")

## now with agglomerative clustering
aggres <- aggExCluster(sim)

## label (names of exemplars)
labels(cutree(aggres, 2))