apcluster (version 1.2.1)

aggExCluster: Exemplar-based Agglomerative Clustering

Description

Runs exemplar-based agglomerative clustering for a given similarity matrix

Usage

aggExCluster(s, cl)

Arguments

s

an $l\times l$ similarity matrix

cl

a prior clustering; if present, cl must be an object of class APResult or ExClust.

Value

• Upon successful completion, the function returns an AggExResult object.

code

aggExCluster

sQuote

• Cluster 1
• Cluster 2

Details

aggExCluster performs agglomerative clustering. Unlike other methods, e.g., the ones implemented in hclust, aggExCluster is computing exemplars for each cluster and its merging objective is geared towards the identification of meaningful exemplars, too.

For each pair of clusters, the merging objective is computed as follows:

1. An intermediate cluster is created as the union of the two clusters.

The potential exemplar is selected from the intermediate cluster as the sample that has the largest average similarity to all other samples in the intermediate cluster. Then the average similarity of the exemplar with all samples in the first cluster and the average similarity with all samples in the second cluster is computed. These two values measure how well the joint exemplar describes the samples in the two clusters. The merging objective is finally computed as the average of the two measures above. Hence, we can consider the merging objective as some kind of balanced average similarity to the joint exemplar.

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: http://dx.doi.org/10.1093/bioinformatics/btr406{10.1093/bioinformatics/btr406}.

See Also

AggExResult, plot-methods, cutree-methods

Examples

## create two Gaussian clouds
cl1 <- cbind(rnorm(50,0.2,0.05),rnorm(50,0.8,0.06))
cl2 <- cbind(rnorm(50,0.7,0.08),rnorm(50,0.3,0.05))
x <- rbind(cl1,cl2)

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

## compute agglomerative clustering from scratch
aggres1 <- aggExCluster(sim)

## show results
show(aggres1)

## plot dendrogram
plot(aggres1)

## plot heatmap along with dendrogram
plot(aggres1, sim)

## plot level with two clusters
plot(aggres1, x, k=2)

## run affinity propagation
apres <- apcluster(sim, q=0.7)

## create hierarchy of clusters determined by affinity propagation
aggres2 <- aggExCluster(sim, apres)

## show results
show(aggres2)

## plot dendrogram
plot(aggres2)

## plot heatmap
plot(aggres2, sim)

## plot level with two clusters
plot(aggres2, x, k=2)