apcluster (version 1.4.3)

APResult-class: Class "APResult"

Description

S4 class for storing results of affinity propagation clustering. It extends the class .

Objects

Objects of this class can be created by calling apcluster or apclusterL for a given similarity matrix or calling one of these procedures with a data set and a similarity measure.

Slots

The following slots are defined for APResult objects. Most names are taken from Frey's and Dueck's original Matlab package:

sweeps:

number of times leveraged clustering ran with different subsets of samples

it:

number of iterations the algorithm ran

p:

input preference (either set by user or computed by apcluster or apclusterL)

netsim:

final total net similarity, defined as the sum of expref and dpsim (see below)

dpsim:

final sum of similarities of data points to exemplars

expref:

final sum of preferences of the identified exemplars

netsimLev:

total net similarity of the individual sweeps for leveraged clustering; only available for leveraged clustering

netsimAll:

vector containing the total net similarity for each iteration; only available if apcluster was called with details=TRUE

exprefAll:

vector containing the sum of preferences of the identified exemplars for each iteration; only available if apcluster was called with details=TRUE

dpsimAll:

vector containing the sum of similarities of data points to exemplars for each iteration; only available if apcluster was called with details=TRUE

idxAll:

matrix with sample-to-exemplar indices for each iteration; only available if apcluster was called with details=TRUE

Extends

Class "ExClust", directly.

Methods

plot

signature(x="APResult"): see plot-methods

plot

signature(x="ExClust", y="matrix"): see plot-methods

heatmap

signature(x="ExClust"): see heatmap-methods

heatmap

signature(x="ExClust", y="matrix"): see heatmap-methods

show

signature(object="APResult"): see show-methods

labels

signature(object="APResult"): see labels-methods

cutree

signature(object="APResult"): see cutree-methods

length

signature(x="APResult"): gives the number of clusters.

sort

signature(x="ExClust"): see sort-methods

as.hclust

signature(x="ExClust"): see coerce-methods

as.dendrogram

signature(object="ExClust"): see coerce-methods

Accessors

In the following code snippets, x is an APResult object.

x[[i]]: Returns the i-th cluster as a list of indices of samples belonging to the i-th cluster.

x[i]: Returns a list of integer vectors with the indices of samples belonging to this cluster. The list has as many components as the argument i has elements. A list is returned even if i is a single integer.

similarity

signature(x="APResult"): gives the similarity matrix.

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: 10.1093/bioinformatics/btr406.

Frey, B. J. and Dueck, D. (2007) Clustering by passing messages between data points. Science 315, 972-976.

See Also

apcluster, apclusterL, show-methods, plot-methods, labels-methods, cutree-methods

Examples

## create two Gaussian clouds
cl1 <- cbind(rnorm(100,0.2,0.05),rnorm(100,0.8,0.06))
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)

## run affinity propagation
apres <- apcluster(sim, details=TRUE)

## show details of clustering results
show(apres)

## plot information about clustering run
plot(apres)

## plot clustering result
plot(apres, x)

## plot heatmap
heatmap(apres, sim)