This function performs repeated soft clustering for a range of cluster numbers c and reports the minimum centroid distance.
1 
eset 
object of class ExpressionSet. 
m 
value of fuzzy cmeans parameter 
crange 
range of number of clusters 
repeats 
number of repeated clusterings. 
visu 
If 
The minimum centroid distance is defined as the minimum distance between two cluster centers produced by the cmeans clusterings.
The average minimum centroid distance for the given range of cluster number is returned.
The minimum centroid distance can be used as cluster validity
index. For an optimal cluster number, we may see a ‘drop’ of minimum centroid distance
wh plotted versus a range of cluster number and a slower
decrease of the minimum centroid distance for higher cluster number.
More information and some examples can be found in the study of
Schwaemmle and Jensen (2010).
However, it should be used with care, as the determination remains
difficult especially for short time series and overlapping
clusters. Alternatively, the function cselection
can be used or
functional enrichment analysis (e.g. using Gene Ontology) can help to
adjust the cluster number.
Matthias E. Futschik (http://www.cbme.ualg.pt/mfutschik_cbme.html)
M.E. Futschik and B. Charlisle, Noise robust clustering of gene expression timecourse data, Journal of Bioinformatics and Computational Biology, 3 (4), 965988, 2005
L. Kumar and M. Futschik, Mfuzz: a software package for soft clustering of microarray data, Bioinformation, 2(1) 57,2007
Schwaemmle and Jensen, Bioinformatics,Vol. 26 (22), 28412848, 2010
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22  if (interactive()){
data(yeast)
# Data preprocessing
yeastF < filter.NA(yeast)
yeastF < fill.NA(yeastF)
yeastF < standardise(yeastF)
#### parameter selection
# For fuzzifier m, we could use mestimate
m1 < mestimate(yeastF)
m1 # 1.15
# or the function partcoef (see example there)
# For selection of c, either cselection (see example there)
# or
tmp < Dmin(eset,m=m1,crange=seq(4,40,4),repeats=3,visu=TRUE)# Note: This calculation might take some time
# It seems that the decrease for c ~ 20  25 24 and thus 20 might be
# a suitable number of clusters
}

Questions? Problems? Suggestions? Tweet to @rdrrHQ or email at ian@mutexlabs.com.
All documentation is copyright its authors; we didn't write any of that.