Hi. The clustering algorithms present in R assume i have a data matrix.
However my data is expressed as a graph (nodes and weights between these
nodes). Generally i can work around this by replacing the weights as
similarities in the dissimilarity matrix, but as R clustering algorithms
like MClust work from the original data matrix i cant do this in this case.
Is there any way i could obtain a suitable matrix to replace the data matrix
starting just from the knowledge i have from the graph? I have no idea if
this is even mathematically feasible as you seem to lose information when
going from the data matrix to the dissimilarity matrix.
Well, thanks
[[alternate HTML version deleted]]
______________________________________________
[EMAIL PROTECTED] mailing list
http://www.stat.math.ethz.ch/mailman/listinfo/r-help
