Is it possible to decide which cluster to be used by CE calculations? I know nbclusters.in but there as far as I understand we can only decide the number of pairs, triplets … to be used.
What I want is something like to get many clusters by corrdump and just choose some which I think can fit the structure better and run and get ECI’s of these clusters by maps.
Is this kind of thing possible?
I’m sorry to piggyback on this thread, but I have a question relating to this. Is there a way to get the cross validation score for a specific choice of clusters using the clusterexpand command or something similar?