I need to form groups/clusters of size 3 to 4 from a list of examples (people with some attributes). I'm using the K-Means Constrained algorithm to limit the cluster size.
For some attributes, I need to make sure that people with the same values do not end up in the same cluster.
My understanding of the clustering algorithm is that it works based on similarity. I need to form clusters based on dissimilarity (of some attributes). How do I achieve this?