When looking at the histogram distribution of medoids from a cluster
analysis, clara{cluster}, they are close to normally distributed around
zero. The cluster plot, clusplot{cluster}, does not suggest distinct
partitions. How should the histogram distribution of medoids be
interpreted? Can one say that there is a relationship between lack of
obvious partitioning, and the distribution of the mediods? If clustering
was apparent, would it then be expected that medoids would not be
distributed normally?
Thanks
[[alternative HTML version deleted]]