octopucs - Statistical Support for Hierarchical Clusters
Generates n hierarchical clustering hypotheses on subsets
of classifiers (usually species in community ecology studies).
The n clustering hypotheses are combined to generate a
generalized cluster, and computes three metrics of support. 1)
The average proportion of elements conforming the group in each
of the n clusters (integrity). And 2) the contamination, i.e.,
the average proportion of elements from other groups that enter
a focal group. 3) The probability of existence of the group
gives the integrity and contamination in a Bayesian approach.