Hierarchical Cluster Comparison

Hierarchical Cluster Comparison by E. B. Fowlkes and C. L. Mallows (1983)

The zip file contains three functions written for the paper "Comparative adenine dinucleotide phosphate
(NADPH)-diaphorase histochemistry in amphibian brain: a cluster analysis" by Claudia Pinelli, Rakesh K. Rastogi, Anna Scandurra, Arun G. Jadhao, Massimo Aria, Biagio D’Aniello, accepted on Journal of Comparative Neurology (2014).

The functions are:

- Cluster_comparison.m that calculates the B measure proposed by E. B. Fowlkes and C. L. Mallows (1983) to compare the results of two different hierarchical clustering performed on the same dataset;
- Bootstrap_compare_cluster.m that estimates the 95% bootstrap confidence interval of B measure;
- PermTest_cluster_compare.m that performs a permutation test on B measure.
Null Hypothesis of the test is: The two hierarchical partitions (C_A and C_B) are completely uncorrelated, in other word, are completely different.
Alternative Hypothesis is: The two hierarchical partitions are correlated, in other word, have significant similarity.
C_A represents a well-known partition of objects provided by theoretical assumptions.
C_B represents a hierarchical partition obtained by a statistical analysis on a real dataset.