Cophenetic correlation coefficient

The cophenetic correlation coefficient measures how well a hierarchical clustering method reflects the actual similarities between data points. Imagine grouping related items in a tree; the coefficient compares the original pairwise distances between items with the distances implied by the tree. A high value (close to 1) means the tree accurately represents the true relationships among data points, while a low value indicates less fidelity. Essentially, it’s a statistic that tells us how reliably the clustering tree captures the real structure in the data.