A 'bottom-up' hierarchical clustering approach where each observation starts in its own cluster, and pairs of clusters are merged as one moves up the hierarchy.
Most common form of hierarchical clustering.
Used when the hierarchy matters (e.g., taxonomy).