Hierarchical Clustering Explainer
Hierarchical clustering creates a sequence of nested groupings rather than selecting one flat partition immediately. Agglomerative methods begin with individual observations and repeatedly merge groups, while divisive methods begin together and split them apart.
A distance measure compares observations, and a linkage rule defines the distance between groups. Single, complete, average, and Ward linkage emphasize different ideas of compactness and can produce substantially different structures.
The result is often displayed as a dendrogram showing the order and relative dissimilarity of merges. Cutting the tree at a chosen level produces a working set of clusters, but that cut is an analytical decision.
Hierarchical clustering is useful for exploratory taxonomy and data with nested organization. It is sensitive to scaling, outliers, distance definitions, and early merge decisions, so clusters require stability checks and subject-matter interpretation.
Hierarchical clustering organizes observations into nested groups by repeatedly merging or splitting clusters according to a distance rule.
Key terms include agglomerative clustering, divisive clustering, linkage, distance metric, dendrogram, merge, cut level and cluster stability.
Different linkage rules compare the closest, farthest, average, or variance-based relationship between groups.
Those choices can turn the same observations into different hierarchies.
The hierarchy contains several clustering levels rather than one automatically correct answer.
Scaling, outliers, and resampling tests help determine whether a chosen grouping is useful and stable.
Explore more "Explainers"
Discover additional explainers across politics, science, business, technology, and other fields. Each explainer breaks down a complex idea into clear, everyday language—helping you better understand how major concepts, systems, and debates shape the world around us.
