Complete Linkage
When hierarchical clustering builds groups step by step, it needs a rule for deciding which two clusters are “closest.” Complete linkage is one of those rules, and it takes a strict view: two clusters are considered close only if even their farthest pair of points is still close.
How it works
In agglomerative hierarchical clustering, every data point starts as its own cluster. At each step, the algorithm merges the pair of clusters with the smallest distance according to a linkage criterion. With complete linkage, the distance between two clusters is defined as the maximum distance between any point in one cluster and any point in the other. That means a merge is allowed only when the two clusters fit together tightly.
Why this matters
This choice strongly affects the shape of the clusters you get. Complete linkage tends to produce:
- Compact, tight clusters rather than long, chain-like ones
- Better separation between groups when you want all members of a cluster to be mutually close
- More resistance to the “chaining effect” seen in single linkage, where clusters can grow by thin bridges of nearby points
The tradeoff is that complete linkage can be sensitive to outliers, because one faraway point can make two otherwise similar clusters look distant.
Practical use
This is useful in tasks like customer segmentation, where you want each segment to be internally consistent, or in document clustering, where a topic group should stay tight instead of stretching across loosely related texts. In Python, you will see it in tools like scikit-learn via AgglomerativeClustering(linkage="complete") or in SciPy’s hierarchical clustering functions. On a dendrogram, complete linkage usually creates merges that reflect stricter cluster boundaries, which helps when choosing where to cut the tree into final groups.
Complete Linkage is a hierarchical clustering criterion that defines the distance between two clusters as the greatest distance between any pair of points, one from each cluster. It merges clusters only when their farthest members are still relatively close, producing compact, tightly bounded groups and resisting chaining effects. This matters because the linkage rule directly shapes cluster structure, separation, and interpretability in agglomerative clustering.
Imagine grouping people at a party into friend circles, but only merging two circles if even the two people who are least alike across those circles are still fairly similar. That is the intuition behind Complete Linkage.
In AI, Complete Linkage is a way of deciding when two groups of data should be treated as one group. It looks at the farthest-apart members of each group and uses that as the standard. This tends to create tighter, more compact groups, because one very different item can stop two groups from being merged too soon. It matters when you want clusters that feel neat and well separated, not loose or stretched out.