Hierarchical clustering

Posted: 13/09/2025

Hierarchical clustering techniques are an important category of clustering methods. They can be divided in:

The first group is most commonly used. A hierarchical clustering can be graphically displayed by using a dendrogram, which is a tree-like graph that displays the order in which the clusters were merged (agglomerative) or split (divisive).

Algorithm 1: Agglomerative hierarchical clustering algorithm.


    1. Compute the proximity matrix.
    2. repeat
    3.    Merge the closest two points.
    4.    Update the proximity matrix to reflect the proximity between the new cluster
    and the original cluster.
    5. until Only one cluster remains.
    

The key issue is defining proximity between clusters. Common measures include:

This clustering algorithm does not assume a particular number of clusters, since the “desired” number can be obtained by simply cutting the dendrogram at the appropriate height.

← Back to Blog List