The smallest number in this matrix is 3.16, which indicates that samples 1 and 2 are nearer than any other pair of samples. Single-linkage clustering (and most other agglomerative algorithms) therefore joins samples 1 and 2 into a group. Note that single-linkage clustering ignores the properties of the groups: only individual samples are compared. All possible distances between entities are calculated, with distance defined as the shortest distance involved in a comparison of two entities (e.g., the distance between samples 1 and 3 is 4.47 and the distance between samples 2 and 3 is 7.07, so the distance is defined as the shorter of the distances, 4.47):
Was this article helpful?