Algorithm 2. Agglomeration.

1: procedure AGGLOMERATION(G = (V,E), an initial clustering, ${\sigma }_A^2$)

2:  Compute the between-cluster evaluation function between any two initial clusters.

3:  for for every cluster from the initial clustering in a random order do

4:   Find the closest initial cluster using the average between-cluster evaluation function $D_{{\sigma }_A^2}$.

5:   Merge both clusters.