Similarity (Agglomerative Clustering)

The similarity between two clusters $C_1=\{a,b\}$ and $C_2=\{c,d\}$ is