Definition:
Cluster separation measures how far apart different clusters are from one another. It is a proxy for cluster distinctiveness.

Common Approaches:

  • Inter-cluster distance
  • Davies–Bouldin index
  • Silhouette score, Silhouette Analysis (combines cohesion and separation)

Exploratory Questions:

  • What does high separation mean if internal density is low?
  • How does Curse of dimensionality affect separation metrics?
  • What clustering algorithms maximize separation explicitly?