# Different Categories of cluster validation:
- external indices (you have labelled data)
- internal indices
- relative indices
Cluster results might be evaluated on :
- compactness
- separability
# External Indices
Score range from -1 to 1
# Internal Indices
In the below image, there are 3 clusters. The further we are from k=3, score decreases
In the below image, we compare Silhouette score on the two ring dataset.
We see that although Single Link and DBScan clustered it well, the score is really low
