Computer scienceData scienceMachine learningClusteringClustering evaluation

Introduction to clustering evaluation

The ranges

Report a typo

Connect the metric to it's score range:

Match the items from left and right columns

The silhouette score

The Davies-Bouldin index

The Calinski-Harabasz index

Rand Index

Mutual information

Unbounded, higher scores indicate better clustering

[0,1], higher scores are better

Unbounded, lower scores indicate better clustering

[-1,1], higher scores indicate better clustering

The original version is unbounded, but there variants in a bounded interval

___

Create a free account to access the full topic