Computer scienceData scienceMachine learningClusteringClustering evaluation

Introduction to clustering evaluation

Theory

Consider the scenarios

Report a typo

Below you have a list that has descriptions of the data and the algorithms used for clustering:

A. There are the ground truth labels. The dataset is balanced (the number of points in each class is roughly similar). The cluster count is >20>20.

The first dataset

B. There are no ground truth labels and there are no significant outliers in the dataset. The classes could be clustered into several distinct convex clusters.

The second data set

Match the scenario to the most suitable metric.

Match the items from left and right columns
A
B
The Davies-Bouldin index
The normalized mutual information
___

Create a free account to access the full topic