Computer scienceData scienceMachine learningClusteringClustering evaluation

Introduction to clustering evaluation

Theory

The internal limitation

Report a typo

Suppose we have multiple datasets:

1. A scatterplot of clusters with non-convex shapes, different densities and noise

2. A scatter plot of 4 round well-separated clusters

3. A scatter plot of 3 non-spherical clusters

4. A scatter plot of non-spherical not well separated clusters

Select the options where the internal metric will output relatively low scores due to the data structure, regardless of the actual clustering quality. Enter the answer, with each option separated by a space.

Pay attention to the various densities and cluster shapes in the graphs. Also, consider the presence of outliers (points that do not belong to any particular cluster) in each case.
Enter a short text
___

Create a free account to access the full topic