ExternalClusterMetricBase<T>
Base Classes · AiDotNet.Clustering.Evaluation
Base class for external cluster evaluation metrics that compare against ground truth.
Implements: IExternalClusterMetric<T>
Known derived types: Purity<T>
For Beginners
This base class handles the common math that all external metrics need:
- Building contingency tables (cross-tabulation of true vs predicted)
- Computing entropy (how “mixed” or “uncertain” clusters are)
- Counting pairs (how many point-pairs agree/disagree)
Individual metrics focus on their specific formulas while this class handles the plumbing.
How It Works
External cluster metrics compare clustering results against known ground truth labels. This base class provides common functionality shared by all external metrics including contingency table construction, entropy calculations, and pair counting.
Methods
| Method | Summary |
|---|---|
double Compute(Vector<T> trueLabels, Vector<T> predictedLabels) |