Metrics normalizedmutualinformation - CyrilB1531/lodestar GitHub Wiki
Development build. This page describes
main, not a released package. The latest published Lodestar.Metrics is 0.3.0 — read its documentation.
Home › Metrics › Clustering metrics
NormalizedMutualInformation
Information-theoretic rather than pair-counting, and equal to VMeasure by construction: scikit-learn's default normalizer is the arithmetic mean of the two entropies, which is what the harmonic mean of homogeneity and completeness reduces to. The frozen corpus shows the two agreeing on every case.
Members
| Member | What it does |
|---|---|
NormalizedMutualInformation.Score |
How much knowing one labelling tells you about the other, scaled into [0, 1]. |