Lodestar.Metrics 0.3.0. This page is frozen at that release. Read the current documentation for what main says now. A link to a decision or a migration page follows main, and leaves the archive.

AdjustedMutualInformation

Shared information between two labellings, corrected for what chance alone would produce.

public static class AdjustedMutualInformation

Example — the same partition under different names.

using Lodestar.Metrics;

double same = AdjustedMutualInformation.Score([0, 0, 1, 1], [2, 2, 0, 0]);  // => 1

Remarks — NormalizedMutualInformation with the chance correction AdjustedRand applies to pair counts. That correction is what makes it safe to compare clusterings with different numbers of clusters: splitting a labelling further raises the raw mutual information, and raises the expected one with it.

Reference behaviour is sklearn.metrics.adjusted_mutual_info_score.

Applies to — net10.0, netstandard2.0.

See also — NormalizedMutualInformation, FowlkesMallows, the clustering index.

Members

Member	What it does
AdjustedMutualInformation.Score	Mutual information, corrected for chance.