Metrics silhouette persample

Development build. This page describes main, not a released package. The latest published Lodestar.Metrics is 0.3.0 — read its documentation.

Silhouette.PerSample

The score of each sample, from the samples themselves.

public static double[] PerSample(ReadOnlySpan<int> labels, ReadOnlySpan<double> features, int featureCount)

Parameters — as Silhouette.Score: labels, the samples features row-major, and featureCount.

Returns — double[], one value per sample in the order the samples were given.

Exceptions — ArgumentException when the inputs disagree in size, and when the number of distinct labels falls outside [2, n - 1] — scikit-learn's own bound, carried with its own sentence: Number of labels is 1. Valid values are 2 to n_samples - 1 (inclusive).

Example — the mean hides which sample is misplaced; this does not.

using Lodestar.Metrics;

double[] features = [0.0, 0.0, 0.2, 0.1, 4.0, 4.0, 4.2, 3.9, 0.1, 0.3];
int[] labels = [0, 0, 1, 1, 1];

double[] scores = Silhouette.PerSample(labels, features, 2);
double stranger = scores[4];   // => -0.9501…

Remarks — the negative value is the point. Sample 4 was labelled into the far cluster while sitting among the near one, and no mean would have told you which sample to look at.

A cluster holding one sample scores that sample 0.0 rather than dividing by zero: there is no other member to be close to, so it is neither well nor badly placed. That is scikit-learn's answer, measured.

Applies to — net10.0, netstandard2.0.

See also — Silhouette.Score, Silhouette.PerSampleFromDistances, the Python equivalence table.