Metrics detcurve compute

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

DetCurve.Compute

Draws the detection error tradeoff curve — sklearn.metrics.det_curve.

public static DetCurve Compute(ReadOnlySpan<int> yTrue, ReadOnlySpan<double> yScore, int posLabel = 1, ReadOnlySpan<double> sampleWeight = default, bool dropIntermediate = false)

Parameters — yTrue is the true labels, one per sample. yScore is a score per sample. posLabel is the label counted as positive, 1 by default. sampleWeight is one weight per sample, or empty. dropIntermediate drops points whose true-positive count matches both neighbours; false here, as the reference has it.

Returns — a DetCurve whose FalsePositiveRate, FalseNegativeRate and Thresholds are three parallel arrays of equal length, ordered by ascending threshold.

Exceptions — ArgumentException when the inputs disagree in length, are empty, or hold a NaN score.

Example — the shortest of the three curves on the same four samples.

using Lodestar.Metrics;

int[] truth = [0, 0, 1, 1];
double[] scores = [0.1, 0.4, 0.35, 0.8];

DetCurve curve = DetCurve.Compute(truth, scores);
int points = curve.Thresholds.Count;  // => 3

Remarks — the false-negative rate is one minus the true-positive rate RocCurve reports at the same threshold, so the two curves carry the same information and differ only in what they make easy to see.

Applies to — net10.0, netstandard2.0.

See also — RocCurve.Compute, PrecisionRecallCurve.Compute, the Python equivalence table.