Interface: ConfusionMatrix

Defined in: ml/metrics.ts:10

A confusion matrix: raw counts plus a row-normalized (recall) view.

Properties

classes

classes: number;

Defined in: ml/metrics.ts:17

---

counts

counts: Float64Array;

Defined in: ml/metrics.ts:12

Row-major counts, length classes*classes; row = true label, col = predicted.

---

normalized

normalized: Float64Array;

Defined in: ml/metrics.ts:14

Row-normalized: every row sums to 1 (a row with no support stays all-zero).

---

support

support: Float64Array;

Defined in: ml/metrics.ts:16

Per-true-class support (row totals).