Interface: ClassificationReport

Defined in: ml/metrics.ts:287

A scikit-learn style classification report.

Properties

accuracy

accuracy: number;

Defined in: ml/metrics.ts:289

---

macro

macro: {
  f1: number;
  precision: number;
  recall: number;
};

Defined in: ml/metrics.ts:291

Unweighted mean over classes — every class counts the same.

f1

f1: number;

precision

precision: number;

recall

recall: number;

---

perClass

perClass: ClassScore[];

Defined in: ml/metrics.ts:288

---

weighted

weighted: {
  f1: number;
  precision: number;
  recall: number;
};

Defined in: ml/metrics.ts:293

Support-weighted mean — dominated by the common classes.

f1

f1: number;

precision

precision: number;

recall

recall: number;