📏  Multilabel Precision

Multilabel Precision#

Precision when each instance may carry several labels at once.

Important

✨ AI-generated content. This page was written with the assistance of an AI language model and is provided as a learning aid. Despite careful review, it may still contain mistakes, omissions, or out-of-date information. Whether you are new to the topic, a team lead, or a senior practitioner, treat it as a starting point rather than an authoritative reference: read it critically and independently verify anything you act on (code, commands, figures, and factual claims) against official documentation and primary sources before relying on it.

What it is#

Multilabel precision is precision when each sample can carry several labels at once — the labels are not mutually exclusive. Targets are an indicator matrix (cell [i, j] = 1 if sample i has label j), and precision is computed per label, each label a binary problem.

How it’s aggregated#

As with multiclass, per-label precisions are combined by micro, macro or weighted averaging — but multilabel adds a distinctive ‘samples’ average, which computes precision per instance (across that sample’s labels) and averages over samples.

Why it differs#

Unlike multiclass, where exactly one class is correct, multilabel labels co-occur, so a prediction can be partly right (some labels correct, others missed). The averaging choice — especially samples vs micro — decides whether you’re scoring per-label or per-example accuracy.


Theme: Classification & Averaging Metrics  ·  All terminology



See also

Source article Adapted (context, re-expressed) in our own words from: Multilabel Precision (insightful-data-lab.com).

Tags: purpose: reference topic: terminology level: intermediate