📏  Multiclass Precision

Multiclass Precision#

Precision aggregated across the classes of a multiclass problem.

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 Despite careful review, it may still contain mistakes, omissions, or out-of-date information.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#

Multiclass precision is precision for a problem with more than two mutually exclusive classes. Since precision is defined on a binary positive/negative split, it is computed by treating the task as K one-vs-rest binary problems — the per-class precisions — then reduced to a single number.

How it’s reduced#

The per-class values are combined by an averaging scheme — micro (pool counts, majority-dominated), macro (equal weight per class), or weighted (by support). The choice determines whether rare classes are surfaced or hidden, so it must be stated with the score.

The caveat#

Because each class is scored against “the rest,” each binary split is imbalanced; reporting the per-class precisions guards against an average that looks good only because the majority class does.


Theme: Classification & Averaging Metrics  ·  All terminology



See also

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

Tags: purpose: reference topic: terminology level: intermediate