# What is One-class SVM? Definition and meaning

A kernel method that fits a boundary around normal samples and flags points outside it.

## Definition

**One-class SVM**: A kernel method that fits a boundary around normal samples and flags points outside it.

In short: Learning the boundary of normal data.

## When it is used

Anomaly detection when only normal examples are available.

## Learn more

The full guide to **[One-class SVM](/tools/one-class-svm)** explains how it works, when to use it and how it compares with the alternatives, with examples. The [glossary](/glossary) lists every term on the site from A to Z.

Previous term: [OIDC](/glossary/oidc) · Next term: [Online travel](/glossary/online-travel)
