# Anomaly detection · Glossary
Identifying data points, events or patterns that differ significantly from what is normally expected.
[Glossary](/glossary) · Machine learning and MLOps

# Anomaly detection

     Identifying data points, events or patterns that differ significantly from what is normally expected. It is used to spot fraud, equipment faults, data errors, security incidents and unusual operational behaviour.

## Why it matters

     Many important events are rare and poorly labelled, which makes them hard to learn with supervised methods. Anomaly detection highlights the unusual so people can investigate, even when past examples are scarce.

     Its value depends on the investigation process around it. Too many alerts cause fatigue; too few miss real problems. Thresholds should be tuned with the teams who act on the alerts.

## In practice

     For example, a UK food manufacturer might monitor temperature and vibration sensors on its production lines and flag readings that depart from each machine’s normal pattern, prompting a maintenance check before a fault stops the line.

## Where Rodan fits

     Rodan builds anomaly detection into operational monitoring and decision workflows in [AI and Decision Systems](https://rodan.io/what-we-build/ai-decision-systems) work.

## Related terms

- [Unsupervised learning](/glossary#unsupervised-learning)

- [Machine learning (ML)](/glossary#machine-learning)

- [Data observability](/glossary/data-observability)

- [Real-time analytics](/glossary/real-time-analytics)

- [Digital twin](/glossary#digital-twin)

- [Stream processing](/glossary/stream-processing)
HTML: https://rodan.io/glossary/anomaly-detection