FTB-C000091 / Data quality concept

Outlier

An outlier is an observation that is unusually distant or inconsistent under a specified comparison model.

Also known asanomalous valueextreme observation

Definitions

In plain terms

It may be an error, a rare genuine event, or evidence that the chosen model is inappropriate; detection alone cannot decide which.

Technical

Outlier status depends on population, feature, window, robust statistic or model, threshold, causal direction, and boundary policy.

Scope

An outlier flag is not proof of bad data and should preserve the original observation.

Examples

  • A governed dataset records outlier explicitly and preserves the source context needed to interpret it.

Common misconceptions

  • Outlier does not have one universal meaning or policy without its declared source, scope, and data contract.

Concept relationships

Where this concept is used

Evidence and governance

  1. Fintech Builder Glossary Definition Contract The Fintech Builder · internal governed contract

    Supports: preferred label, short definition, technical definition

    Limits: Defines governed platform behavior rather than claiming these policies are universally prescribed by external McClellan methodology.

Reviewed by
fintech-builder-owner-approved
Last reviewed
2026-07-26
Next review
2027-07-26
Record status
published