k-Anonymity
Definition
k-Anonymity requires that every combination of the specified quasi-identifiers in the published data appears at least k times. The dataset under consideration counts as well.
Quasi-identifiers are generalized or suppressed for this purpose. The method reduces certain re-identification risks but does not automatically protect against attribute disclosure or background knowledge.
Simply explained
How to publish data so that each row belongs to a group of at least k rows based on the specified quasi-identifiers.
Why it matters in practice
Relevant for understanding data protection and data releases, but often not sufficient alone for sensitive data.
Example from business
Before external analysis, age, location, and occupation are generalized so that each combination includes at least ten records.
Limits and pitfalls
- Treating k as a complete proof of data protection; ignoring sensitive attributes within groups; not considering external auxiliary data.
Origin & history
Samarati and Sweeney worked on k-Anonymity in 1998. Sweeney further formalized the model in 2002 and is credited with the naming Samarati.
Use in mid-sized companies
- Possible application: risk analysis before data release, typically combined with additional data protection measures.
Frequently asked questions
Does k-Anonymity prevent all re-identification?
No. What are quasi-identifiers? Attributes that can indirectly identify individuals when combined.
Sources
- k-Anonymity: A Model for Protecting Privacy · Sweeney (2002), k-Anonymity: A Model for Protecting Privacy
- Volltext · Sweeney (2002), Volltext
Source check: 10./11.10.2026.
Editorially reviewed on: