k-anonymity

Crowd-based data protection

K-anonymity is a privacy standard for datasets requiring that any individual's record be indistinguishable from at least k-1 other records.

If a medical database has 5-anonymity, every combination of identifying traits (age, zip code, gender) matches at least five people, preventing someone from being singled out.

The concept guides how organizations anonymize data before sharing it.

technique

generalization

Deriving principles from examples

hierarchy · bucketing · binning · rounding · aggregation · coarsening · swapping

privacy

anonymization

Removing identifying information

re-identification · microdata · confidentiality · sensitive data · background knowledge · adversary

database

dataset

Organized collection of information

  • recordPreserved account or achievement
  • tupleOrdered finite sequence
  • attributeQuality assigned to something

column · row · tabular data · census · medical records · survey data · demographic data · public records · voter registration · data sharing

Categories
Opposites
privacy protection level

k-anonymityre-identification

Linking data back to someone

personal identification · singling out

data identifiability

k-anonymityPII

Identity-revealing information

fingerprinting · tracking · profiling

Family

anonymous, anonymity, pseudonymous, pseudonymity, anonymized, anonymize, anonymizing

Familiarity
Top 2,000,000
Profile
Used worldwide · Timeless · Abstract · Rated G