A technique used to release person-specific data such that the ability to link to other information using the quasi-identifier is limited. K-anonymity achieves this through suppression of identifiers and output perturbation.