k-anonymity is a traditional anonymization technique that preserves privacy by creating groups consisting of k records that are indistinguishable from each other, so that the probability that the person is identified based on the quasi-identifiers is not more than 1/k.