K Anonymity Overview
K anonymization is privacy preserving algorithm.
k-anonymity objective is to make it difficult to identify individuals in a dataset.
It ensure there are at least K individuals (or at least k-1 other similar records)
Select value of K
To achieve k-anonymity, a dataset is first divided into equivalence classes. An equivalence class is a group of records that are indistinguishable from each other.
If there are less than K record in equivalence class, then data is purturb in such a manner to make them equal. To purturb the data one has to use approaches to generalize data or suppress sensitive info.