Beyond k-Anonymity: l-Diversity and t-Closeness

Abstract

The main advantage of k-anonymity is that it provides an intuitive notion of disclosure risk limitation. The principle that underlies k-anonymity is that an individual’s privacy must be protected if the corresponding record is hidden within a group of k records. However, this principle fails to provide sufficient protection when the records in the k-anonymous group have a similar value for the confidential attribute. In other words, k-anonymity provides protection against identity disclosure but that is not enough to prevent attribute disclosure when the values of the confidential attribute are similar across records.