How to Find an Appropriate K for K-Anonymization

Abstract

Personalization has been implemented in a variety of services. k-anonymity is the most common definition for the anonymization of personal data sets, and is considered to be a normal feature of personal data exchanges. However, there is an important issue: How to find an appropriate k for k-anonymity. In this paper, we present a model for finding an appropriate k in k-anonymization. The model suggests that an optimal k exists that is appropriate to the balance between value and risk when personal data are published. This is the first step towards realizing an optimized configuration for publication of personal data.