Differentially Private k-Anonymity: Achieving Query Privacy in Location-Based Services

Abstract

While location based services (LBS) are facilitating our daily life, the privacy issues including location privacy and query privacy are also arising. Existing efforts aiming to preserving query privacy rely on third party servers or fail to provide guaranteed privacy degree. In this paper we combine the concepts of differential privacy and k-anonymity to establish a novel notion of differentially private k-anonymity (DPkA) for query privacy in LBS. The sufficient and necessary condition for the availability of 0-DPkA is recognized and a mechanism M0 to achieve 0-DPkA is given. For other cases that 0-DPkA is impossible, this paper presents an algorithm to achieve ε-DPkA with bounded ε.