Spectrum Privacy Preserving for Social Networks: A Personalized Differential Privacy Approach

Abstract

AbstractThe characteristics of social networks has always been a hot topic of scientific research. In order to protect the privacy of users, the owner of the private data need to provide privacy protection when providing inquiries or publishing data. Local differential privacy (LDP) is difficult to construct a highly available social networks graph due to its independent perturbation process. Centralized differential privacy usually adds excessive noise due to the structural characteristics of social networks graphs. Higher security usually results in lower availability. Simply implementing any differential privacy mechanism will cause a large amount of data to be disturbed by noise. On the other hand, some spectral based privacy protection methods provide accurate spectrum, however ignore the disclosure of privacy data in spectrum query. Therefore, we propose a spectrum query algorithm based on personalized differential privacy. The algorithm effectively improves data availability by taking advantage of different privacy preferences of users in the social network and the characteristics of the spectrum. To verify the availability of these methods, experimental tests have been carried out in both model networks and actual networks, which shows that the algorithm improves the availability of data when it has the same security.KeywordsSocial networksDifference privacySpectrum