Abstract:

The development of information technology facilitates people’s life, but it also introduces risk of disclosure of personal privacy as well. In general, data anonymization is an effective way to prevent privacy disclosure. However, few existing anonymity principles conside the hostile attacks against datasets with function dependence. So we study the problems of privacy preserving data publishing (PPDP) while functional dependency exists in the datasets, and illustrate how to preserve privacy when privacy information is vulnerable with function dependence. First, we propose a (l,α)diversity privacy model to protect the privacy of individuals. To achieve better privacy protection for users and increase data utility, we use a hybrid method of perturbation and generalization/suppression to achieve an effective anonymous algorithm. We conduct experiments and make a detailed theoretical analysis for the experimental results. Experimental results verify the effectiveness and efficiency of the proposed algorithm.

Key words: privacy-preserving;data publishing;functional dependency;generalization;utility;perturbation