An enhanced structure-based de-anonymization of online social networks

Abstract

To protect users’ privacy, online social network data are usually anonymized before being sold to or shared with third parities. Various structure-based approaches have been proposed to de-anonymize the social network data. In this paper, we study the limitations of the existing structure-based de-anonymization methods and propose an enhanced de-anonymization algorithm. The basic idea of our algorithm is to leverage the structural transformations of the social graph to de-anonymize the social network data. We also define a new similarity measure that is more robust for de-anonymization. We use the arXiv dataset to evaluate our algorithm, and the experiment results show that our method can significantly improve the de-anonymization rate.