A Brief Survey on Privacy-Preserving Methods for Graph-Structured Data

As one of the main manifestations of big data, graph-structured data widely exists in various fields such as social networks, smart cities, medical health, and finance, and is characterized by high dimension, nonlinear, scale-free, small world, etc. Extensive graph-structured data provides sufficient data resources for scientific research and commercial applications. However, graph data mining not only reveals the intrinsic value of data, but also brings the risk of privacy disclosure. Therefore, how to protect graph-structured data privacy is of great significance. In this paper, we survey the very recent research development on privacy preserving methods for graph-structured data. We introduce the common sensitive information and the related privacy risks in graph data, and elaborate the background knowledge for privacy inference attack. Then, the privacy inference attack methods and privacy preservation methods for graph-structured data are summarized. Finally, the shortcomings of the current research about graph-structured data privacy preservation and the possible research directions are discussed.