ANNM: A New Method for Adding Noise Nodes Which are Used Recently in Anonymization Methods in Social Networks

Abstract

One of the main concerns at the time of production or share of information on social networking sites for scientific research and business analysis is privacy. Recently, different models of privacy such as k-anonymity have been created by researchers to avoid detection by using structural information. But still, attackers may be able to access private information by observing the behavior of some nodes in social networks. Current approaches that mainly focus on creating anonymity by edge editing or clustering may significantly change the properties of the social network graph. According to studies of Yuan et al. (IEEE Trans Knowl Data Eng 25(3):633–647, 2013), that makes anonymity with adding noise nodes, we decided to present a new method for adding noise nodes with least changes in main graph attributes. We used betweenness centrality measurement to prioritize the creation of noise nodes and considered the amount of their impact on graph properties. The result of comparing our proposed solution and other related works shows that the structural properties of the original social network graph have had very little change.