Authors: Gautam Srivastava, Mykel Shumay and Evan Citulsky
In recent years, there has been a significant increase in the use of graph-formatted data. Many applications such as so-cial networks, recommendation systems, and other collabo-rative applications are built on top of graph infrastructures. These infrastructures in turn represent relationships among users and present interesting information for researches and other third-parties. The problem appears when someone wants to publicly release this information, especially in the case of social or healthcare networks. In these cases, it is essential to implement an anonymization process in the data in order to preserve the privacy of users who appears in the network. We focus here on the recent explosion of gaming in social network sites. Given that many people are playing games and linking to other players using social networking, the risk has become even higher for data breaches as these gaming platforms tend not to guarantee much security. The contributions of this paper are two fold: we first look to build on our previous work in anonymization of graph data using Ant Colony Otimizations to help aid in the proce-dure. We introduce a new method to anonymize labeled graphs using Ant Colony Systems (ACS). Furthermore, we also take on a systematic study of small world networks and apply those finding to the creation and testing of graph data as it applied to the graph infrastructures described above.
Keywords: graphs ant colony social network privacy degree anonymization securityl