Seed-and-Grow Algorithm to Identify users from Anonymized Social Networks

Digital traces left by users of on-line social networking services, even once Anonymization, square measure at risk of privacy breaches. This is often exacerbated by the increasing overlap in user-bases among varied services. To alert fellow researchers in each the domain and also the trade to the practicability of such AN attack, we have a tendency to propose an formula, Seed-and-Grow, to spot users from AN anonymzed social graph, primarily based only on graph structure. The formula initial identifies a seed subgraph, either planted by AN offender or divulged by a collusion of a little cluster of users, then grows the seed larger supported the attacker's existing data of the users ’social relations. Our work identifies and relaxes implicit assumptions taken by previous works, eliminates discretional parameters, and improves identification effectiveness and accuracy. Simulations on real-world collected information sets verify our claim.