Graph-based Mechanism to Prevent Structural Attack over Social Media

Abstract

Social media sites contain the personal information of the users, which entice the attackers. The attacker performs different types of attack on the social media site to get the users sensitive information. User privacy may be breached as other passive and active attacks are performed on social media sites; to prevent such a scenario, the network operator releases the data anonymized. Social media operators fetch and store social media users' data to share among various third-party consumers. The network operator releases the complete graph in anonymized and sanitized versions because the fetched information often contains sensitive data. But it does not provide a full guarantee of user privacy. This paper proposed a solution that provides a neighborhood adjacency matrix-based anonymization process for the social network graph. This anonymization process may be used to counter the neighborhood attack over the social network graph. The proposed anonymization process increases the number of isomorphic neighborhood networks by adding dummy edges in the social network graph. Therefore, a user may not be re-identified in a social network graph based on their unique neighborhood.