k-Degree Anonymity Model for Social Network Data Publishing

Abstract

Publicly accessible platform for social networking has gained special attraction because of its easy data sharing. Data generated on such social network is analyzed for various activities like marketing, social psychology, etc. This requires preservation of sensitive attributes before it becomes easily accessible. Simply removing the personal identities of the users before publishing data is not enough to maintain the privacy of the individuals. The structure of the social network data itself reveals much information regarding its users and their connections. To resolve this problem, k-degree anonymous method is adopted. It emphasizes on the modification of the graph to provide at least k number of nodes that contain the same degree. However, this approach is not efficient on a huge amount of social data and the modification of the original data fails to maintain data usefulness. In addition to this, the current anonymization approaches focus on a degree sequence-based graph model which leads to major modification of the graph topological properties. In this paper, we have proposed an improved k-degree anonymity model that retain the social network structural properties and also to provide privacy to the individuals. Utility measurement approach for community based graph model is used to verify the performance of the proposed technique.

Keywords

• data privacy
• data processing
• publishing
• social network services
• utility programs