Title: Rule based anonymisation against inference attack in social networks
Authors: Nidhi Desai; Manik Lal Das
Addresses: DA-IICT, Gandhinagar – 382007, India ' DA-IICT, Gandhinagar – 382007, India
Abstract: Social networks information has evolved as a powerful decision-maker in various facets of society. Meticulous scrutinisation of gigantic volumes of social data provides insightful solutions for forecasting, government policies, societal problems, business and strategic goals. The presence of sensitive information makes social network's information vulnerable to privacy concerns, leading to users' resistance to sharing their information confidently. Recently, inference attack using rule-based mining technique has posed a challenging privacy concern, particularly in social networks. This paper presents rule anonymity, a privacy model against inference attack using rule-based mining techniques in social networks. The proposed model considers adversary with strong knowledge of rule generation. The proposed rule-based anonymisation technique incorporates Rule Anonymity and ensures a strong privacy guarantee against an adversary with rule mining capability. The experimental results of the proposed anonymisation technique manifest the idea of rule anonymity on the social dataset.
Keywords: social networks; data privacy; rule mining; inference attack; anonymity; privacy-preserving; social networks data publishing; adversarial model.
International Journal of Social Computing and Cyber-Physical Systems, 2021 Vol.2 No.3, pp.212 - 228
Received: 30 May 2020
Accepted: 27 Sep 2020
Published online: 05 Oct 2021 *