Title: Optimised K-anonymisation technique to deal with mutual friends and degree attacks

Authors: Amardeep Singh; Monika Singh; Divya Bansal; Sanjeev Sofat

Addresses: Computer Science and Engineering, Punjab Engineering College, Chandigarh, India ' Computer Science and Engineering, Punjab Engineering College, Chandigarh, India ' Computer Science and Engineering, Punjab Engineering College, Chandigarh, India ' Computer Science and Engineering, Punjab Engineering College, Chandigarh, India

Abstract: Online social networks have become a predominant service on the web, collecting the huge amount of users' information. It is drastically revolutionising the way people interact with each other. Publishing data of social network users for researchers, academicians, advertising organisations, etc. has raised many serious privacy implications. Lots of techniques have been proposed for preserving the privacy of individuals handling different types of attack scenarios used by adversaries. In this paper, we address a new attack model, i.e., mutual friends attack model, in which an adversary can identify the victim nodes by using information about the number of their mutual friends. An algorithm 'optimised K-anonymisation' has been devised that can deal with two types of attacks, i.e., degree attacks and the number of mutual friends attacks. The experimental results illustrate that our proposed algorithm can preserve the identification of individuals and subsequently maintain the utility of data.

Keywords: privacy preserving; social networks; degree attacks; mutual friends attacks; K-anonymisation; Twitter; APL; information loss.

DOI: 10.1504/IJICS.2021.114706

International Journal of Information and Computer Security, 2021 Vol.14 No.3/4, pp.281 - 299

Accepted: 04 Jun 2018
Published online: 04 May 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article