Title: A cluster-based approach for distributed anonymisation of vertically partitioned data

Authors: Antonios Xenakis; Zhiyuan Chen; George Karabatis

Addresses: Department of Information Systems, University of Maryland, Baltimore County (UMBC), Baltimore, MD, USA ' Department of Information Systems, University of Maryland, Baltimore County (UMBC), Baltimore, MD, USA ' Department of Information Systems, University of Maryland, Baltimore County (UMBC), Baltimore, MD, USA

Abstract: In modern organisations, data is often spread across different sites, posing challenges for effective analysis. Transferring data to a centralised server may jeopardise privacy and leak sensitive/proprietary information. Therefore, organisations hesitate adopting this solution despite its potential to fully utilise, and analyse the data, for better decision making. Current approaches concentrate on distributed privacy-preserving techniques for data analysis, where data does not leave each site, but incurs substantial computational and communication overhead. This paper focuses on distributed data that is anonymised on site, then merged and sent to a centralised server for analysis. Two new approaches on cluster-based distributed anonymisation are introduced for vertically partitioned data, one based on distributed coordinated anonymisation, and the other based on top-down distributed anonymisation, resulting in low initial onsite anonymisation overhead. Experiments show these approaches preserve data privacy with very minor loss of utility of anonymised data and impose minimal computational overhead.

Keywords: privacy; distributed anonymisation; differential privacy; K-anonymity; cluster-based anonymisation.

DOI: 10.1504/IJWET.2024.143360

International Journal of Web Engineering and Technology, 2024 Vol.19 No.4, pp.397 - 420

Received: 23 Oct 2023
Accepted: 03 Jan 2024
Published online: 16 Dec 2024 *

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