Model for a Common Notion of Privacy Leakage on Public Database

Shinsaku Kiyomoto¹ and Keith M. Martin²

¹KDDI R&D Laboratories, Inc.

2-1-15 Ohara, Fujimino, Saitama

356-8502, Japan

kiyomoto@kddilabs.jp

²Information Security Group,

Royal Holloway University of London

Egham, Surrey TW20 0EX, UK

keith.martin@rhul.ac.uk

Abstract

Privacy is an increasingly important aspect of data publishing services.

If personal private information is leaked from the data, the service will be regarded
as unacceptable by the original owners of the data. Two different approaches
to defining a notion of database privacy, the generalization method and the perturbation method,
have been independently studied. These two approaches have significantly differences,
making it hard to compare related research. In this paper, we propose a unified model
that is based on the perturbation method, but which is applicable to generalized data sets.

In particular, this model applies the notion of differential privacy to data sets that satisfy k-anonymity.

We demonstrate this approach through a simple case study. This is a first step towards a common

notion for protecting database privacy

Keywords: privacy leakage, database privacy, perturbation method

Journal of Wireless Mobile Networks, Ubiquitous Computing, and Dependable Applications (JoWUA),

Vol. 2, No. 1, pp. 50-62, June 2011 [pdf]