Protecting the publishing identity in multiple tuples

Abstract

Current privacy preserving methods in data publishing always remove the individually identifying attribute first and then generalize the quasi-identifier attributes. They cannot take the individually identifying attribute into account. In fact, tuples will become vulnerable in the situation of multiple tuples per individual. In this paper, we analyze the individually identifying attribute in the privacy preserving data publishing and propose the concept of identity-reserved anonymity. We develop two approaches to meet identity-reserved anonymity requirement. The algorithms are evaluated in an experimental scenario, demonstrating practical applicability of the approaches. © 2008 Springer-Verlag Berlin Heidelberg.