L-diversity
From Cohen Courses
Revision as of 10:31, 6 November 2012 by Zsheikh (talk | contribs) (Created page with 'An adversary can easily discover the values of sensitive attributes in a k-anonymized dataset if there is little diversity in the sensitive attributes. l-diversit…')
An adversary can easily discover the values of sensitive attributes in a k-anonymized dataset if there is little diversity in the sensitive attributes. l-diversity is a new definition of privacy that takes care of this problem. A set of records in an equivalence class C is l-diverse if it contains at least l “well-represented” values for each sensitive attribute.
Related papers
A. Machanavajjhala, J. Gehrke, D. Kifer, and M. Venkitasubramaniam. l-diversity: Privacy beyond k-anonymity. In Proc. 22nd Intnl. Conf. Data Engg. (ICDE), 2006. pdf
