Loose associations to increase utility in data publishing

• DOI: 10.3233/jcs-140513
• Publication Date: 2016-05-18T07:46:57Z
• Authors: Sabrina De Capita...; Sara Foresti; Sushil Jajodia; Giovanni Livraga; Stefano Paraboschi; Pierangela Samarati
• Keywords: Heuristics; Fragmentation; Association (psychology); Information loss; Information sensitivity

Data fragmentation has been proposed as a solution for protecting the confidentiality of sensitive associations when releasing data publishing or external storage. To enrich utility fragments, recent approach put forward idea complementing pair fragments with some (non-precise, hence loose) information on association between them. Starting from observation that in presence multiple publication several independent pairs can cause improper leakage information, this paper we extend loose to operate over an arbitrary number fragments. We first illustrate how different potentially expose associations, and describe defining among set investigate tuples be grouped producing so increase queries executed then provide heuristics performing such grouping satisfying given level protection while achieving also result extensive experimental effort both synthetic real datasets, which shows efficiency enhanced provided by our proposal.