Anonymizing Transaction Data for Publication
Young-Hoon Kim, Hyoung-Min Park, Kyuseok Shim · 2011
Transaction data, which is a table of item sets where each item set is associated with an individual, is very common in databases such as basket data and query log in search engine. When a table containing individual data is published, disclosure of sensitive information should be prohibitive. Since simply removing identifiers such as name and social security number may reveal the sensitive information by linking attacks which join the published table with other public tables on some sets of items, several privacy preserving models such as k-anonymity and 1-diversity are proposed, and anonymization algorithms are also suggested previously. In this paper, we propose a novel privacy preserving model that prohibits the linking attacks with the information of absent items, which we call (h,k,p,n)-coherence, and suggest an approximation algorithm that guarantees (h,k,p,n)-coherence using item generalization and transaction appending. Experimental results confirm that our approximation algorithm performs significantly better than traditional approximation algorithms.