Border Revision for Knowledge Hiding

Abstract

In this chapter, we highlight the process of border revision, which plays a significant role on both border-based and exact approaches for association rule hiding. Simply stated, the process of border revision captures those itemsets of the original database which need to remain frequent and those that need to become infrequent in the sanitized version of the database, in order to allow furnishing an optimal hiding solution. After presenting the theory behind border revision, we introduce a set of algorithms (adopted from [27]) that can be applied for the efficient computation of both the original and the revised borders in a transactional database, when given a minimum support threshold.