An Improved Algorithm for Completely Hiding Sensitive Association Rule Sets
Shan-Tai Chen, Shih‐Min Lin, Chi-Yii Tang, Guei-Yu Lin · 2009
Association analysis is a powerful technology for discovering hidden information in huge databases. Nowadays, strategic alliances often share their information for exploring more useful decision-making information and for enhancing their core competitiveness. However, information sharing will potentially involve threats that privacy or sensitive information could be unconsciously divulged. In this study, we propose a novel algorithm, ADSSI, for privacy preservation against association rule analysis. Through entirely analyzing all types of frequent items, ADSSI reduces the risk of sensitive information leakage by slightly modifying transactions in a database. ADSSI is able to completely hide all sensitive association rule sets while no new rules generated in the shared database. Moreover, compared to the current-best works, experimental results show that ADSSI outperforms other published algorithms in terms of side effects generated.