Enhanced flip and additive rotation perturbation approaches for privacy preserving data mining

Abstract

In the digital era, data is moving around the world in a rapid way, creating huge vulnerabilities to sensitive and private information during the mining process. Privacy issues on the web are based on the fact that most users want to maintain strict anonymity on web applications and activities. Privacy Preserving Data Mining (PPDM) methods are evolved to share sensitive data with external parties to facilitate data analysis by balancing utility and privacy. newlineThe main objective of this work is to preserve the data efficiently with less computation time, better privacy and utility with minimal information loss. Perturbation processes are performed with the additive and flip rotations after condensation in a streaming manner. Perturbed data are classified by fast, scalable and efficient classification algorithms. newlineIn the first part of the work, Additive Rotation Perturbation (ARP) and Flip Rotation Perturbation (FRP) schemes are applied effectively on HIggs DataSet (HIDS), Letter Recognition DataSet (LRDS), Heart Disease DataSet (HDDS) and Page Block DataSet (PBDS) after performing Fuzzy C Means (FCM) clustering. Naïve newline