International Journal of Engineering and Technology

As the amount of data available and shared in the electronic media increases, the threat for its privacy and security also increases. Every organization publishes its data to many recipients for various reasons .Thus preserving the privacy of data during the process of data mining is the aim of Privacy Preserving Data Mining [PPDM]. Multi-trust level is a scenario in privacy preserving data mining where different versions of privacy preserved data are distributed to the users based on their trust level. This work presents three trigonometric based rotation perturbation algorithms for privacy preservation at multi-trust level. These algorithms are applied on three bench marked datasets to generate multiple sequential dataset versions at various privacy levels. The perturbed datasets are evaluated on utility, distortion rate and the ability to prevent linking attacks. The results show that the perturbed datasets have utility comparable to the original datasets and linking attacks are prevented.