Indian Journal of Science and Technology

Abstract

Background/Objectives: This paper proposes an Optimized Social Ant Based Sensitive Item Hiding (OSA-SIH) technique and expands the scope of quality privacy preservation for distributed data mining with optimal side effects on the original dataset. Methods/Statistical Analysis: in OSA-SIH technique, initially sensitive items for the given distributed dataset are evaluated using the social ant based relative item set distribution. Based on the evaluated dataset, optimal hiding of sensitive item is arrived with social ant based relative item set distribution even for larger item sets, ensuring time for optimal hiding. Next, sensitive item hiding is performed through multiplicative and transformational data perturbation. This data perturbation is based on socially cohesive relational rate between sensitive and non sensitive item sets, ensuring privacy preservation accuracy. The side effects on the modified dataset are checked for several users’ requested item set distribution. Findings: The experimental results demonstrated that proposed technique out performed than the existing state of the art works in terms of privacy preservation accuracy, rate of side effects on the modified dataset, and time for optimal hiding. Improvement/Application: Experiments revealed that the proposed OSA-SIH technique is able to reduce the rate of side effects on modified dataset as compared to the state-of-the-art works. 

Keywords: Perturbation, Privacy Preserving Data Mining, Social Ant, Sensitive Item Hiding, Transformational Data Perturbation, Multiplicative Data Perturbation