Survey on Recent Developments in Privacy Preserving Models

Abstract

Privacy preserving in data mining [1] is one of the major and increasingly interested area of research under data security. Privacy will be provided for data at different levels such as, while publishing the data, at the time of retrieving result by preserving sensitive data without disclosing it. It is not just sufficient to preserve sensitive data without disclosing it, but also need to manipulate and present data so that, certain inference channels are blocked. Numbers of techniques are proposed to achieve privacy protection for sensitive data. But, most of these methods are facing side effects such as reduced utility, less accuracy, data mining efficiency down-graded, disclosure risk, etc. In this paper we analyze all these different techniques how they handle data in turn to provide privacy and points out their merits and demerits.