Analysis of privacy preserving echanisms for outsourced data mining

Abstract

The emergence of data mining techniques have revolutionized the information-centric world. New tools and techniques have been announced in a quick session. In recent years, the amount of digital data collected from various resources has increased tremendously. Multiple organizations and individuals having fewer computation resources and lack of expertise can outsource their data mining jobs to the third party service provider/server. The Data-Mining-as-a-Service (DMaaS) paradigm is steadily gaining impetus. Privacy and security issues are the primary concern of DMaaS, as the third party is assumed to be semi-trusted. In this scenario, the data owner may not want to share its sensitive data either with the server or other data owners. In this paper, we propose a cloud-aided privacy preserving data mining solution for outsourced data in the multi-party environment with less leakage of raw data. Our solution is designed for applications, where high-level privacy-preservation is required by the data owners.