Geometric Data Perturbation for Privacy Preserving in Data Stream Mining
Mayur Prajapati, Aniket Patel · Zenodo (CERN European Organization for Nuclear Research) · 2018
Today as we have tendency to live within the era of information explosion. It's become important to search for helpful data from large dataset. Additionally advance in web communication and hardware technology has lead to raise within the capability of storing personal information of people. Huge quantity of data stream are generated from completely different applications like shopping record, medical, network traffic etc. Sharing such type of information is incredibly important plus to business decision but the worry is that when the non-public information is leaked it may be abused for a different purposes. Therefore some quantity of privacy preserving must be done on the information before it is free to others. Ancient ways of Privacy Preserving Data Mining (PPDM) area unit designed for static information sets that makes its unsuitable for dynamic data streams. In this paper an economical and effective information perturbation methodology is proposed that aims to protect the privacy of sensitive attributes and obtaining information bunch with minimum information loss.