Privacy Preserving in Data Stream Mining using Geometric Data Perturbation Approach

Today as we are living in the era of information explosion, it has become very important to find out useful information from large massive data. Also advances in internet, communication and hardware technology has lead to an increase in the capability of storing personal data of individuals. Massive amount of data streams are generated from different applications like medical, shopping record, network traffic, etc. Sharing such data is very important asset to business decision making but the fear is that once the personal data is leaked it can be misused for a variety of purposes. Hence some amount of privacy preserving needs to be done on the data before it is released to others. Traditional methods of privacy preserving data mining (PPDM) are designed for static data sets which makes it unsuitable for dynamic data streams. In this paper an efficient and effective data perturbation method is proposed that aims to protect privacy of sensitive attribute and obtaining data clustering with minimum information loss.