Multiple Sensitive Attributes Based Privacy Preserving Data Publishing

Abstract

The advances in digital information applications facilitates the collection of huge amount of data about governments, healthcare, other organizations and individuals. To make this data available for researchers, businesses and other users, it needs to be released. This in turn increases the demand of exchanging and publishing this collected data. However, data in its original form, typically contains sensitive information about individuals and/or organizations, and publishing such data will violate individual or organizational privacy. Hence, Privacy Preserving Data Publishing (PPDP) provides methods and tools for publishing useful information while preserving data privacy. Before data is published to the concerned users, it is altered to maintain its privacy without compromising data utility, using various anonymization techniques. Real-time datasets contain different types of Multiple Sensitive Attributes (MSAs) (which could be numerical or categorical). Anonymization for only Single Sensitive Attribute is not suitable for functional usage. Thus, it is important to maintain the association between these MSAs and to preserve the privacy of Mixed (numerical and categorical) MSAs efficiently while working with high dimensional data. The main focus of this paper is to analyse the different schemes proposed in literature for PPDP of MSAs.