Abstract
Privacy preserving data publishing techniques focus on providing a sanitized view of a private data set to the recipients, e.g., government institutions, research organizations, statisticians, etc. The private data set contains the sensitive data about the individuals, e.g., hospital releases data about the patients for research or funding purposes. There exists ample work on preserving the privacy of individuals contained in these private data sets but when it comes to skewed data publication, these approaches fail to preserve the privacy of individuals. In this paper, we present a frequency based Anatomization technique termed as Skewed Data Anatomizer (SDA), for preserving the privacy of published sensitive attributes in skewed data. Our technique publishes two sets of data table. Grouping is used by which privacy is preserved and correlation of a large set of data is also captured from the microdata table. Extensive experimentation confirms that our technique is allowing more effective analysis of data than its counterparts.