A non-grouping anonymity model for preserving privacy in health data publishing

Abstract

Publishing health data may jeopardize privacy breaches, since they contain sensitive information about the individuals. Privacy preserving data publishing (PPDP) addresses the problem of revealing sensitive data when extracting the useful data. The existing privacy models are group based anonymity models. Hence, these models consider the privacy of the individual only in a group based manner. And those groups are the hunting ground for the adversaries. All data re-identification attacks are based on the group of records. The root cause behind our approach is that the k-anonymity problem can be viewed as a clustering approach. Though the k-anonymity problem does not insist on the number of clusters, it requires that each group must contain at least k-records. We propose a Non-Grouping Anonymity model; this gives a basic level of anonymization that prevents an individual being re-identified from their published data.