An Approach to Mitigate Composition Attacks on Privacy in Non-coordinated Environments

Nowadays individual's private and confidential information is available at hospitals, banks, or any other organization. Organizations share that information in a manner to protect individuals' privacy. Many well-known techniques have been proposed to do that such as k-anonymity, l-diversity, etc. But the majority fails to protect privacy when one patient goes to two hospitals for medical care; and his/her confidential information disclosed by tracking two disparate published data sets with a composition attack. Though this attack can be easily mitigated when organizations coordinate prior data publication. However, such coordination is not always possible. In this paper, we introduce a model that uses a public knowledge domain to reduce the risk of composition attacks without coordination. The result shows that our approach reduces the rate of composition attack compared with other well-known models.