Outsourced Database Security - Inference attacks on statistical databases represent a complex issue in institutions and corporates since it is hard to detect and prevent, especially when it is committed by an internal adversary. The issue has been manifested further with the widespread of data analytics techniques in industry and academia, besides outsourced services. Even when the released statistical data has been anonymized and the identifying attributes are removed, targeted individuals can be spotted in such data. Therefore, preventing sensitive statistical data leakage is crucial for protecting the privacy of individuals or events, but such measures should not form utilization obstacles or degrade the data utility. This paper proposes an antiinference technique for preserving the privacy of sensitive data in statistical databases. Unlike existing solutions, which either require considerable computing resources or trade-off between statistical data accuracy and its privacy, our solution is designed to maintain the accuracy while privacy is ensured.

jan

Tabuk, Saudi Arabia

https://ieeexplore.ieee.org/document/9711549/