Known sample attacks on relation preserving data transformations

Official URL: http://dx.doi.org/10.1109/TDSC.2017.2759732

Many data mining applications such as clustering and $k$k-NN search rely on distances and relations in the data. Thus, distance preserving transformations, which perturb the data but retain records' distances, have emerged as a prominent privacy protection method. In this paper, we present a novel attack on a generalized form of distance preserving transformations, called relation preserving transformations. Our attack exploits not the exact distances between data, but the relationships between the distances. We show that an attacker with few known samples (4 to 10) and direct access to relations can retrieve unknown data records with more than 95 percent precision. In addition, experiments demonstrate that simple methods of noise addition or perturbation are not sufficient to prevent our attack, as they decrease precision by only 10 percent.