On the Privacy of Euclidean Distance Preserving Data Perturbations

We examine Euclidean distance preserving data perturbation as a tool for privacy-preserving data mining. Such perturbations allow many important data mining algorithms, with only minor modification, to be applied to the perturbed data and produce exactly the same results as if applied to the original data, e.g., hierarchical cluatering and k-means clustering. However, the issue of how well the original data is hidden needs careful study. We take a step in this direction by assuming the role of an attacker armed with two types of prior information regarding the original data. We examine how well the attacker can recover the original data from the perturbed data and prior information. Our results offer insight into the vulnerabilities of Euclidean distance preserving transformations.
Date: October 18, 2008
Book Title: IEEE Transactions on Knowledge and Data Engineering
Type: InProceedings


  author = "Kun Liu and Chris Giannella and Hillol Kargupta",
  title = "{On the Privacy of Euclidean Distance Preserving Data Perturbations}",
  month = "October",
  year = "2008",
  booktitle = "IEEE Transactions on Knowledge and Data Engineering",