Visible to the public "Discovery of De-identification Policies Considering Re-identification Risks and Information Loss"Conflict Detection Enabled

Title"Discovery of De-identification Policies Considering Re-identification Risks and Information Loss"
Publication TypeConference Paper
Year of Publication2015
AuthorsH. M. Ruan, M. H. Tsai, Y. N. Huang, Y. H. Liao, C. L. Lei
Conference Name2015 10th Asia Joint Conference on Information Security
Date PublishedMay
ISBN Number978-1-4799-1989-5
Accession Number15293307
KeywordsComputational modeling, Data analysis, data privacy, de-identification, deidentification policies, deidentified data, Entropy, HIPPA, information entropy, information loss, Lattices, learning (artificial intelligence), privacy, privacy leakage, pubcrawl170105, reidentification risks, risk analysis, Safe Harbor, Synthetic aperture sonar, UCI machine learning repository, Upper bound

In data analysis, it is always a tough task to strike the balance between the privacy and the applicability of the data. Due to the demand for individual privacy, the data are being more or less obscured before being released or outsourced to avoid possible privacy leakage. This process is so called de-identification. To discuss a de-identification policy, the most important two aspects should be the re-identification risk and the information loss. In this paper, we introduce a novel policy searching method to efficiently find out proper de-identification policies according to acceptable re-identification risk while retaining the information resided in the data. With the UCI Machine Learning Repository as our real world dataset, the re-identification risk can therefore be able to reflect the true risk of the de-identified data under the de-identification policies. Moreover, using the proposed algorithm, one can then efficiently acquire policies with higher information entropy.

Citation Key7153938