Skip to Main content Skip to Navigation
Theses

Publication de données individuelles respectueuse de la vie privée : une démarche fondée sur le co-clustering

Résumé : There is a strong economic and civic demand for the opening of individual data. However, the publication of such data poses a risk to the individuals represented in it. This thesis focuses on the problem of anonymizing multidimensional data tables containing individual data for publishing purposes. In particular, two data anonymization approaches families will be focused on: the first aims to merge each individual into a group of individuals, the second is based on the addition of disruptive noise to the original data. Two new approaches are developed in the context of group anonymization. They aggregate the data using a co-clustering technique and then use the produced model, to generate synthetic records, in the case of the first solution. While the second proposal seeks to achieve the formalism of k-anonymity. Finally, we present a new anonymization algorithm “DPCocGen” that ensures differential privacy. First, a data-independent partitioning on the domains is used to generate a perturbed multidimensional histogram, a multidimensional co-clustering is then performed on the noisy histogram resulting in a partitioning scheme. Finally, the resulting schema is used to partition the original data in a differentially private way. Synthetic individuals can then be drawn from the partitions.
Complete list of metadatas

Cited literature [146 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-02053043
Contributor : Marc Gelgon <>
Submitted on : Thursday, February 28, 2019 - 10:05:33 PM
Last modification on : Wednesday, June 24, 2020 - 4:19:33 PM
Document(s) archivé(s) le : Wednesday, May 29, 2019 - 5:33:08 PM

File

these_Tarek_Benkhelif.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-02053043, version 1

Citation

Tarek Benkhelif. Publication de données individuelles respectueuse de la vie privée : une démarche fondée sur le co-clustering. Cryptographie et sécurité [cs.CR]. Université de Nantes, 2018. Français. ⟨tel-02053043⟩

Share

Metrics

Record views

132

Files downloads

342