Full Citation

Title: Clustering-Based k-Anonymity

Citation Type: Journal Article

Publication Year: 2012

ISBN:

ISSN:

DOI:

NSFID:

PMCID:

PMID:

Abstract: Privacy is one of major concerns when data containing sensitive information needs to be released for ad hoc analysis, which has attracted wide research interest on privacy-preserving data publishing in the past few years. One approach of strategy to anonymize data is generalization. In a typical generalization approach, tuples in a table was first divided into many QI (quasi-identifier)-groups such that the size of each QI-group is no less than k. Clustering is to partition the tuples into many clusters such that the points within a cluster are more similar to each other than points in different clusters. The two methods share a common feature: distribute the tuples into many small groups. Motivated by this observation, we propose a clustering-based k-anonymity algorithm, which achieves k-anonymity through clustering. Extensive experiments on real data sets are also conducted, showing that the utility has been improved by our approach.

User Submitted?: No

Authors: Chen, Yefang; Dong, Yihong; Chen, Huahui; Wang, Peng; He, Xianmang; Huang, Zhenhua

Periodical (Full): Lecture Notes in Artificial Intelligence

Issue:

Volume: 7301

Pages: 405 - 417

Data Collections: IPUMS USA

Topics:

Countries: