Towards privacy preserving information retrieval through semantic microaggregation
Publication Type:
Conference ProceedingsSource:
Web Intelligence and Intelligent Agent Technology (WI-IAT), 2010 IEEE/WIC/ACM International Conference on, IEEE, Volume 3, Toronto, Canada, p.296 - 299 (2010)ISBN:
978-1-4244-8482-9URL:
http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=5614556Keywords:
Web indexing task; classification task; frequency term vector; k-anonymity preservation; privacy preserving information retrieval; semantic microaggregation; Internet; data privacy; information retrieval; pattern classification; vectorsAbstract:
In this paper we introduce the problem of providing privacy preserving information for Web indexing, classification, and other information retrieval task. Web pages are represented by a frequency term vector that preserves k-anonymity for all the Web pages. This vector can then be used, for example, to build indexes of classifiers. Our proposal makes use of semantic micro aggregation.
Towards Semantic Microaggregation of Categorical Data for Confidential Documents
Publication Type:
Conference ProceedingsSource:
Modeling Decisions for Artificial Intelligence, MDAI 2010, Springer, Volume 6408, Perpignan, France, p.266-276 (2010)URL:
http://www.springerlink.com/content/f41402862155w6t4/Keywords:
Web indexing task; classification task; frequency term vector; k-anonymity preservation; privacy preserving information retrieval; semantic microaggregation; Internet; data privacy; information retrieval; pattern classification; vectorsAbstract:
In the data privacy context, specifically, in statistical disclosure control techniques, microaggregation is a well-known microdata protection method, ensuring the confidentiality of each individual. In this paper, we propose a new approach of microaggregation to deal with semantic sets of categorical data, like text documents. This method relies on the WordNet framework that provides complete semantic relationship taxonomy between words. Therefore, this extension aims ensure the confidentiality of text documents, but at the same time, it should preserve the general meaning. We apply some measures to evaluate the quality of the protection method relying on information loss.
