Please use this identifier to cite or link to this item:

http://hdl.handle.net/10609/61486
Title: Triangulation-Based Multivariate Microaggregation
Author: Machín Casañas, Juvenal
Director: Solana Gómez, Agustí
Keywords: microaggregation, microdata protection, Statistical disclosure control, privacy, triangulation
Issue Date: 2-Jan-2017
Publisher: Universitat Oberta de Catalunya
Abstract: Microaggregation is a Statistical Disclosure Control technique in which similar records are clustered into groups containing a minimum of k records that are later replaced by group centroids, so that released data preserve some of their statistical properties while reducing the risk of re-identification. A fixed-size microaggregation method clusters data into groups of size k except perhaps one group with size between k and 2k -1, whereas a data-oriented (variable-size) method allows group size to vary between k and 2k -1. Heuristic clustering methods are needed since the minimum information loss microaggregation problem is NP-hard. In this paper we studied various microaggregation methods in the literature and we have proposed a new heuristic approach for multivariate fixed-size microaggregation based on the triangulation of a set of points in R^2. A reference data set and a random generated one are used to compare the method outcomes, in terms of information loss, with other previous proposals and the results summarized.
Language: English
URI: http://hdl.handle.net/10609/61486
Appears in Collections:Bachelor thesis, research projects, etc.

Share:
Export:
Files in This Item:
File Description SizeFormat 
jmachin0117Triangulation-Based Multivariate Microaggregation.pdfTriangulation-Based Multivariate Microaggregation419.61 kBAdobe PDFView/Open

This item is licensed under a Creative Commons License Creative Commons