Beyond multivariate microaggregation for large record anonymization

Producció científica: Capítol de llibreContribució a congrés/conferènciaAvaluat per experts

Resum

Microaggregation is one of the most commonly employed microdata protection methods. The basic idea of microaggregation is to anonymize data by aggregating original records into small groups of at least k elements and, therefore, preserving k-anonymity. Usually, in order to avoid information loss, when records are large, i.e., the number of attributes of the data set is large, this data set is split into smaller blocks of attributes and microaggregation is applied to each block, successively and independently. This is called multivariate microaggregation. By using this technique, the information loss after collapsing several values to the centroid of their group is reduced. Unfortunately, with multivariate microaggregation, the k-anonymity property is lost when at least two attributes of different blocks are known by the intruder, which might be the usual case. In this work, we present a new microaggregation method called one dimension microaggregation (Mic1D − k). With Mic1D − k, the problem of k-anonymity loss is mitigated by mixing all the values in the original microdata file into a single non-attributed data set using a set of simple pre-processing steps and then, microaggregating all the mixed values together. Our experiments show that, using real data, our proposal obtains lower disclosure risk than previous approaches whereas the information loss is preserved.

Idioma originalAnglès
Títol de la publicacióCitizen in Sensor Networks - 2nd International Workshop, CitiSens 2013, Revised Selected Papers
EditorsJordi Nin, Daniel Villatoro
EditorSpringer Verlag
Pàgines87-107
Nombre de pàgines21
ISBN (electrònic)9783319041773
DOIs
Estat de la publicacióPublicada - 2014
Publicat externament
Esdeveniment2nd International Workshop on Citizen in Sensor Networks, CitiSens 2013 - Barcelona, Spain
Durada: 19 de set. 201319 de set. 2013

Sèrie de publicacions

NomLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volum8313
ISSN (imprès)0302-9743
ISSN (electrònic)1611-3349

Conferència

Conferència2nd International Workshop on Citizen in Sensor Networks, CitiSens 2013
País/TerritoriSpain
CiutatBarcelona
Període19/09/1319/09/13

Fingerprint

Navegar pels temes de recerca de 'Beyond multivariate microaggregation for large record anonymization'. Junts formen un fingerprint únic.

Com citar-ho