A tutorial on blocking methods for privacy-preserving record linkage

Dimitrios Karapiperis, Vassilios S. Verykios, Eleftheria Katsiri, Alex Delis

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper, we first present five state-of-the-art private blocking methods which rely mainly on random strings, clustering, and public reference sets. We emphasize on the drawbacks of these methods, and then, we present our L-fold redundant blocking scheme, that relies on the Locality-Sensitive Hashing technique for identifying similar records. These records have undergone an anonymization transformation using a Bloom filter-based encoding technique. Finally, we perform an experimental evaluation of all these methods and present the results.

Original languageEnglish (US)
Title of host publicationAlgorithmic Aspects of Cloud Computing - 1st International Workshop, ALGOCLOUD 2015, Revised Selected Papers
EditorsPeter Triantafillou, Ioannis Karydis, Spyros Sioutas, Dimitrios Tsoumakos
PublisherSpringer Verlag
Pages3-15
Number of pages13
ISBN (Print)9783319299181
DOIs
StatePublished - 2016
Event1st International Workshop on Algorithmic Aspects of Cloud Computing, ALGOCLOUD 2015 - Patras, Greece
Duration: Sep 14 2015Sep 15 2015

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume9511
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference1st International Workshop on Algorithmic Aspects of Cloud Computing, ALGOCLOUD 2015
CountryGreece
CityPatras
Period9/14/159/15/15

Keywords

  • Blocking
  • Bloom filter
  • Locality-sensitive hashing

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Fingerprint Dive into the research topics of 'A tutorial on blocking methods for privacy-preserving record linkage'. Together they form a unique fingerprint.

Cite this