Anonymizing bipartite graph data using safe groupings

Graham Cormode, Divesh Srivastava, Ting Yu, Qing Zhang

Research output: Contribution to journalArticle

41 Citations (Scopus)

Abstract

Private data often come in the form of associations between entities, such as customers and products bought from a pharmacy, which are naturally represented in the form of a large, sparse bipartite graph. As with tabular data, it is desirable to be able to publish anonymized versions of such data, to allow others to perform ad hoc analysis of aggregate graph properties. However, existing tabular anonymization techniques do not give useful or meaningful results when applied to graphs: small changes or masking of the edge structure can radically change aggregate graph properties. We introduce a new family of anonymizations for bipartite graph data, called (k, ℓ)-groupings. These groupings preserve the underlying graph structure perfectly, and instead anonymize the mapping from entities to nodes of the graph. We identify a class of "safe" (k, ℓ)-groupings that have provable guarantees to resist a variety of attacks, and show how to find such safe groupings. We perform experiments on real bipartite graph data to study the utility of the anonymized version, and the impact of publishing alternate groupings of the same graph data. Our experiments demonstrate that (k, ℓ)-groupings offer strong tradeoffs between privacy and utility.

Original languageEnglish
Pages (from-to)115-139
Number of pages25
JournalVLDB Journal
Volume19
Issue number1
DOIs
Publication statusPublished - 1 Feb 2010
Externally publishedYes

Fingerprint

Experiments

Keywords

  • Graph
  • Microdata
  • Privacy
  • Query answering

ASJC Scopus subject areas

  • Hardware and Architecture
  • Information Systems

Cite this

Anonymizing bipartite graph data using safe groupings. / Cormode, Graham; Srivastava, Divesh; Yu, Ting; Zhang, Qing.

In: VLDB Journal, Vol. 19, No. 1, 01.02.2010, p. 115-139.

Research output: Contribution to journalArticle

Cormode, G, Srivastava, D, Yu, T & Zhang, Q 2010, 'Anonymizing bipartite graph data using safe groupings', VLDB Journal, vol. 19, no. 1, pp. 115-139. https://doi.org/10.1007/s00778-009-0167-9
Cormode, Graham ; Srivastava, Divesh ; Yu, Ting ; Zhang, Qing. / Anonymizing bipartite graph data using safe groupings. In: VLDB Journal. 2010 ; Vol. 19, No. 1. pp. 115-139.
@article{80b2af543d534bed9c0cdfa7fe630033,
title = "Anonymizing bipartite graph data using safe groupings",
abstract = "Private data often come in the form of associations between entities, such as customers and products bought from a pharmacy, which are naturally represented in the form of a large, sparse bipartite graph. As with tabular data, it is desirable to be able to publish anonymized versions of such data, to allow others to perform ad hoc analysis of aggregate graph properties. However, existing tabular anonymization techniques do not give useful or meaningful results when applied to graphs: small changes or masking of the edge structure can radically change aggregate graph properties. We introduce a new family of anonymizations for bipartite graph data, called (k, ℓ)-groupings. These groupings preserve the underlying graph structure perfectly, and instead anonymize the mapping from entities to nodes of the graph. We identify a class of {"}safe{"} (k, ℓ)-groupings that have provable guarantees to resist a variety of attacks, and show how to find such safe groupings. We perform experiments on real bipartite graph data to study the utility of the anonymized version, and the impact of publishing alternate groupings of the same graph data. Our experiments demonstrate that (k, ℓ)-groupings offer strong tradeoffs between privacy and utility.",
keywords = "Graph, Microdata, Privacy, Query answering",
author = "Graham Cormode and Divesh Srivastava and Ting Yu and Qing Zhang",
year = "2010",
month = "2",
day = "1",
doi = "10.1007/s00778-009-0167-9",
language = "English",
volume = "19",
pages = "115--139",
journal = "VLDB Journal",
issn = "1066-8888",
publisher = "Springer New York",
number = "1",

}

TY - JOUR

T1 - Anonymizing bipartite graph data using safe groupings

AU - Cormode, Graham

AU - Srivastava, Divesh

AU - Yu, Ting

AU - Zhang, Qing

PY - 2010/2/1

Y1 - 2010/2/1

N2 - Private data often come in the form of associations between entities, such as customers and products bought from a pharmacy, which are naturally represented in the form of a large, sparse bipartite graph. As with tabular data, it is desirable to be able to publish anonymized versions of such data, to allow others to perform ad hoc analysis of aggregate graph properties. However, existing tabular anonymization techniques do not give useful or meaningful results when applied to graphs: small changes or masking of the edge structure can radically change aggregate graph properties. We introduce a new family of anonymizations for bipartite graph data, called (k, ℓ)-groupings. These groupings preserve the underlying graph structure perfectly, and instead anonymize the mapping from entities to nodes of the graph. We identify a class of "safe" (k, ℓ)-groupings that have provable guarantees to resist a variety of attacks, and show how to find such safe groupings. We perform experiments on real bipartite graph data to study the utility of the anonymized version, and the impact of publishing alternate groupings of the same graph data. Our experiments demonstrate that (k, ℓ)-groupings offer strong tradeoffs between privacy and utility.

AB - Private data often come in the form of associations between entities, such as customers and products bought from a pharmacy, which are naturally represented in the form of a large, sparse bipartite graph. As with tabular data, it is desirable to be able to publish anonymized versions of such data, to allow others to perform ad hoc analysis of aggregate graph properties. However, existing tabular anonymization techniques do not give useful or meaningful results when applied to graphs: small changes or masking of the edge structure can radically change aggregate graph properties. We introduce a new family of anonymizations for bipartite graph data, called (k, ℓ)-groupings. These groupings preserve the underlying graph structure perfectly, and instead anonymize the mapping from entities to nodes of the graph. We identify a class of "safe" (k, ℓ)-groupings that have provable guarantees to resist a variety of attacks, and show how to find such safe groupings. We perform experiments on real bipartite graph data to study the utility of the anonymized version, and the impact of publishing alternate groupings of the same graph data. Our experiments demonstrate that (k, ℓ)-groupings offer strong tradeoffs between privacy and utility.

KW - Graph

KW - Microdata

KW - Privacy

KW - Query answering

UR - http://www.scopus.com/inward/record.url?scp=75949123863&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=75949123863&partnerID=8YFLogxK

U2 - 10.1007/s00778-009-0167-9

DO - 10.1007/s00778-009-0167-9

M3 - Article

AN - SCOPUS:75949123863

VL - 19

SP - 115

EP - 139

JO - VLDB Journal

JF - VLDB Journal

SN - 1066-8888

IS - 1

ER -