Back to search

University of Waterloo

Differential Privacy for Nearest Neighbor Queries

Abstract

dc:description.abstract

We examine the problem of providing differential privacy for nearest neighbor queries. Very few mechanisms exist that achieve this, most notable geo-indistinguishability in the context of location privacy. However it uses an extended definition of differential privacy and restricts the sensitivity of queries. This work presents a new mechanism for DP nearest neighbor queries that is general to many applications and is based on tree data-structures and traversal. The biggest challenge with existing local differential private solutions is poor utility, requiring the addition of a restriction on the sensitivity of queries. We provide two variations, one which uses a similar restriction and one that does not. We explore different tree traversal algorithms. We evaluate our method on artificial datasets as well as real world location data. The results show that the variant using a restricted sensitivity does not perform better than geo-indistinguishability, while the unrestricted variant offers a method with good utility.

Degree

thesis:*
Grantor dc:publisher
University of Waterloo
Year dc:date.issued
2022

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Lepert, Emily
Advisor dc:contributor.advisor
  • Kerschbaum, Florian

Subjects

dc:subject × 4

Rights

Language dc:language.iso
en

Identifiers

dc:identifier.*
Handle dc:identifier.uri
http://hdl.handle.net/10012/18550

Chain of custody

source
Harvested from
University of Waterloo
Base URL
dspacemainprd01.lib.uwaterloo.ca/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Lepert, Emily. Differential Privacy for Nearest Neighbor Queries. University of Waterloo, 2022. http://hdl.handle.net/10012/18550