Abstract

Privacy leakage is an important issue for social relationships-based recommender systems (i.e., social recommendation). Existing privacy preserving social recommendation approaches usually allow the recommender to fully control users’ information. This may be problematic since the recommender itself may be untrusted, leading to serious privacy leakage. Besides, building social relationships requires sharing interests as well as other private information, which may lead to more privacy leakage. Although sometimes users are allowed to hide their sensitive private data using personalized privacy settings, the data being shared can still be abused by the adversaries to infer sensitive private information. Supporting social recommendation with least privacy leakage to untrusted recommender and other users (i.e., friends) is an important yet challenging problem. In this paper, we aim to achieve privacy-preserving social recommendation under personalized privacy settings. We propose PrivSR, a novel privacy-preserving social recommendation framework, in which user can model user feedbacks and social relationships privately. Meanwhile, by allocating different noise magnitudes to personalized sensitive and non-sensitive feedbacks, we can protect users’ privacy against untrusted recommender and friends. Theoretical analysis and experimental evaluation on real-world datasets demonstrate that our framework can protect users’ privacy while being able to retain effectiveness of the underlying recommender system.

Original languageEnglish (US)
Pages (from-to)1-29
Number of pages29
JournalWorld Wide Web
DOIs
StateAccepted/In press - Jul 14 2018

Fingerprint

Recommender systems
Feedback

Keywords

  • Differential privacy
  • Personalized privacy settings
  • Ranking
  • Social recommendation

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this

Towards privacy preserving social recommendation under personalized privacy settings. / Meng, Xuying; Wang, Suhang; Shu, Kai; Li, Jundong; Chen, Bo; Liu, Huan; Zhang, Yujun.

In: World Wide Web, 14.07.2018, p. 1-29.

Research output: Contribution to journalArticle

Meng, Xuying ; Wang, Suhang ; Shu, Kai ; Li, Jundong ; Chen, Bo ; Liu, Huan ; Zhang, Yujun. / Towards privacy preserving social recommendation under personalized privacy settings. In: World Wide Web. 2018 ; pp. 1-29.
@article{6fa4b53eb136498ea8393ca9b45d8611,
title = "Towards privacy preserving social recommendation under personalized privacy settings",
abstract = "Privacy leakage is an important issue for social relationships-based recommender systems (i.e., social recommendation). Existing privacy preserving social recommendation approaches usually allow the recommender to fully control users’ information. This may be problematic since the recommender itself may be untrusted, leading to serious privacy leakage. Besides, building social relationships requires sharing interests as well as other private information, which may lead to more privacy leakage. Although sometimes users are allowed to hide their sensitive private data using personalized privacy settings, the data being shared can still be abused by the adversaries to infer sensitive private information. Supporting social recommendation with least privacy leakage to untrusted recommender and other users (i.e., friends) is an important yet challenging problem. In this paper, we aim to achieve privacy-preserving social recommendation under personalized privacy settings. We propose PrivSR, a novel privacy-preserving social recommendation framework, in which user can model user feedbacks and social relationships privately. Meanwhile, by allocating different noise magnitudes to personalized sensitive and non-sensitive feedbacks, we can protect users’ privacy against untrusted recommender and friends. Theoretical analysis and experimental evaluation on real-world datasets demonstrate that our framework can protect users’ privacy while being able to retain effectiveness of the underlying recommender system.",
keywords = "Differential privacy, Personalized privacy settings, Ranking, Social recommendation",
author = "Xuying Meng and Suhang Wang and Kai Shu and Jundong Li and Bo Chen and Huan Liu and Yujun Zhang",
year = "2018",
month = "7",
day = "14",
doi = "10.1007/s11280-018-0620-z",
language = "English (US)",
pages = "1--29",
journal = "World Wide Web",
issn = "1386-145X",
publisher = "Springer New York",

}

TY - JOUR

T1 - Towards privacy preserving social recommendation under personalized privacy settings

AU - Meng, Xuying

AU - Wang, Suhang

AU - Shu, Kai

AU - Li, Jundong

AU - Chen, Bo

AU - Liu, Huan

AU - Zhang, Yujun

PY - 2018/7/14

Y1 - 2018/7/14

N2 - Privacy leakage is an important issue for social relationships-based recommender systems (i.e., social recommendation). Existing privacy preserving social recommendation approaches usually allow the recommender to fully control users’ information. This may be problematic since the recommender itself may be untrusted, leading to serious privacy leakage. Besides, building social relationships requires sharing interests as well as other private information, which may lead to more privacy leakage. Although sometimes users are allowed to hide their sensitive private data using personalized privacy settings, the data being shared can still be abused by the adversaries to infer sensitive private information. Supporting social recommendation with least privacy leakage to untrusted recommender and other users (i.e., friends) is an important yet challenging problem. In this paper, we aim to achieve privacy-preserving social recommendation under personalized privacy settings. We propose PrivSR, a novel privacy-preserving social recommendation framework, in which user can model user feedbacks and social relationships privately. Meanwhile, by allocating different noise magnitudes to personalized sensitive and non-sensitive feedbacks, we can protect users’ privacy against untrusted recommender and friends. Theoretical analysis and experimental evaluation on real-world datasets demonstrate that our framework can protect users’ privacy while being able to retain effectiveness of the underlying recommender system.

AB - Privacy leakage is an important issue for social relationships-based recommender systems (i.e., social recommendation). Existing privacy preserving social recommendation approaches usually allow the recommender to fully control users’ information. This may be problematic since the recommender itself may be untrusted, leading to serious privacy leakage. Besides, building social relationships requires sharing interests as well as other private information, which may lead to more privacy leakage. Although sometimes users are allowed to hide their sensitive private data using personalized privacy settings, the data being shared can still be abused by the adversaries to infer sensitive private information. Supporting social recommendation with least privacy leakage to untrusted recommender and other users (i.e., friends) is an important yet challenging problem. In this paper, we aim to achieve privacy-preserving social recommendation under personalized privacy settings. We propose PrivSR, a novel privacy-preserving social recommendation framework, in which user can model user feedbacks and social relationships privately. Meanwhile, by allocating different noise magnitudes to personalized sensitive and non-sensitive feedbacks, we can protect users’ privacy against untrusted recommender and friends. Theoretical analysis and experimental evaluation on real-world datasets demonstrate that our framework can protect users’ privacy while being able to retain effectiveness of the underlying recommender system.

KW - Differential privacy

KW - Personalized privacy settings

KW - Ranking

KW - Social recommendation

UR - http://www.scopus.com/inward/record.url?scp=85049886244&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85049886244&partnerID=8YFLogxK

U2 - 10.1007/s11280-018-0620-z

DO - 10.1007/s11280-018-0620-z

M3 - Article

AN - SCOPUS:85049886244

SP - 1

EP - 29

JO - World Wide Web

JF - World Wide Web

SN - 1386-145X

ER -