TY - GEN
T1 - dK-Personalization
T2 - 26th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2022
AU - Iftikhar, Masooma
AU - Wang, Qing
AU - Li, Yang
N1 - Publisher Copyright:
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
PY - 2022
Y1 - 2022
N2 - Preserving privacy of an individual in network structured data while enhancing utility of published data is one of the most challenging problems in data privacy. Moreover, different individuals might have different privacy levels based on their own preferences, thereby personalization needs to be considered to achieve personal data protection. In this paper, we aim to develop a privacy-preserving mechanism to publish network statistics, particularly degree distribution, and joint degree distribution, which guarantees personalized (edge or node) differential privacy while enhancing network data utility. To this extend we propose four approaches to handle personal privacy requirements of individuals in a differentially private computation. We have empirically verified the utility enhancement and privacy guarantee of our proposed approaches on four real-world network datasets. To the best of our knowledge, this is the first study to publish network data distributions under personalized differential privacy, while enhancing network data utility.
AB - Preserving privacy of an individual in network structured data while enhancing utility of published data is one of the most challenging problems in data privacy. Moreover, different individuals might have different privacy levels based on their own preferences, thereby personalization needs to be considered to achieve personal data protection. In this paper, we aim to develop a privacy-preserving mechanism to publish network statistics, particularly degree distribution, and joint degree distribution, which guarantees personalized (edge or node) differential privacy while enhancing network data utility. To this extend we propose four approaches to handle personal privacy requirements of individuals in a differentially private computation. We have empirically verified the utility enhancement and privacy guarantee of our proposed approaches on four real-world network datasets. To the best of our knowledge, this is the first study to publish network data distributions under personalized differential privacy, while enhancing network data utility.
KW - Graph data utility
KW - Network data distributions
KW - Personalized differential privacy
KW - Privacy-preserving graph data publishing
UR - http://www.scopus.com/inward/record.url?scp=85130368882&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-05933-9_16
DO - 10.1007/978-3-031-05933-9_16
M3 - Conference contribution
SN - 9783031059322
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 194
EP - 207
BT - Advances in Knowledge Discovery and Data Mining - 26th Pacific-Asia Conference, PAKDD 2022, Proceedings
A2 - Gama, João
A2 - Li, Tianrui
A2 - Yu, Yang
A2 - Chen, Enhong
A2 - Zheng, Yu
A2 - Teng, Fei
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 16 May 2022 through 19 May 2022
ER -