Privacy preserving two-party k-means clustering over vertically partitioned dataset

Lin, Zhenmin; Jaromczyk, Jerzy W.

doi:10.1109/isi.2011.5983998

Cited by 5 publications

(1 citation statement)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are also other studies on vertically partitioned data distribution. 34,35 In horizontally partitioned data distributions, each entity is owned by the user side so that the users can calculate the distance to the cluster center while maintaining privacy. However, privacy leaks may occur during the calculation of intermediate results by the cluster centers.…”

Section: Related Workmentioning

confidence: 99%

Secure grid‐based density peaks clustering on hybrid cloud for industrial IoT

Sun

Liu

et al. 2020

Int J Network Mgmt

View full text Add to dashboard Cite

Cloud computing gives clients the convenience of outsourcing data calculations. However, it also brings the risk of privacy leakage, and datasets that process industrial IoT information have a high computational cost for clients. To address these problems, this paper proposes a secure grid-based density peaks clustering algorithm for a hybrid cloud environment. First, the client utilizes the homomorphic encryption algorithm to construct encrypted objects with client dataset. Second, the client uploads the encrypted data to the cloud servers to implement our security protocol. Finally, the cloud servers return the clustering results with the disturbance to the client. The experimental results on the UCI datasets and the smart power grid dataset reveal that the secure algorithm presented in this paper can improve upon the precision and efficiency of other clustering algorithms while also preserving user privacy. Moreover, it only performs encryption and removes the disturbance operation on the client, so that the client has lower computational complexity. Therefore, the secure clustering scheme proposed in this paper is applicable to industrial IoT big data and has high security and scalability.

show abstract

Section: Related Workmentioning

confidence: 99%