Density Peak Clustering Algorithm Based on Differential Privacy Preserving

Clustering by fast search and find of density peaks (CFSFDP) is an efficient algorithm for density-based clustering. However, such algorithm inevitably results in privacy leakage. In this paper, we propose DP-CFSFDP to address this problem with differential privacy, which adds random noise in order to distort the data but preserve its statistical properties. Besides, due to the poor performance of CFSFDP on evenly distributed data, we further optimize the clustering process with reachable-centers and propose DP-rcCFSFDP. The experimental results show that, under the same privacy budget, DP-rcCFSFDP can improve the clustering effectiveness while preserving data privacy compared with DP-CFSFDP.