Improving the Utility of Differentially Private Clustering through Dynamical Processing
Preprint
Abstract
This study aims to alleviate the trade-off between utility and privacy in the task of differentially private clustering. Existing works focus on simple clustering methods, which show poor clustering performance for non-convex clusters. By utilizing Morse theory, we hierarchically connect the Gaussian sub-clusters to fit complex cluster distributions. Because differentially private sub-clusters are obtained through the existing methods, the proposed method causes little or no additional privacy loss. We provide a theoretical background that implies that the proposed method is inductive and can achieve any desired number of clusters. Experiments on various datasets show that our framework achieves better clustering performance at the same privacy level, compared to the existing methods.Keywords
All Related Versions
This publication has 23 references indexed in Scilit:
- Distributed K-Means clustering guaranteeing local differential privacyComputers & Security, 2020
- Outlier-eliminated k-means clustering algorithm based on differential privacy preservationApplied Intelligence, 2016
- A quadratic string adapted barrier exploring method for locating transition statesComputer Physics Communications, 2007
- An online classification EM algorithm based on the mixture modelStatistics and Computing, 2007
- Marketing segmentation using support vector clusteringExpert Systems with Applications, 2007
- Fuzzy c-means clustering with spatial information for image segmentationComputerized Medical Imaging and Graphics, 2006
- A database for handwritten text recognition researchIEEE Transactions on Pattern Analysis and Machine Intelligence, 1994
- A classification EM algorithm for clustering and two stochastic versionsComputational Statistics & Data Analysis, 1992
- Nonlinear oscillations, dynamical systems, and bifurcations of vector fields: Applied Mathematical Science, Vol. 42, J. Guckenheimer and P. Holmes, Springer-Verlag, New York, Berlin, Heidelberg, Tokyo (1983). XVI + 453 pp., 206 figs, DM 104International Journal of Engineering Science, 1988
- Nonlinear Oscillations, Dynamical Systems, and Bifurcations of Vector Fields1983