Privacy-Preserving Clustering with High Accuracy and Low Time Complexity

This paper proposes an effficient solution with high accuracy to the problem of privacy-preserving clustering. This problem has been studied mainly using two approaches: data perturbation and secure multiparty computation. In our research, we focus on the data perturbation approach, and propose an algorithm of linear time complexity based on 1-d clustering to perturb the data. Performance study on real datasets from the UCI machine learning repository shows that our approach reaches better accuracy and hence lowers the distortion of clustering result than previous approaches