Privacy preserving two-party k-means clustering over vertically partitioned dataset
Zhenmin Lin, Jerzy W. Jaromczyk · 2011
We propose a secure approximate comparison protocol and develop a practical privacy-preserving two-party k-means clustering algorithm over vertically partitioned dataset. Experiments with to real datasets show that the accuracy of clustering achieved with our privacy preserving protocol is similar to the standard (non-secure) kmeans function in MATLAB.