A Distributed Solution for Privacy Preserving Outlier Detection

In this paper, we study some parties - each has a private data set - want to conduct the outlier detection on their joint data set, but none of them want to disclose its private data to the other parties. We propose a linear transformation technique to design protocols of secure multivariate outlier detection in both horizontally and vertically distributed data models. While different from the most of previous techniques in a privacy preserving fashion for distance-based outliers detection, our focus is the technique in statistics for detecting outliers.