Privacy Preserving Boosting in the Cloud with Secure Half-Space Queries

摘要

This paper presents a preliminary study on the PerturBoost approach that aims to provide efficient and secure classifier learning in the cloud with both data and model privacy preserved.