Privacy-Preserving SVM Classification on Vertically Partitioned Data without Secure Multi-party Computation

Abstract

With the development of information science and modern technology, it becomes more important about how to protect privacy information. In this paper, a novel privacy-preserving support vector machine (SVM) classifier is put forward for a vertically partitioned data. The proposed SVM classifier, which is public but does not reveal the privately-held data, has accuracy comparable to that of an ordinary SVM classifier based on the original data. We prove the feasibility of our algorithms by using matrix factorization theory and show the security without using the secure multi-party computation.