Privacy-Preserving Discovery of Multivariate Linear Relationship
Ningning Wu, Jing Zhang, Li Ning
The fast development of network and database techniques makes the data collecting and storing much easy and convenient. With more data being collected and available, there come the increasing requirements and huge opportunities for cooperative computation, where data are distributed across sites, and each site holds a portion of the data and wishes to collaborate to detect globally valid multivariate linear relationship.
This paper considers the privacy-preserving cooperative linear system of equations (PPC-LSE) problem in a large, heterogeneous, distributed database scenario, in which two parties would like to conduct cooperative computation from their private database while keeping their own data secret. The paper proposes a privacy-preserving algorithm to discover multivariate linear relationship based on factor analysis. Compared with other PPC_LSE algorithms, the proposed algorithm not only significantly reduces the communication cost, but also avoids the random