Secure and efficient publicly verifiable ridge regression outsourcing scheme
Ou Ruan, Shanshan Qin · 2022
Ridge regression is an important statistical method that is widely used in real life, such as health prediction. As clients with limited computing resources may not be able to handle large training datasets, it is necessary to outsource this operation to a powerful cloud server. Data privacy is raised due to outsourcing. Some previous works on secure outsourcing ridge regression were provided for privacy preserving. However, there are some issues such as having heavy workloads and inefficient. In this paper, we propose an efficient privacy-preserving protocol for outsourcing ridge regression. In our design, we utilize blinding technology to protect the privacy of clients by transforming the original problem into an encryption problem with orthogonal and diagonal matrices as the secret keys. The following advantages can be seen from the experimental result and theoretical analysis: (1) Our protocol is more efficient than related works. When the calculation dimension is greater than 6000×1500, clients computations greatly reduce to 60 percent of other schemes and the total computing is close to the original calculation; (2) Edge server provides the public verifiability, which allows any verifier to verify whether the result returned by the cloud server is correct; (3) We also give a formal proof of security based on the standard simulation model.