Efficient and privacy-preserving online diagnosis scheme based on federated learning in e-healthcare system

Publication Name

Information Sciences


Electronic healthcare (e-healthcare) system has brought great convenience for people to seek medical treatment. However, data security, user privacy and online diagnosis efficiency have also aroused widespread public concern. In this paper, we propose an efficient and privacy-preserving online diagnosis scheme for e-healthcare system based on federated learning mechanism (FLM). Specifically, we first transform the data owner's data sharing problem into machine learning problem through using FLM. By sharing calculated local model parameters instead of actual data, the privacy of training data sets can be well protected. Then, we use a homomorphic cryptosystem and the support vector machine (SVM) algorithm to classify patients' physiological data efficiently without leaking their privacy. Furthermore, we design a novel approach to recover decision function of SVM model, which can efficiently prevent model parameters from leaking. Security analysis shows that the proposed scheme can protect data privacy under the defined threat models. Numerical results derived from experiments show that the proposed scheme is highly efficient. Therefore, our scheme is of practical significance in the e-healthcare system.

Open Access Status

This publication is not available as open access



Article Number




Link to publisher version (DOI)