An Expert Hybrid Federated Learning and Trust Management for Security, Efficiency, and Power Optimization in Smart Health Systems

Abstract

Health care systems play an important role in smart city infrastructure and seem very beneficial to citizens. The large numbers of health devices are connected with each other and share the patient’s data, AI doctors analyze the data and give recommendations to patients. The challenges associated with the integration of the health system bring significant security and privacy issues to the forefront, especially with respect to sensitive patient information. Ensuring the security and privacy of the health system is necessary. To overcome these challenges, the author proposed a novel and practical model consisting of a hybrid federated SVM and trust management model. First, the system computes the trust, using the parameters of cooperativeness, honesty, and community trust. The proposed model achieves an overall accuracy of 95%, linear kernel accuracy of 95%, RBF kernel accuracy of 93%, and polynomial kernel accuracy of 95% against anomaly detection and provides security and privacy to the health system. The proposed approach is lightweight and reduces 52.5% computational. Our design also promotes savings on unnecessary energy consumption and computational overhead. As a result, our novel strategy opens the door to enhancing the security of smart health infrastructures, ensuring optimal performance and economical use of resources.