Federated learning-assisted intelligent yellow fever outspread prediction framework

Deep learning has demonstrated strong potential in the detection of yellow fever. However, its effectiveness is often hindered by privacy concerns and limitations in data sharing across medical institutions. To address this challenge, this study proposes a novel Federated Learning-inspired framework for privacy-preserving, collaborative symptom analysis aimed at containing the spread of yellow fever. The proposed system integrates multimodal inputs specifically audio and visual data to detect key symptoms and utilizes a Dynamic Integration Approach for evaluating patient health severity. Comparative experiments were conducted against several state-of-the-art methods. Results show that the proposed model achieves superior performance, with an accuracy of 98.57%, precision of 99.85%, F-score of 96.53%, and recall of 94.72%. Additionally, the system exhibits enhanced delay latency (11.80 s) and operational stability (71%), indicating its robustness and real-time applicability in medical environments.