Design of Cross-Domain Medical Data Privacy-Preserving Classification Algorithm Based on Federated Learning
Yang bo  Gong yichen

Mongolian National University ，Ulaanbaatar city Bayangol District 11th Khoroo National University of Mongolia，16060；

Abstract：Against the backdrop of rapid medical informatization, the sharing and utilization of cross-domain medical data have become key to improving the accuracy and efficiency of medical diagnosis, but the risk of data privacy leakage has also intensified. Traditional centralized machine learning algorithms require aggregating multi-domain medical data onto a unified platform, making it difficult to meet privacy protection requirements. Federated learning, as a distributed machine learning technique, enables collaborative model training without sharing raw data, providing a new solution for cross-domain medical data privacy protection. This paper designs a cross-domain medical data privacy-preserving classification algorithm based on federated learning. It first analyzes the characteristics of cross-domain medical data and privacy protection requirements, then designs an algorithmic framework combining federated learning architecture, including key steps such as local model training, encrypted transmission of model parameters, and global model aggregation and updating. Finally, it discusses the challenges and optimization directions of the algorithm in practical applications, aiming to provide technical support for the secure utilization of data and classification tasks in cross-domain medical scenarios, and to promote the compliant development of medical artificial intelligence.

Keywords: Privacy Protection; Classification Algorithm; Distributed Training

References

[1] Sun Changyong. Research and Implementation of Cross-Domain Task Scheduling Algorithm Based on Federated Deep Reinforcement Learning [D]. Beijing University of Posts and Telecommunications, 2023.

[2] Zhang Shenyilang. Research on Cross-Domain sEMG Gesture Recognition Based on Transfer Learning [D]. Hangzhou Dianzi University, 2023.

[3] Zhou Chuanxin. Research on Key Technologies for Data Security Sharing in Federated Learning [D]. Strategic Support Force Information Engineering University, 2022.

[4] Pu Liyuan. Research on Federated Learning Load-Aware Intelligent Management of RAN Slicing [D]. Beijing University of Posts and Telecommunications, 2023.

[5] Software Engineering. Research on Cross-Domain Recommendation Method Based on Transfer Federated Learning [D]. 2024.