Research on Federated Learning Model for Assisted Diagnosis in Cancer Rehabilitation-Based on AMO-XGBoost Model
Year: 2025
Type: article
Abstract: With the application of AI in healthcare, privacy and security issues are a growing concern. Protecting the privacy of patient data and controlling the loss of model accuracy becomes critical when training AI models. Cancer is an increasing threat to human health, and postoperative patients are at high risk of recurrence; predicting the time and location of recurrence is critical. This study uses federated learning combined with localised differential privacy and the AMO-XGBoost model to investigate a privacy-preserving model suitable for cancer recovery patient data. The model is tested on a dataset of 5 types of cancer in rehabilitation patients, aiming to construct a privacy-preserving and accurate model for assisted diagnosis in cancer rehabilitation.
DOI: 10.54097/gqqkp488
Authors: Zebao Zong, Qinyu Zhang, Li Yuan, Zihan Xiao, Kenan Wang
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4408933207