Federated Learning for Privacy-Preserving Hospital Readmission Prediction in Regional Hospitals

juc_2_20261011162403 20261011162403 JURNAL UJI COBA purwan78@gmail.com Karoteh Utama Publisher JURNAL UJI COBA JUC 1234-273X 09212026 1 1 Federated Learning for Privacy-Preserving Hospital Readmission Prediction in Regional Hospitals Siti Aminah Institut Teknologi Samudra Marco Bellini Università di Città Nuova Regional hospitals hold valuable patient data but cannot share it because of privacy rules. We train a 30-day readmission model across nine hospitals using federated averaging with differential privacy, so raw records never leave each site. The federated model achieves an AUC of 0.81, compared with 0.73 for the average single-hospital model, at a privacy budget of epsilon 3. We release the training pipeline and a governance checklist that hospital IT teams can adopt. 09212026 1529 https://creativecommons.org/licenses/by/4.0/ https://uji.purwantoro.my.id/article/2