EUVIMEDEUVIMED entdecken
Crossref

A Differentially Private Federated Learning Framework of Secure Diabetes Risk Prediction

Madeeha Laiq, Gaurav Pandey, Surya Prakash Dwivedi

International Journal of Drug Delivery Technology · 29.05.2026 · 16 (35s)

DOI: 10.25258/ijddt.16.35s.59 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Drug Delivery Technology, online erschienen am 29.05.2026. Verfasst von Madeeha Laiq und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The high rate of expansion of health-related information provided by hospitals, wearables, and electronic health records (EHRs) has provided an opportunity to use machine learning models to predict illness and provide personalized care. However, when handling important medical data, centralized data collecting presents a serious privacy and security risk.In an effort to address the issue, this paper suggests a Differentially Private Federated Learning (DP-FL) model of making diabetes risk predictions securely. In the proposed framework, Federated Averaging (FedAvg) will be used to jointly learn a deep neural network with various hospital nodes (H1, H2, and H3) without the actual transfer of raw patient data to a main server. Every client does regional training with clinical and lifestyle characteristics and communicates only encrypted updates to the model. In order to guarantee high privacy levels, prior to aggregation, Gaussian noise is introduced to train gradients with the help of differential privacy mechanisms. As shown by experimental assessment of 2,200 patient records, the suggested DP-FL model can be ranked high in predictive performance with an accuracy approaching 97 percent, and simultaneously, it maintains patient privacy. The model offers privacy-friendly and scalable collaborative healthcare analytics in distributed settings.

Zitieren

APA

Laiq, M., Pandey, G., & Dwivedi, S. P. (2026). A Differentially Private Federated Learning Framework of Secure Diabetes Risk Prediction. International Journal of Drug Delivery Technology, 16(35s). https://doi.org/10.25258/ijddt.16.35s.59

Vancouver

Laiq M, Pandey G, Dwivedi SP. A Differentially Private Federated Learning Framework of Secure Diabetes Risk Prediction. International Journal of Drug Delivery Technology. 2026;16(35s). doi:10.25258/ijddt.16.35s.59

BibTeX
@article{Laiq2026,
  author = {Laiq, Madeeha and Pandey, Gaurav and Dwivedi, Surya Prakash},
  title = {A Differentially Private Federated Learning Framework of Secure Diabetes Risk Prediction},
  journal = {International Journal of Drug Delivery Technology},
  year = {2026},
  volume = {16},
  number = {35s},
  doi = {10.25258/ijddt.16.35s.59}
}
Kostenlos starten →