Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Videos

Machine Unlearning Based on Globally Refined Convergent Clustering for Health Survey Data-driven Prediction Model.

Ziyan Wang, Yuqin Zhou, Xufeng Lang

IEEE Journal of Biomedical and Health Informatics

|July 13, 2026

Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

BCP-YOLO: a lightweight and interpretable YOLOv8-based framework for real-time coloreactal polyp detection.

BMC medical imaging·2026

Same author

Electronic and molecular dynamics mechanisms of wogonin adsorption on graphene nanocarriers.

Scientific reports·2026

Same author

Construction and application of a traditional Chinese medicine syndrome differentiation and treatment model grounded in knowledge distillation and reinforcement learning.

Health information science and systems·2026

Same author

In-silico prediction of multi‑target mechanisms of Pinellia ternata phytochemicals in lung cancer: Evidence from a graph‑attention‑guided virtual screening and multi‑scale simulations.

PloS one·2026

Same author

Computational profiling of vernomenin from Vernonia amygdalina via network pharmacology and molecular modeling: a preliminary assessment for fever targeting.

Journal of molecular graphics & modelling·2026

Same author

A Dual Domain Collaborative Network for Polyp Segmentation.

IEEE journal of biomedical and health informatics·2025

We developed a subgroup-aware unlearning method for health surveys, ensuring data privacy while maintaining model accuracy. This approach effectively removes participant data impacts, addressing GDPR requirements for machine learning in healthcare.

Area of Science:

Health Informatics
Machine Learning
Data Privacy

Background:

Machine learning models from national health surveys aid risk stratification.
The GDPR's
right to be forgotten
requires complete data erasure, posing challenges for existing privacy methods.
Current unlearning techniques are often computationally expensive, offer approximate privacy, or struggle with heterogeneous health data.

Purpose of the Study:

To propose a subgroup-aware exact unlearning framework for heterogeneous health survey data.
To address the limitations of existing unlearning methods in preserving privacy and utility.
To enable effective data erasure compliant with regulations like GDPR.

Main Methods:

Related Experiment Videos

A multi-stage refinement process identifies epidemiologically coherent subgroups.

Proportion-preserving shards are created for localized unlearning, maintaining cohort representation.

Affected shards undergo localized retraining from clean snapshots, with global consistency restored via soft-voting ensemble.

Main Results:

The proposed framework achieves exact unlearning with bounded retraining scope.
Demonstrated strong forgetting performance while preserving predictive utility on depression-risk prediction tasks.
Evaluations on NHANES and CHARLS datasets confirmed improved Zero-Retrain-Forgetting performance.

Conclusions:

The subgroup-aware exact unlearning framework offers a practical solution for health survey data.
It effectively balances data privacy requirements with the need for accurate predictive models.
This approach supports robust and compliant machine learning applications in public health research.