Explainable Machine Learning Models for Type-2 Diabetes Prediction

Conference: IMCOM 2026
Publisher: IEEE Xplore

This research addresses the challenge of trust and interpretability in healthcare AI. Many diabetes prediction models function as black boxes, limiting their usefulness in clinical and preventive settings. This study proposes explainable machine-learning models that balance high predictive accuracy with transparent decision logic.

Key Contributions:

• Developed explainable ML models for early Type-2 diabetes risk prediction
• Applied interpretable techniques to highlight key contributing risk factors
• Improved clinician and patient understanding of model decisions
• Supported proactive, preventive healthcare interventions
• This work demonstrates how explainability can be embedded into predictive models to support ethical and trustworthy AI deployment.