Predicting cardiovascular risk with hybrid ensemble learning and explainable AI.

Area of Science:

• Cardiology
• Artificial Intelligence
• Machine Learning

Background:

• Cardiovascular diseases (CVDs) remain a leading global cause of mortality.
• Accurate early risk prediction is crucial for effective prevention and treatment strategies.
• Existing models may lack the robustness and interpretability needed for clinical application.

Purpose of the Study:

• To develop an innovative hybrid ensemble learning framework for cardiovascular disease (CVD) risk prediction.
• To enhance model interpretability using explainable AI (XAI) techniques.
• To improve the accuracy and trustworthiness of AI in healthcare settings.

Main Methods:

• A stacked ensemble architecture combining Gradient Boosting, CatBoost, and Neural Networks was employed.
• Publicly accessible datasets were utilized for model training and validation.
• Explainable AI methods, including SHAP values, t-SNE, and PCA, were used for visualization and interpretation.

Main Results:

• The hybrid model achieved a high Area Under the Receiver Operating Characteristic Curve (AUC-ROC) score of 0.82.
• Classification metrics demonstrated strong performance: Precision 81%, Recall 83%, and F1-Score 82%.
• Visualizations revealed multidimensional relationships between risk factors (e.g., blood pressure, BMI, cholesterol-glucose ratio) and lifestyle parameters.

Conclusions:

• Ensemble learning offers a powerful approach for complex medical prediction tasks like CVD risk assessment.
• Model interpretability is essential for building trust in AI systems within clinical practice.
• The developed framework provides a promising tool for healthcare stakeholders to identify and manage CVD risk effectively.