Improving an Intelligent Detection System for Coronary Heart Disease Using a Two-Tier Classifier Ensemble

Area of Science:

• Cardiology
• Machine Learning
• Data Science

Background:

• Coronary heart disease (CHD) is a leading global cause of mortality.
• Heart attacks often occur without warning symptoms, necessitating advanced detection methods.
• Traditional diagnostic approaches may lack the precision for early detection.

Purpose of the Study:

• To develop and evaluate a novel machine learning-based method for detecting coronary heart disease (CHD).
• To enhance the accuracy and generalisability of CHD detection models.
• To improve early identification of heart conditions for timely intervention.

Main Methods:

• A two-tier ensemble machine learning architecture was designed, integrating Random Forest, Gradient Boosting Machine, and Extreme Gradient Boosting classifiers.
• Particle Swarm Optimization was employed for feature selection to identify the most significant predictive variables.
• The model was rigorously evaluated on multiple established heart disease datasets (Z-Alizadeh Sani, Statlog, Cleveland, Hungarian) using 10-fold cross-validation.

Main Results:

• The proposed stacked ensemble model demonstrated superior performance compared to individual base classifiers and traditional ensemble methods.
• The model achieved high accuracy, F1-score, and Area Under the Curve (AUC) across diverse datasets.
• Statistical tests confirmed the significance of the performance improvements and the generalisability of the model.

Conclusions:

• The developed two-tier ensemble model offers a robust and accurate solution for coronary heart disease detection.
• This machine learning approach represents a significant advancement over existing methods, enhancing early diagnosis capabilities.
• The findings suggest a considerable contribution to the field of automated cardiovascular disease detection and risk stratification.