Ensemble Learning Based on Hybrid Deep Learning Model for Heart Disease Early Prediction

Area of Science:

• Cardiology
• Artificial Intelligence
• Machine Learning

Background:

• Heart disease remains a leading cause of mortality globally.
• Lifestyle factors like poor diet, smoking, and inactivity contribute to its prevalence.
• The 'silent killer' nature of heart disease necessitates improved early detection methods.

Purpose of the Study:

• To propose a deep stacking ensemble model for enhanced heart disease prediction.
• To leverage hybrid deep learning architectures and Support Vector Machine (SVM) for improved accuracy.
• To optimize the model through feature selection and comparison with existing methods.

Main Methods:

• Developed a deep stacking ensemble integrating Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) and Convolutional Neural Network-Gated Recurrent Unit (CNN-GRU) models.
• Employed Recursive Feature Elimination (RFE) for feature selection optimization.
• Trained and validated the ensemble model on two distinct heart disease datasets.

Main Results:

• The proposed deep stacking ensemble model achieved the highest performance metrics.
• The model demonstrated superior predictive accuracy compared to traditional machine learning and individual hybrid models.
• Optimization techniques further enhanced the performance of all evaluated models.

Conclusions:

• The developed deep stacking ensemble model offers a promising approach for accurate early heart disease detection.
• Integration of advanced deep learning architectures with SVM provides a robust predictive tool.
• Further research can explore this ensemble approach for other complex health predictions.