Coronary heart disease prediction based on hybrid deep learning

Area of Science:

• Cardiology
• Artificial Intelligence
• Medical Informatics

Background:

• Machine learning aids medical experts in diagnosing coronary heart disease (CAD).
• Traditional methods face limitations in prediction accuracy and overfitting.
• Deep learning offers potential for enhanced diagnostic capabilities.

Purpose of the Study:

• To propose a novel deep learning hybrid model for improved CAD prediction.
• To enhance the accuracy and reliability of CAD diagnosis using advanced AI.
• To address the overfitting issue prevalent in single-model machine learning approaches.

Main Methods:

• Development of a hybrid deep learning model combining two deep neural networks and a recurrent neural network.
• Utilizing a k-nearest neighbor model for secondary training to refine predictions.
• Training and validation on a dataset comprising 7291 patient records.

Main Results:

• The hybrid model achieved a prediction accuracy of 82.8% on the test set.
• Key performance metrics include precision (87.08%), recall (88.57%), and F1-score (87.82%).
• The model demonstrated an Area Under the Curve (AUC) value of 0.8, indicating strong predictive power.

Conclusions:

• The proposed hybrid deep learning model significantly improves CAD prediction accuracy compared to single models.
• The model effectively mitigates overfitting, offering a more robust diagnostic tool.
• This approach provides valuable auxiliary support for clinical CAD diagnosis by exploring complex feature inter-relationships.