Efficient Model for Coronary Artery Disease Diagnosis: A Comparative Study of Several Machine Learning Algorithms

Area of Science:

• Cardiology
• Medical Informatics
• Artificial Intelligence

Background:

• Coronary artery disease (CAD) is a leading cause of mortality in industrialized nations.
• Early detection and intervention are crucial for managing CAD complications and reducing mortality.
• Machine learning (ML) offers advanced computational solutions for disease diagnosis.

Purpose of the Study:

• To compare the performance of various ML algorithms for early CAD diagnosis.
• To develop an effective diagnostic model for CAD using clinical examination features.
• To identify the most efficient ML algorithms for CAD prediction.

Main Methods:

• Applied descriptive study involving 303 patient records and 26 selected clinical features.
• Evaluation of seven classification algorithms: Multilayer Perceptron (MLP), Support Vector Machine (SVM), Logistic Regression (LR), J48, Random Forest (RF), K-Nearest Neighborhood (KNN), and Naive Bayes (NB).
• Comparison of algorithm performance metrics to determine the most effective models for CAD diagnosis.

Main Results:

• Support Vector Machine (SVM) achieved an AUC of 0.88 and F-measure of 0.88.
• Random Forest (RF) demonstrated an AUC of 0.87 and ROC of 0.91.
• K-Nearest Neighborhood (KNN) showed the lowest efficiency with an AUC of 0.81 and F-measure of 0.81.

Conclusions:

• ML algorithms play a significant role in the diagnosis of coronary artery disease.
• SVM and RF algorithms are highly effective in diagnosing CAD based on patient examination data.
• The proposed ML models can serve as a foundation for developing clinical decision support systems for CAD.