[Study on application of SVM in prediction of coronary heart disease]

Area of Science:

• Biomedical Engineering
• Computational Biology
• Cardiovascular Disease Research

Background:

• Coronary heart disease (CHD) poses a significant global health challenge.
• Accurate and early diagnosis is crucial for effective prevention and treatment.
• Traditional diagnostic methods may benefit from advanced computational approaches.

Purpose of the Study:

• To develop and validate a Support Vector Machine (SVM) model for identifying CHD in a South China population.
• To compare the performance of an optimized SVM model against other classification algorithms.
• To assess the utility of SVM in assisting CHD diagnosis based on clinical data.

Main Methods:

• Utilized clinical data including blood pressure, plasma lipids, glucose (Glu), and uric acid (UA).
• Implemented Support Vector Machine (SVM) with radial basis function (RBF), linear, and polynomial kernels.
• Optimized SVM parameters (penalty factor C, kernel sigma) using Particle Swarm Optimization (PSO).
• Compared optimized SVM performance against Artificial Neural Network (BP), Linear Discriminant Analysis, Logistic Regression, and non-optimized SVM.

Main Results:

• The optimized RBF-SVM model achieved superior classification performance.
• Achieved high accuracy (94.51%), sensitivity (92.31%), and specificity (96.67%).
• Demonstrated significantly better results compared to alternative classification algorithms.

Conclusions:

• Optimized SVM, particularly the RBF-SVM model, is a valid and effective method for assisting CHD diagnosis.
• This approach offers potential for improved early detection and management of coronary heart disease.
• The findings support the integration of machine learning in cardiovascular diagnostics.