Development of Health Parameter Model for Risk Prediction of CVD Using SVM.

Area of Science:

• Cardiology
• Medical Informatics
• Machine Learning

Background:

• Current cardiovascular disease (CVD) risk assessment relies on models like the Framingham study.
• These traditional methods exhibit limitations in sensitivity and specificity for accurate risk prediction.
• There is a need for improved CVD risk assessment tools.

Purpose of the Study:

• To compare the predictive performance of traditional Framingham risk factors with machine learning approaches.
• To evaluate the effectiveness of linear regression and Support Vector Machine (SVM) models trained on local data.
• To determine if machine learning can overcome the limitations of existing CVD risk assessment tools.

Main Methods:

• Linear regression analysis was used to train a model on a local database, comparing its performance to the Framingham equation.
• Support Vector Machine (SVM) was employed to assess the efficacy of a machine learning approach using Framingham health parameters for CVD risk prediction.
• Model performance was evaluated based on sensitivity and specificity.

Main Results:

• The linear regression model, trained on local data, showed improvement over the standard Framingham model.
• The SVM-based risk assessment model demonstrated high sensitivity and specificity in predicting CVD.
• Machine learning effectively addressed the low sensitivity and specificity issues inherent in the Framingham model.

Conclusions:

• Machine learning models, particularly SVM, offer a superior approach to cardiovascular disease risk assessment.
• Training models on local databases enhances prediction accuracy compared to generalized models.
• SVM utilizing Framingham health parameters provides a more sensitive and specific method for CVD risk prediction.