Cardiovascular Disease Prediction Using Machine Learning

Cardiovascular Diseases (CVDs) remain among the leading causes of death worldwide, highlighting the urgent need for tools that can support early diagnosis and intervention.This research explores the use of supervised machine learning techniques to predict cardiovascular risk using commonly available clinical and demographic data.The goal of this research is to develop accurate, interpretable models that help the clinical department identify individuals at risk and improve decision-making in preventive cardiology.The research focuses on three widely used algorithms: Random Forest, XGBoost, and Support Vector Machine (SVM).The dataset used contains 14 key attributes collected from patient health records CDC (CDC: Centers for Disease Control and Prevention), including main factors age, gender, blood pressure, cholesterol levels, chest pain type, and maximum heart rate.A thorough data preprocessing pipeline is applied prior to model training.This includes handling missing values with imputation, removing duplicates, detecting.These steps help ensure clean, consistent input for the models.Results demonstrate that Random Forest outperformed the other models, achieving an accuracy of 84%, ROC-AUC score of 92%, precision 84%, F1-Score 84% and recall/sensitivity of 84%.XGBoost exhibited comparable performance with slightly higher precision of 86%, accuracy of 82%, recall/sensitivity of 78%, F1-score of 82% and ROC-AUC of 91%, while Support Vector Machine (SVM) showed promise with accuracy of 70%, precision of 67%, F1-Score of 76%, sensitivity/recall 88% and ROC-AUC 84% but is sensitive to feature scaling and lacked robustness in handling non-linear patterns in the data.The models are further tested on hypothetical patient profiles to assess their ability to detect less obvious patterns in the data.These results reinforce the potential of ensemble machine learning methods to support early risk detection and assist clinicians in identifying high-risk patients.With proper integration, such models can serve as valuable tools in real-world healthcare settings.