Predicting Coronary Heart Disease Using Data Mining and Machine Learning Solutions

This research focuses on predicting cardiovascular disease using machine learning classification strategies. The study presents a unique approach by integrating multiple machine learning techniques, leveraging the strengths of Random Forest and Gradient Boosting. The authors developed a novel ensemble learning model, combining Linear Regression, Random Forest, and Gradient Boosting algorithms, optimized using Bayesian hyperparameter tuning. The model demonstrated superior performance in predicting CVD outcomes, with classification accuracy of 95.5%, 94.26%, and 98.3% for Linear Regression, Decision Tree, and Gradient Boosted methods, respectively. The true positive rate for the GB algorithm's predictions of patients was 98.3%. The study hypothesizes that the GB method predicts the Framingham dataset better than other algorithms using 4240 samples.