Published Paper Details:

Abstract-- Cardiovascular diseases (CVDs) continue to be the primary cause of death globally, highlighting the critical need for efficient prevention and early detection. Traditional diagnostic approaches, though clinically valuable, often involve high costs, delayed diagnosis, and limited personalization. Leveraging extensive hospital records, this work proposes a robust machine learning-driven framework for automated feature selection, risk prediction, and clinical decision support in CVD management. A comprehensive prepossessing pipeline, including normalization, missing value handling, and feature selection, ensured reliable model inputs. Multiple machine learning algorithms classified patients into cardiovascular risk categories based on clinical and lifestyle parameters such as cholesterol, blood pressure, and fasting blood sugar, and ECG. To enhance predictive accuracy, a Super Learner ensemble algorithm was employed to combine multiple base models, achieving an F1-score of 92.3% with a 90.94% average accuracy, outperforming current methods. Metrics like accuracy and precision were employed in performance validation. recall, F1-score, and ROC-AUC, demonstrating robustness across datasets. The integrated recommendation engine provides personalized lifestyle and medical guidance, including dietary changes, exercise plans, and consultation needs, through a user-friendly graphical interface for instant predictions and recommendations. The novelty of this work lies in combining the Super Learner ensemble approach with a personalized recommendation system, thereby improving predictive accuracy and supporting preventive cardiovascular care to alleviate the overall healthcare burden.