Effective Feature Engineering Technique for Heart Disease Prediction With Machine Learning

Abstract

Heart Disease, a chronic ailment impacting millions globally, underscores the significance of early detection. An innovative feature engineering technique, utilizing Principal Component analysis, is introduced to identify and enhance the most crucial features Utilizing machine learning, the project aims to predict Heart Diseases health status promptly and initiate essential actions. Include project, an ensemble method is implemented, specifically a Stacking Classifier, which combines the predictions of Random Forest (RF), Multilayer Perceptron (MLP), and LightGBM models. This approach synergistically leverages the strengths of individual models, resulting in a highly robust and accurate final prediction, achieving an impressive 100% accuracy. The selected features based on Principal Component Heart Failure (PCHF) were utilized for model building, and the Stacking Classifier was trained to be deployed in the front end. The integration of Flask framework with user authentication ensures an effective and secure platform for user testing, enhancing the accessibility and usability of our machine learning-based heart disease prediction system.