Harnessing Machine Learning for Accurate Heart-Disease Prediction

Cardiovascular disease (CVD) is a major health concern affecting many populations. Accurate prediction using machine learning can assist in early detection and management. The project titled “Cardiac Clarity: Harnessing Machine Learning for Accurate Heart-Disease Prediction” develops an improved model for CVD prediction using ensemble learning. The existing model uses a stacking classifier combining Random Forest, Logistic Regression, XGBoost, Support Vector Classifier, and Decision Tree Classifier based on clinical data.

The proposed model enhances this by employing CatBoost, LightGBM, ExtraTrees, and MLPClassifier as base learners. Instead of directly combining predictions, the model calibrates the probability outputs of these base learners and inputs them to a meta-model, either RidgeClassifier or CatBoostClassifier. This layered approach improves prediction accuracy and robustness. The model is trained and tested on a public heart disease dataset, using standard metrics for evaluation. The project also includes a web application where users can input health parameters and obtain predictions. The system is designed to be modular, scalable, and user-friendly, with a clean frontend and backend implemented using Django and Python.

Keywords: Heart Disease, Machine Learning, Stacking Classifier, CatBoost, LightGBM, RidgeClassifier, MLPClassifier, ExtraTrees, Prediction, Ensemble Learning.