Cardiovascular Risk Prediction Through Stacking Classifier

Enormous amounts of patient data, including information on their liver, diabetes, heart, and maternal health, are being generated by medical sectors.Among them, cardiovascular diseases (CVD) are thought to be the deadliest, posing the greatest health risk to mortality by causing silent and not transmissible death.The preventive approach requires an examination of health records which is quite difficult; therefore, adopting machine learning (ML) technology that identifies early CVD is necessary.In this work, we used three different boosting techniques, namely, AdaBoost, gradient boosting, XgBoost, and stacked boosting algorithms, as well as a combination of these algorithms with and without hyperparameter tuning to predict CVD patients.Our analysis showed that, compared to a single boosting model, the recall score of the stacked boosting algorithm is much higher, demonstrating a 92.85% recall score after using hyperparameter tuning.Our predicted model will help hospital associates to take preventive measures in these cases as well as properly allocate the resources for hospital managers and, lastly, reduce the financial burdens of the patients.