Cardiovascular disease prediction based on Stacking integrated strategy

Abstract

According to the report of the World Health Organization, cardiovascular disease causes at least 17.1 million deaths worldwide every year, ranking second among the top ten causes of death for many years, and is still a problem to be solved in China and even in the world. No less dangerous than cancer on the list. Early diagnosis is of great significance for patients with cardiovascular diseases, which can diagnose patients with cardiovascular diseases as early as possible to achieve the purpose of early treatment and reduce the cost and pain of patients. In order to achieve accurate early diagnosis of cardiovascular diseases, Logistic Regression, Decision Tree, Random Forest, Gradient Boosting (GBDT), Adaptive Boosting (AdaBoost), eXtreme Gradient Boosting (XGBoost), Deep Neural Network (DNN), and Stacking integrated models were used to predict cardiovascular diseases. The comparison results showed that the Stacking model is the optimal prediction model. The precision reached 86.80%, the recall reached 84.78%, and the f1 reached 85.76%. The proposed model can be used in cardiovascular disease prediction to reduce the incidence of cardiovascular disease.