Survival Prediction of Heart Failure Patients using Stacked Ensemble Machine Learning Algorithm
Published in 2021 IEEE International Women in Engineering (WIE) Conference on Electrical and Computer Engineering (WIECON-ECE), 2021
Heart failure (HF) is the leading cause of global death from chronic diseases. Data mining using machine learning (ML) converts massive volumes of raw data created by healthcare institutions into meaningful information that can aid in making predictions and crucial decisions. After an HF, collecting and analyzing follow-up data from patients is critical to monitor their health recovery. The aim of this study is to use ML and predict the survival possibility of patients after HF based on the follow-up data. Three supervised classifiers i.e., Random Forest (RF), XGBoost (XGB), and Decision Tree (DT) have been used in our study. Moreover, we proposed to design a supervised stacked ensemble learning model that can achieve a prediction accuracy, precision, recall, and F1 score of 99.98%, 100%, 99.98%, and 99.98%, respectively. The ensemble structure has three base learners (DT, RF & XGB) and one meta learner (RF). The integration of multiple algorithms boosted the prediction performance, which is a significant improvement from the other contemporary studies. Our results demonstrate that Ensemble ML can be a powerful intervention tool to predict accurately, beforehand, the recovery status of a chronic HF patient and prevent potential fatalities.
Recommended citation: Zaman, S.M., Qureshi, W.M., Raihan, M.M.S., Shams, A.B. and Sultana, S., 2021, December. Survival Prediction of Heart Failure Patients using Stacked Ensemble Machine Learning Algorithm. In 2021 IEEE International Women in Engineering (WIE) Conference on Electrical and Computer Engineering (WIECON-ECE) (pp. 117-120). IEEE.