A Proposed Stacked Machine Learning Model to Predict the Survival of a Patient with Heart Failure

Abstract

Now a days heart failure is one of the most common chronic diseases that cause death. As it possesses high risk of death, it is important to predict patient’s survival and optimize treatment strategies. Machine learning techniques have come to light as useful tools for evaluating enormous quantities of patient data and deriving important patterns and insights in recent years. The purpose of the study is to investigate the feasibility of using the machine learning methods for predicting heart failure patient’s chances of survival. We have worked on a dataset with 2029 heart failure patients and the dataset comprises 13 features. To conduct this research, we suggested a model (Stacked machine learning model using scikit-learn using Decision Tree, Naive Bias, Random Forest, Linear Regression, SVM, XGBoost, ANN) using which we got better results than previously existed researches. We believe the suggested model will help advance our understanding of heart attack prediction.