Predicting Survival of Heart Failure Patients Using Classification Algorithms

Abstract

Heart failure is a situation that occurs when the heart is unable to pump enough blood to meet the needs of other organs in the body. It is responsible for the annual death of approximately 17 million people worldwide. Series of studies have been done to predict heart failure survival with promising results. Hence, the purpose of this study is to increase the accuracy of previous works on predicting heart failure survival by selecting significant predictive features in order of their ranking and dealing with class imbalance in the classification dataset. In this study, we propose an integrated method using machine learning. The proposed method shows promising results as it performs better than previous works and this study confirms that dealing with imbalanced dataset properly increases accuracy of a model. The model was evaluated based on metrics such as F-measure, Precision-Recall curve and Receiver Operating Characteristic Area Under Curve. This discovery has the potential to impact on clinical practice, when health workers aim at predicting if a patient will survive heart failure. Attention may be focused on mainly serum creatinine, ejection fraction, smoking status and age.