COMPARISON OF SUPERVISED MACHINE LEARNING ALGORITHMS IN HEART FAILURE DISEASE

Abstract

The heart is a vital organ in the human body that functions to pump blood throughout the body and to the lungs. The heart is located in the chest cavity. The heart is the main force that drives human life. Therefore, if there is a disturbance in heart function, this can cause a decrease in quality of life to death, one of which is heart failure. Heart failure, if not diagnosed and treated quickly, will result in death. Based on findings showing the high death rate due to heart failure, a classification is needed to predict heart failure using machine learning methods. Machine learning can help predict this disease to improve early detection and more accurate medical decision-making. This study focuses on predicting the likelihood of a patient experiencing heart failure. The machine learning algorithm method used is supervised machine learning classification, including decision trees, random forests, naïve bayes, SVM, and K-NN. The results showed that the best method for predicting heart failure was Random Forest with an accuracy of 74.35%, followed by SVM with an accuracy of 69.23%. Meanwhile, Naïve Bayes had the lowest accuracy of 51.28%. Based on these findings, Random Forest is recommended as the best method for heart failure prediction due to its ability to handle data complexity and provide more stable results. Once the best algorithm is obtained, the prediction results and early detection of heart failure will be more accurate.