Predicting Mortality in Heart Failure Patients: A Comparative Study of Decision Tree C4.5 and Naïve Bayes Algorithms

The heart is one of many vital organs of humans that is used to pump blood to the entire part of the body. One of many diseases that attack the heart is heart failure. Heart failure arises from either structural or functional abnormalities that hinder the ventricle’s capability to effectively pump blood throughout the body. Based on the data released by the Health Research Department, the prevalence of heart failure in Indonesia is $1.5 \%$, which means 1 million Indonesians or more have heart problems. The mortality rate of heart failure patients can be reduced by data mining classification. The classification process uses decision tree C4.5 and the naïve Bayes algorithm. From 299 data of heart failure patients that were classified using a decision tree, the C4.5 algorithm got an $83.26 \%$ accuracy rate, and the naïve Bayes algorithm got a $79.26 \%$ accuracy rate. From the classification process that has been done, it can be seen that the decision tree C4.5 algorithm has better performance than the naïve Bayes algorithm in the classification process of heart failure patient data.