Application of the K-Nearest Neighbors (K-NN) Algorithm for Classification of Heart Failure

Abstract

Heart failure is a type of disease that has the largest number of patients in the world. Based on information from the data center, there were 229,696 people with heart failure in 2013. Lack of public knowledge about what indications of a person having heart failure make the main cause not handled properly by heart failure patients. In this study, data classification was carried out using KNN algorithm because it has a simple calculation and has a fast time. This study only uses 12 attributes, while the previous study compared 6 algorithms with 13 attributes from 299 data. The highest algorithm with 94.31% accuracy by Random Forest while KNN had an accuracy rate of 86.95% with the same data. In this study, the accuracy of the sample data was compared between 20 data and 299 total data. Both of them have different accuracy. 20 sample data has an accuracy rate of 89.29% while 299 data has an accuracy rate of 96.66%.