Performa Klasifikasi K-NN dan Cross Validation pada Data Pasien Pengidap Penyakit Jantung

Abstract

Globally, the number one cause of death each year is cardiovascular disease. Cardiovascular disease is a disease caused by impaired function of the heart and blood vessels, such as coronary heart disease, heart failure or heart failure, hypertension and stroke. The purpose of this study was to measure the performance of accuracy, precision, recall and f-measure of the K-NN and Crossvalidation methods on a dataset of cardiovascular patients. The dataset used was 1000 records consisting of 11 attributes (age, gender, height, etc.) cardiovascular and non cardiovascular patient data, the dataset was obtained from the UCI Machine Learning Repository managed by the Hungarian Institute of Cardiology Budapest: Andras Janosi, MD, University Hospital, Zurich, Switzerland. The steps taken are: dividing the simulation ratio of the dataset to 20:80, 50:50 and 80:20, applying crossvalidation (k-fold = 10) and classification using the K-NN method (k = 2 to K = 900). The research results from the simulation of the dataset ratio 50:50 obtained an accuracy value of 82%, 82% precision, 82% recall and 80% f-measure at a value of K = 13, then the research results from the simulation of the dataset ratio 20:80 obtained an accuracy value of 87%, 87% precision, 97% recall and 92% f-measure at the value of K = 3, and the results of research from the simulation of the dataset ratio 80:20 obtained an accuracy value of 91%, 92% precision, 60% recall and 72% f-measure at the value K = 5.