OPTIMASI PREDIKSI PENYAKIT JANTUNG DENGAN NAÏVE BAYES DAN PARTICLE SWARM OPTIMIZATION (PSO)

Abstract

Heart disease is a chronic and highly fatal condition that is common in both industrialized and non-industrialized countries, including Indonesia. Looking at the report from the Indonesian Health Survey (IHS), the state of heart disease in Indonesia is at an alarming level. The disease is spread in almost all regions in Indonesia. Heart disease has no age limit. According to the Indonesian Health Survey (SKI), about 0.8% of every 100 Indonesians suffer from heart disease. Optimization of the use of Naïve Bayes as an algorithm used to predict heart disease is needed in the world of health so that the results obtained from the use of the Naïve Bayes algorithm are more optimal. This study aims to determine whether the use of Particle Swarm Optimization (PSO) metrode for feature selection will be able to improve the efficiency of the features of heart disease prediction results, experiments have been carried out using datasets and get a confusion matrix of 79.12%. Better results were obtained after applying feature selection optimization using PSO to perform feature selection on the dataset and the results obtained increased to 86.37%.