OPTIMASI ALGORITMA NAÏVE BAYES BERBASIS PARTICLE SWARM OPTIMIZATION (PSO) DAN STRATIFIED UNTUK MENINGKATKAN AKURASI PREDIKSI PENYAKIT DIABETES

Abstract

Diabetes mellitus is a disease that causes uncontrolled blood sugar levels due to a lack of insulin levels in the body. Based on data, the results of diabetes laboratory tests can be predicted with data mining that can help medical personnel. Data mining is a process of identifying data to become information and decisions. This study uses the Naïve Bayes algorithm based on Particle Swarm Optimization (PSO) and Stratified. The results of the Naïve Bayes algorithm get an accuracy value of 75.40% and an AUC value of 0.829%. Meanwhile, the results of the Naïve Bayes algorithm based on Particle Swarm Optimization (PSO) and Stratified get an accuracy value of 90.00% and an AUC value of 0.926. From this study, the Particle Swarm Optimization (PSO) and Stratified based Naïve Bayes algorithm obtained a higher accuracy value with an increase of 14.60% in predicting diabetes disease.

Keyword : Diabetes, Data Mining, Naïve Bayes, PSO, Stratified