PREDIKSI PENYAKIT DIABETES MENGGUNAKAN MACHINE LEARNING DENGAN ALGORITMA NAÏVE BAYES

Abstract

Diabetes Mellitus (DM) is a chronic disease characterized by hyperglycemia and glucose intolerance that occurs because the pancreatic glands are unable to produce insulin adequately or because the body cannot use insulin produced effectively. Data at Harapan Mulya Pharmacy is influenced by the number of patients who perform health checks such as diabetes mellitus so that it has an effect in terms of data classification that will make it difficult for the dispensary. By utilizing data mining. classification to determine patients who have done the examination including diabetics or not. This study aims to create a predictive model using the Nae Bayes Algorithm which produces classification and prediction of diagnosis of diabetes mellitus done using Rapidminner so that prevention against diabetes can be done as soon as possible. accuracy of diagnostic classification. Diabetes mellitus using Nae Bayes algorithm of 95,94% with a true positive score with diabetes with a precesion value is 97,64% and false negative 94,54% of 1035 diagnoses of diabetes. Nae Bayes Algorithm method used, then the resulting prediction model has a true negative recall class of 93,63% and a false negative of 97,99%. Nilai accuracy dari model Algoritma Nae Bayes for prediksi diagnose penyakit diabetes adalah  95,94%, Data yang diprediksi positif sebanyak 487 true positip dan 548 true negative