SUPERVISED LEARNING METODE K-NEAREST NEIGHBOR UNTUK PREDIKSI DIABETES PADA WANITA

Abstract

Supervised learning is a technique of machine learning by doing learning that has a reference value to direct something, one of which is the K-Nearest Neighbor (KNN) method. This method is for object classification through learning data that is closest to the object (neighbor) using euclidean distance to calculate the distance. KNN can be used for data classification that already has a reference, in this case the dataset used is the diabetes mellitus dataset in women. DM is a disease that can cause complications in parts of the body that cause death. DM in women can be seen from several parameters such as glucose levels, blood pressure, skin thickness, insulin hormone, body index mass, age, number of pregnancies, and the number of family history of diabetes. In this research, KNN will be used for the classification of diabetes in women with two classes, namely DM Positive and DM Negative, in other words, a woman can be predicted to suffer from DM disease or not. This method will be implemented into a system with PHP programming language and Codeigniter Framework. KNN testing is carried out with three test scenarios, the 1st test with 150 test data gets an 82% accuracy rate, the 2nd data test with 200 test data gets an 84% accuracy rate, and the 3rd data test with 300 test data gets an 82% accuracy rate.