Diabetes Mellitus Early Detection Simulation using The K-Nearest Neighbors Algorithm with Cloud-Based Runtime (COLAB)

Abstract

Diabetes Mellitus is a genetically and clinically heterogeneous metabolic disorder with manifestations of loss of carbohydrate tolerance characterized by high blood glucose levels as a result of insulin insufficiency. Public knowledge of diabetes mellitus 39.30% is influenced by public health education and information about diabetes mellitus that the public has ever received. Early detection of diabetes mellitus can prevent the development of chronic complications and allow timely and rapid treatment. The aim of this study is to simulate the early detection of diabetes mellitus with the K-Nearest Neighbors (K-NN) algorithm using Cloud-Base Runtime (COLAB). The highest accuracy is 76% in K=3, the highest precision is 68% in K=3 and the highest recall is 60% in K=3. The researchers used K-NN as a method to classify data from the Pima Indians Diabetes Database and obtained a fairly good accuracy value of 76% with a value of k = 3.