Prediction Of Early-Stage Diabetes using machine learning

Diabetes is one of the most chronic diseases for many years. It occurs when the pancreas loses its functionality, where the production of insulin will no longer be sufficient to move the blood glucose to the cells, or when the cells stop reacting with the produced insulin. People that are overweight, living a bad lifestyle, and are lazy and inactive are more likely to be diabetic. Diabetes has multiple symptoms, and the most common symptom is Polyuria which is the production of a large amount of urine. Moreover, other symptoms can appear such as Polydipsia, sudden weight loss, and visual blurring. In Healthcare industries, Data-driven decision-making (DDDM) has been incredibly used for predicting diseases due to its high accuracy and low cost, it will reduce the health risk of the patients by obtaining its symptoms. Machine learning has been implemented significantly in predicting diseases, especially the chronic disease. In this existing work, the utilized machine learning models namely are, a Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Multilayer Perceptron (MLP), and Radial-Basis Neural Network (RBF) to predict diabetes using Polyuria, Polydipsia, and sudden weight loss from the diabetes risk prediction dataset. K-Nearest Neighbors (KNN) shows better performance compared to the other classifiers with f1-score of 0.98%.