Diabetes mellitus early stage risk prediction using machine learning algorithms

Abstract

Diabetic patients are on the rise. Diabetes is one of the most debilitating diseases. Undiagnosed and untreated diabetes can lead to a number of health issues, including heart disease and stroke. It is necessary for the patient to visit a diagnostic institution and contact a doctor. With the advent of machine learning, this important problem has been overcome. A primary objective of this work is to build a model that can reliably predict a person's probability of developing diabetes. To detect diabetes at an early stage, six supervised machine learning classification methods and a hybrid model based on the top three findings are employed. UCI's machine learning repository provides access to the Pima Indians Diabetes Database, which is used in the experiments. All of them are evaluated based on a variety of measures. It is highlighted that the hybrid model which got an accuracy of 90,62% performs better than other state-of-the-art methods.