Predicting Diabetes Mellitus Using Artificial Neural Network Through a Simulation Study

Abstract

Diabetes mellitus (DM) is a diverse group of metabolic disorders that is frequently associated with a high disease burden in developing countries such as Nigeria. It also needs continuous blood glucose monitoring and self-management. This research is aimed to predict diabetes mellitus using artificial neural network. In this research, 100 patients were considered from Ahmadu Bello University Teaching Hospital who have undergone diabetes screening test and 29 risk factors were used. Back propagation algorithm was used to train the artificial neural network for the original and simulated data sets. The results show that the models achieved 98.7%, 57.0%, 73.3%, and 63.0% accuracy for training the original, simulated at 100, simulated at 150 and simulated at 200 data sets respectively. The results also shows that the areas covered under receiver operating curves are 0.997, 0.587, 0.849 and 0.706 for training the original, simulated at 100, simulated at 150 and simulated at 200 data sets respectively. The research therefore concludes that in order to predict diabetes mellitus in patients, the simulated data can be used in place of the original data since the simulated ANN models have been able to discriminate between diabetic and non-diabetic patients.