An ANN Model for Early Prediction of Diabetes

Abstract

Diabetes a metabolic disease with the botanical name diabetes mellitus is diagnosed in a person who has high sugar levels in the bloodstream which could be either because of the cells not reacting to the insulin that is created or the pancreas does not deliver insulin by any stretch of the imagination. According to research (World Health Organization, Geneva 2014), Worldwide 194 million people have already been diagnosed with diabetes and this rate is expanding quickly and is estimated to achieve 333 million by 2025. In Africa, more than 5 million people are already diagnosed with diabetes, and it has been estimated that by 2025 diabetes patients in the continent would be up to 15 million. (KMV Narayan). Ogbera also reported that in Nigeria patients confirmed diabetes up to 158 million, and as such, the need to study the prediction of diabetes is very paramount. In this way, there is an incredible need to concentrate on the prediction of diabetes. So that precautions can be taken to control this deadly disease. This research work wishes to present the prediction of diabetes using ANN. The model used in this work considered a group of fifteen factors and identified the factors that are very influential in the diagnosis of diabetes using regression analysis so as to achieve better accuracy of prediction.