Diagnosis And Forecast Of Diabetes Mellitus Using Neural Network

Abstract

One among the chronic disease that is estimated in the year of 2015 is Diabetes. It projects 415 million
people was affected with diabetes and estimated to have the death rate of 5 million each and every decade.
The tool forecasted and utilized to determine the person is having diabetes or not. The proposed method is
used for accurate detection with the technique of deep learning. There is particularly an important
technique for exact prediction such as neural network using Back propagation. The input layer has eight
parameters as a first layer, hidden layer with 10 neurons & 1 output layer is used to display the accurate
outcome. An interactive tool is implemented by graphical user interface for getting exact outcome when the
absence of medical expert. The medical expert will get the patients reports within a second and it will be
used for saving the time for the patients to get further treatment. In this paper, MATLAB is used for
implementing and developing the results of the identification of diabetic patients. This can easily predict
that the patient is having diabetes or not. The accuracy of the given dataset is predicted as 85% by using the
deep learning technique, which is the improvement of the existing work.