Artificial Neural Network based approach to Diabetes Prediction using Pima Indians Diabetes Dataset

Diabetes is a serious health condition that is described as having high blood sugar for long periods of time. It is a rapidly growing disease affecting people all over the world. For effective treatment, detection of the disease at the early stages is very crucial. In this work we have proposed a system available to the end-users to detect the presence of diabetes using different health parameters. We have employed Artificial Neural Network to design the prediction classifier for this system. ANN is chosen because it does not require the explicit application of feature extraction, which is done automatically at the time of training the ANN Model. This results in significant reduction of training time which is more prominent while training with bigger datasets. This reduction in training time has little to no effect on the classification performance which is evident from the model's classification accuracy of 0.8079 when compared to similar works using feature extraction.