Machine Learning Approaches for Type-2 Diabetes Software Predictor

Abstract

Diabetes is one of the common diseases that affect our health, which results in high glucose level in blood. Diabetes can affect the functioning of various parts of our body including heart, kidney, eyes and nerves. Diagnosis of diabetes is performed by checking the blood sugar level and if detected earlier, controlling will be much easier. Prediction in healthcare field is a challenging task, since timely precautions and decisions are to be taken based on the predicted result, for treatment of the patient. Here, performance and accuracy of the predictive algorithms play a vital role. Machine learning is a popular research area, which finds immense application in medical field and remote healthcare. In this paper we analyze six machine learning algorithms for predicting type-2 diabetes mellitus and perform experiments to choose the algorithm, which gives best accuracy compared to others. We also develop a prediction software (prediction application) which facilitates prediction of type-2 diabetes mellitus, at a very early stage.