A Comprehensive Study of Deep Learning Techniques to Predict Dissimilar Diseases in Diabetes Mellitus Using IoT

Abstract

India has evaluated 77 million people with diabetes, which makes it the second most elaborated disease in the world. Diabetes is a chronic syndrome that occurs with increased sugar levels in the blood cells. Once diabetes is diagnosed and untreated by physicians, it may affect the internal organs slowly, so there is a necessity for early prediction. Popular Machine Learning (ML) techniques existed for the early prediction of diabetes mellitus. A significant perspective is to be considered in total management by machine learning algorithms, but it is not a good enough model to predict DMT2. Therefore, Deep learning (DL) models are utilized to produce enhanced prediction accuracy. The ML methods are evaluated and analyzed distinctly on the inconspicuous test information. DL is a subpart of ML with many data sets recurrently used to train the system. IoT was another emerging technology-based Healthcare Monitoring System (HMS) built to support the vision of patients and doctors in the healthcare domain. This paper aims to survey ML and DL techniques relevant to Dissimilar Disease prediction in Diabetes Mellitus. Finally, by doing a study on it, deep learning methods performed well in predicting the dissimilar diseases related to diabetes and also other disease predictions using m-IoT devices. This study will contribute to future deep-learning ideas that will assist in detecting diabetic-related illnesses with greater accuracy.