Internet of Things Based-Early Diagnosis System of Diabetes Using Optimized Neural Network

Abstract

One of the most common diseases today is diabetes. This disease has an initial phase, if it is diagnosed on time, it helps to control and treat this disease to a great extent. But early diagnosis of this disease is very challenging due to the difficulty of repeated tests, especially in some geographical areas. Therefore, in smart health systems, a series of information is frequently taken from users by mobile phone, and by analyzing this information using artificial intelligence and machine learning algorithms and with the help of the Internet of Things, the doctor can diagnose diabetes in the first steps remotely. In this paper, we used a database that has 16 features related to early diabetes. Our proposed method for early diagnosis of diabetes is based on feature ranking along with MLP neural network optimized with whale optimization algorithm (WOA). also, the NCA algorithm is used to rank the features, and the WOA algorithm is used to optimize the parameters of the MLP neural network. Finally, based on the simulation results of the proposed method, we achieved 98.1% accuracy in the early diagnosis of diabetes.