An efficient prediction of diabetes using artificial neural networks
Awab Fakih, A. Narasima Venkatesh, V. Vani, Mohd Naved, Pravin R. Kshirsagar, P. Vijayakumar
Abstract
Diabetes in the population is one of the world's main diseases. Diabetes is a chronic condition if there is either insufficient insulin in the pancreas or insulin cannot be used properly in the body. Long-term risks such as cardiac disease and liver cancer could lead to diabetes if not treated. Therefore, a timely prevalence of diabetes is very important to people around the world. In particular, the challenges to newer people and the workforce were identified as diabetes. Diabetes is tracked if patient changes in diet and lifestyle in the early days could be identified. Type 1 and type 2 diabetes were one of the most common forms of diseases; however, other types of diabetes, such as gestational diabetes pregnancy and other forms of diabetes. This paper outlines a method for predicting early diabetes, taking important risk effects into consideration. This approach uses the strategy to display a feature selection procedure. In order to construct a good subset to improve predictive performance, the selected features are assessed through the rating of design models. Ultimately, these characteristics are learned from the neural network classifier and defined.