Prediction of Diabetics Using Chemical Reaction Optimization

Diabetes Mellitus (DM) is ordinarily conversant like diabetes that is a group of metabolic diseases. It has become a common disease and its degree and prevalence are becoming greater in degree in most inhabitants. Physicians want a reliable system for diagnosing diabetes. Many researchers have used many techniques to diagnose Diabetes. One of the recently established population-based metaheuristic algorithms is Chemical Reaction Optimization (CRO) for problem optimization, which generates from the concept of the nature of chemical reactions. In our study, the main goal of our paper is to find out the best result whether the patient has diabetes or not, using Chemical Reaction Optimization (CRO) with minimum features. We use five different classifiers such as Decision Tree, XGboost, Random Forest, and Support Vector Machine as our objective functions of CRO. We achieved the highest accuracy to predict diabetics compared to the existing methods with optimized features.