The Performance Analysis of Several Machine Learning Methods in Predicting Diabetes

Diabetes is a chronic disease characterized by hyperglycemia. Early screening of diabetes patients will help reduce the incidence of diabetes. Accurately modeling the early symptoms of diabetes to reduce the incidence of diabetes has become an urgent problem. Therefore, it is a valuable research direction to mine and study the information of diabetes patients, predict the occurrence of noninfectious diseases, and assist doctors to make correct diagnoses. The precision of different data mining techniques has been compared in this research, employing machine learning techniques to predict the incidence of diabetes. The University of California, Irvine’s open-source dataset has been used as the main database for experiment. Different machine learning classifiers such as random forest, neural network and support vector machine have been employed in this experiment. This research gives a technology review of machine learning algorithms that can be used for diabetes prediction through classification. The experimental results indicate that random forest performs better than other machine learning methods.