Comparison of Classifiers for the Risk of Diabetes Prediction

This paper applied a use of algorithms to classify the risk of diabetes mellitus. Four well known classification models that are Decision Tree, Artificial Neural Networks, Logistic Regression and Naive Bayes were first examined. Then, Bagging and Boosting techniques were investigated for improving the robustness of such models. Additionally, Random Forest was not ignored to evaluate in the study. Findings suggest that the best performance of disease risk classification is Random Forest algorithm. Therefore, its model was used to create a web application for predicting a class of the diabetes risk.