Comparison of Classification Algorithm Performance for Diabetes Prediction Using Orange Data Mining

Abstract

Diabetes is a disease that contributes to a relatively high mortality rate. The human death rate due to diabetes is a widespread issue globally. The primary goal of this research is to predict individuals suffering from diabetes using a publicly available dataset from the UCI Repository with the Diabetes Disease dataset. To obtain the best classification algorithm, a comparison is made among three algorithms: KNN, Naive Bayes, and Random Forest, commonly used for predicting diabetes. The comparison results indicate that the Random Forest algorithm is the appropriate and accurate algorithm for predicting individuals with diabetes, with an accuracy rate of 97%.