Diabetes Prediction Using Machine Learning Classification Algorithms

Diabetes is a disease that develops when body's glucose levels are too high. If left untreated, it can result in serious complications for a person, including damage to the eyes, kidneys, heart, and other organs. However, its early prediction can help to reduce its adverse effects. Diabetes can be predicted using machine learning. In this paper four classification algorithms including support vector machine (SVM), extreme gradient boosting (XGB), decision tree (DT), and random forest (RF) are used to predict diabetes. PIMA Indian diabetes dataset has been used to perform experiments. Dataset has been balanced using synthetic minority oversampling technique (SMOTE). K-fold validation method is used in experiments. Balancing has increased the performance of all the classifiers. SVM obtained 77.40% accuracy, DT obtained 74.69% accuracy, XGBoost obtained 78.29% accuracy and RF obtained 82.70^% accuracy.