A Comparative Analysis of Machine Leaming-Based Classifiers for Predicting Diabetes

Diabetes, caused by insulin resistance, is a global health issue. Mismanagement of this chronic condition is impacted by a sedentary lifestyle, aging, food, dietary habits, and elevated blood pressure, which can cause serious complications. These include heart and vascular disease, renal impairment, neuropathy, and visual difficulties. Thus, early diabetes detection and management are crucial. Machine learning can provide essential insights from diabetic patients' diagnostic medical data. This work applies supervised machine learning algorithms to PIMA dataset for optimum diagnosis accuracy. In this paper, we implemented eight machine learning algorithms including Logistic Regression, Decision Tree, ADA Boost, Gradient Boosting, K nearest Neighbor (KNN), Random Forest (RF), Support Vector Machine (SVM), and Naive Bayes on PIMA database and compared them for early diabetes prediction using a variety of evaluation criteria to evaluate the prediction results and to predict the best model.