DIABETES PREDICTION USING DIFFERENT ML MODELS
Abstract - Diabetes is one of the biggest causes of mortality, disability, and economic loss worldwide. Type 2 diabetes is the most frequent type of diabetes (90–95 percent globally). However, it can be avoided or postponed by receiving competent treatment and therapies, which include an early diagnosis. There has been significant progress in the field of various machine learning methods, especially for diagnosis. Several computerized information systems for predicting and diagnosing diabetes were described, each utilizing a different classifier. Choosing valid classifiers clearly improves the system’s precision and proficiency. During this study, we present an approach that mixes four classifiers (i.e., a random forest classifier, a decision tree classifier, a logistic regression model, and the K-nearest neighbor model) to diagnose diabetes. The accuracy obtained for the Random Forest Classifier is 90.47 percent, which is higher than that of Logistic Regression, K-Nearest Neighbor, and Decision Tree.
Keywords - Diabetes prediction, Hybridization, Decision Tree, Logistic Regression, K-Nearest Neighbor, Random Forest Classifier.