Type 2 Diabetes Classification and Prediction: An Ensemble and Hyper Parameter Optimization Approach in Machine Learning

This work presents the machine learning-based classification and prediction of T2DM (Type 2 Diabetes Mellitus) via the use of an ensemble and hyper parameter optimization approach. The study starts out with a thorough literature analysis that identifies the positive and negative aspects of current T2DM prediction techniques. The main goal is to build a strong ensemble model that incorporates several machine-learning techniques to improve prediction accuracy. Performance metrics for different machine learning models used to classify Type 2 Diabetes are shown in the table. At 78.57% accuracy, the Voting Ensemble demonstrated superior precision and memory. Decision Tree came in second, with an accuracy of 76.62%. In this challenge, Support Vector Machines and Logistic Regression both performed admirably. Methods for hyper parameter optimization have been employed to fine-tune the ensemble model in order to enhance its functionality. Measuring important parameters including precision, recall, F1-score, and accuracy makes it possible to conduct an empirical evaluation of real-world T2DM datasets. The study finds a gap in the literature that presents opportunities for future improvements in the state of T2DM prediction. This research advances early T2DM diagnosis and ultimately enhances patient treatment.