Prediction of Diabetes Mellitus Using Ensemble Technique for The Enhancement of Machine Learning Algorithms
Diabetes mellitus is a chronic metabolic disorder that requires early and accurate prediction for effective management and prevention. However, conventional machine learning models often struggle with limited predictive accuracy and issues of class imbalance. This study aimed to enhance the predictive performance of diabetes diagnosis models and address class imbalance by applying an ensemble learning approach. Three machine learning algorithms: Logistic Regression, K-Nearest Neighbor (KNN), and Random Forest were implemented and combined using the Gradient Boosting technique. The dataset, obtained from Kaggle, underwent preprocessing, feature scaling, and balancing using the Synthetic Minority Over-sampling Technique (SMOTE) to improve model reliability. Model performance was evaluated using accuracy and area under the curve (AUC) metrics. The ensemble model achieved 98% accuracy and 99% AUC, outperforming all individual algorithms. These findings demonstrate the effectiveness of Gradient Boosting in enhancing predictive accuracy and model stability, offering a reliable framework for early diabetes detection and supporting healthcare professionals in informed decision-making.
Download Paper