Enhancing Type 2 Diabetes Diagnosis with Evolutionary Algorithms and Machine Learning

Type 2 diabetes is a widespread chronic disease that requires early and accurate diagnosis to prevent serious complications. Traditional diagnostic methods often lack sufficient accuracy, which highlights the need for more reliable computational solutions. In this study, we propose a model that integrates the Random Forest classifier with the Bat Optimization Algorithm for simultaneous feature selection and hyperparameter tuning. The SMOTE-ENN method was first applied to the Pima Indians Diabetes Database to correct class imbalance and remove noisy or ambiguous samples, producing a more balanced and cleaner dataset. The optimized model trained on this refined dataset achieved 89% accuracy, 88% precision, 90% recall, and an F1-score of 89%, clearly outperforming the baseline Random Forest and other existing approaches. These results demonstrate the potential of combining evolutionary algorithms with ensemble learning to provide a practical and cost-effective tool for early detection of type 2 diabetes in clinical practice.