Hyperparameter Optimization of Decision Tree Using Enhanced Grey Wolf Optimization for Diabetes Prediction

Diabetes is one of the most prevalent chronic diseases, requiring accurate predictive approaches to support diagnosis and clinical decision-making. This study aims to enhance the performance of the Decision Tree (DT) model through hyperparameter optimization using Enhanced Grey Wolf Optimization (EGWO). The proposed method integrates Lévy flight, adaptive exploration–exploitation control, and local search mechanisms to prevent premature convergence and improve solution stability. Experiments were conducted on the Pima Indians Diabetes Dataset (PIDD) using 10-fold cross-validation. The results demonstrate that the DT model optimized with EGWO achieved an accuracy of 92.11%, outperforming various benchmark methods. These findings indicate that Enhanced Grey Wolf Optimization effectively balances exploration and exploitation in hyperparameter search, enabling the Decision Tree to achieve superior classification performance compared to conventional approaches.