查看文章

Diabetes, a common metabolic disease with serious health effects, is the focus of this investigation. A unique method to increase predicted accuracy is presented in the study. We use ensemble learning methods like Grey Wolf Optimization (GWO) with Adaboost and Ant Colony Optimization (ACO) with XGBoost. After data preparation, GWO and ACO algorithms pick features, and model training is performed. An analysis of a dataset from the National Institute of Diabetes and Digestive and Kidney Diseases found that Grey Wolf Optimizer (GWO) with AdaBoost outperforms Ant Colony Optimization (ACO) with XGBoost in accuracy, precision, and AUC. Ant Colony Optimization (ACO) using XGBoost improves recall, detecting actual positives more accurately. The models’ slight performance differences emphasize the need to select them depending on healthcare goals. This study shows how ensemble learning and feature selection improve diagnostic accuracy and healthcare decision-making, advancing diabetes prediction models.