The Ensemble Learners: Machine Learning Classification of Diabetic Patient Readmission

Diabetes is a major health problem that affects a large number of people worldwide. We proposed an effective and reliable method to diagnose diabetic readmission in this research based on machine learning models. The model was developed on tree-based ensemble classification algorithms, such as Decision Tree, Extreme Gradient Boost, Ada Boost, and CatBoost. In addition, to increase accuracy, we stacked the models with Stack Classifier using Catboost classifier as the final estimator. We carried out the experiments on the diabetic readmission dataset obtained from the VCI machine learning repository. We used the Grid Search technique to learn the best practices for model evaluation and hyperparameter tuning. The performances of all six algorithms are measured using different metrics such as AUC, Accuracy, and Recall. AUC measures a classifier's ability to distinguish between classes. CatBoost Classifier outperforms other algorithms in terms of AUC, Recall, and Accuracy, with 68 percent, 58.7 percent, and 63.2 percent, respectively. These findings are validated using ROC (Receiver Operating Characteristic) curves correctly and systematically.