Enhanced Approach for Chronic Disease Diagnosis and Prediction Using Ensemble Deep Learning
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Chronic kidney disease (CKD) poses a major global health challenge due to its silent progression and late-stage diagnosis. Early detection is essential to prevent complications and improve patient outcomes. This paper proposes a diagnostic framework that integrates machine learning (ML) and explainable artificial intelligence (XAI) techniques to enhance CKD prediction accuracy and transparency. Two ensemble-based stacking classifiers were developed: one leveraging bagging-like methods and the other boosting-like methods, aiming to harness their complementary strengths. The framework was validated using the University of California, Irvine (UCI) CKD dataset, incorporating key biomarkers relevant to diagnosis. Rigorous data preprocessing, including balancing, normalization, and careful prevention of data leakage, ensured robust model generalizability. Hyperparameter tuning and regularization techniques minimized overfitting risks. The models achieved outstanding results: the ExtraTrees Classifier and stacking models reached perfect scores (100%) in accuracy, precision, recall, and F1 score, while the Random Forest Classifier achieved 99.82% accuracy, 97.05% precision, 98.33% recall, and a 98.63% F1 score on test data. To enhance interpretability, XAI techniques such as Local Interpretable Model-Agnostic Explanations (LIME) and Shapley Additive Explanations (SHAP) were employed, providing instance- and global-level insights. Key features such as serum creatinine, blood pressure, and specific biomarkers were identified as critical predictors. Interpretability was further evaluated using stability, prescriptiveness, and fidelity metrics, ensuring clinical relevance and actionable insights. This research highlights the potential of combining ML and XAI to create a robust, transparent, and interpretable framework for early CKD detection, supporting improved patient care across diverse clinical environments.
Fields of Study & Subjects
Artificial Intelligence in Healthcare
Explainable Artificial Intelligence (XAI)
Chronic Kidney Disease and Diabetes
Publication details
- Journal / source
- CRC Press eBooks
- Pages
- 124–181
- Publication date
- 31 March 2026
- Language
- EN
- Document type
- Book chapter