<p>A Comparative Study of Machine Learning Models for Hypertension Risk Assessment With Explainable Artificial Intelligence Integration</p>

Hypertension is a major global health challenge and one of the leading causes of cardiovascular diseases, stroke, kidney failure, and premature death. Early identification and prediction of hypertension risk severity are essential for improving patient outcomes and supporting effective clinical decision-making. However, conventional diagnostic approaches often struggle to capture the complex relationships between clinical and lifestyle-related risk factors associated with hypertension. To address this challenge, this study proposes an explainable machine learning framework for identifying and predicting hypertension risk severity using patients’ clinical and lifestyle data collected from the University of Uyo Teaching Hospital. The proposed framework incorporates data preprocessing techniques such as missing value handling, data normalization, and feature ranking using Principal Component Analysis to improve data quality and model performance. Several machine learning algorithms, including eXtreme Gradient Boosting (XGBoost), Random Forest, Artificial Neural Network, Support Vector Machine, and Decision Tree, were implemented and evaluated using stratified 10-fold cross-validation to reduce overfitting and ensure reliable model evaluation. Performance assessment was carried out using accuracy, precision, recall, F1-score, and area under the curve. The experimental results revealed that XGBoost and Random Forest achieved the highest accuracy of 98%, followed by Support Vector Machine with 95%, Artificial Neural Network with 94%, and Decision Tree with 89%. To enhance transparency and interpretability, Explainable Artificial Intelligence was integrated into the framework using Local Interpretable Model-Agnostic Explanations to identify the most influential features contributing to predictions. XGBoost emerged as the best-performing model due to its ability to effectively learn complex patterns within the dataset. The findings demonstrate that explainable machine learning models can significantly improve hypertension risk prediction while also promoting interpretability, trust, and informed decision-making for healthcare professionals and patients.