Authors: Assistant Professor M.Sai Manasa, L.srivani
Abstract: Hypertension is one of the leading chronic diseases responsible for cardiovascular disorders and premature mortality across the world. Detecting individuals who are at risk during the early stages can significantly reduce severe health complications through timely medical intervention and lifestyle modification. This study presents a machine learning-based framework for predicting hypertension by analyzing demographic, clinical, and lifestyle-related attributes collected from a publicly available dataset. The proposed work investigates the performance of Decision Tree, Gradient Boosting, Extreme Gradient Boosting (XGBoost), and Random Forest classifiers. To improve predictive capability, are incorporated along with appropriate preprocessing and cross-validation strategies. The experimental findings indicate that feature optimization considerably enhances classification performance by reducing redundant information and improving model generalization. Among the evaluated algorithms, Gradient Boosting and XGBoost demonstrate superior prediction capability, while the hybrid learning approach further improves robustness and classification facilitating personalized hypertension management through data-driven decision-making.
