Hypertension Stage Detection Using Ensemble Machine Learning Algorithms

Hypertension, described by high or rising blood pressure, poses significant health risks and is classified into three stages: stage one, stage two, and hypertensive crisis. Hypertensive crisis, an abnormal form of hypertension, contributes to long-term and short-term consequences, including obesity, diabetes, cardiovascular complications during pregnancy, increased mortality and morbidity in older adults, and the development of chronic diseases later in life. To design the proposed approach, first collect the data from the University of Gondar Hospital and then apply preprocessing techniques such as data cleaning, transformation, and optimization to enhance its quality for machine learning tasks. The ensemble model evaluates 92.79% accuracy, 94.3% precision, 92.07% recall, 92.97% F1 score, and 79.07% ROC score. The experimental results demonstrated that the Bagged Decision Tree ensemble algorithm exhibited superior performance with an accuracy of 92.79%. This study highlights the effectiveness of ensemble learning techniques in the healthcare sector, specifically in identifying the stage of hypertension and reducing morbidity and mortality rates in Ethiopia.