Evaluation of machine learning techniques for hypertension risk prediction based on medical data in Bangladesh

Abstract

Hypertension in Bangladesh is a leading cause of cardiovascular diseases, stroke, and kidney failure, resulting in significant morbidity and mortality. Preventive measures and simple health practices can effectively reduce hypertension and its complications. This study utilizes machine learning algorithms (Naive Bayes, support vector machine, logistic regression, random forest) to predict hypertension in high-risk individuals. The proposed hybrid model achieves a prediction accuracy of 78.17%, surpassing other machine learning methods. Random forest has the highest accuracy among the individual algorithms at 73.86%. Classification performance is evaluated using sensitivity, specificity, precision, and F-score, along with receiver operating characteristic analyses and confusion matrices through 10-fold cross-validation. These findings emphasize the importance of managing risk factors for better population health and highlight the efficacy of the hybrid model in hypertension prediction. The study underscores the significance of preventive measures in reducing the burden of hypertension-related diseases and improving overall well-being.