A new hybrid learning model for early diagnosis of hypertension using IoMT technologies

Hypertension is high blood pressure that occurs when the heart is pumping blood. In this study, an e-diagnosis IoMT application has been developed to enable early detection of hypertension. To diagnose and detect hypertension with machine learning algorithms, data from the Behavioral Risk Factor Surveillance System(BRFSS) database were used. Ten different features were identified that can be used to diagnose whether individuals have hypertension or not. By applying parameter optimization for Decision Tree, K-Nearset Neighbor, Random Forest, Support Vector Machine, and Naive Bayes machine learning algorithms with 5 fold cross-validation, the success rates obtained from these models were examined by considering the class imbalance problem. To achieve a better success rate, a hybrid learning model was proposed. The highest success rate of 98.5 % was achieved with the proposed model, thus developing a diagnosis and detection IoMT system that can help doctors with hypertension disease.