A new machine learning method for predicting systolic and diastolic blood pressure using clinical characteristics

Abstract

Listen

Translate article

Hypertension describes elevated blood pressure, which significantly impacts cardiovascular diseases. Typically, a sphygmomanometer, a cuff-like device, is used to measure a patient’s blood pressure. However, new techniques such as phonocardiogram (PPG) and electrocardiogram (ECG) based on cuff signals have been developed. Still, they require complex and expensive multiple sensors. A new machine learning-based method has been proposed to predict both systolic and diastolic blood pressure to overcome this issue. The model considers various clinical characteristics such as gender, blood sugar and cholesterol levels, smoking status, age, alcohol use, weight, and a history of heart disease. A physical activity level metric is used to evaluate the model trained on a dataset of 50,000 blood pressure readings available on Kaggle. Four machine learning techniques, including K-Nearest Neighbors (KNN), logistic regression, decision tree, and random forest, were tested with different training, validation, and testing ratios to enhance the model’s accuracy. The algorithm’s performance was evaluated using accuracy, recall, precision, and F1 scores. Random forest was found to have the highest accuracy and F1 scores.