Application of the Evolutionary Programming for Indirect Evaluation of Blood Pressure based on Heart Rate Variability Parameters

The article presents results of testing the hypothesis that it is possible to predicting the blood pressure using features of the heart rate variability. The hypothesis is based on the previous results associated with the task of arterial hypertension diagnostics by means of heart rate variability. For data, one of the PhysioNet collection is used, which contains simultaneously recorded biomedical signals. The evolutionary programming algorithm, which was successfully used in previous tasks, is updated to solve regression task. The presented results show that evolutionary programming can find such set of heart rate variability parameters that can be used for prediction of blood pressure, taking into account calibration data. Comparison with the state-of-art has been done. It was noted that using heart rate variability is a distinctive feature of the proposed approach. It is discussed that some additional data, such as gender, age, anthropomorphic features, as well as parameters of the raw electrocardiography signals can be used to improve results. The future work plans are discussed.