Continuous Blood Pressure Estimation via Gaussian Regression Model with the Bayesian Optimization

This paper proposes a method for performing the continuous blood pressure estimation. First, the PPGs are normalized. Second, the noise is reduced via the fast Fourier transform filtering approach. Third, the baseline drift is eliminated via the fourth order polynomial. Fourth, the demographic properties of the subjects, the characteristic points in both the first order derivative and the second order derivative of the PPGs as well as the statistics of the PPGs are employed as the features. Fifth, the relief algorithm is used to select the features. Finally, the Gaussian regression model is employed for performing the blood pressure estimation. Here, the model parameters are found via the Bayesian optimization. The computer numerical simulation results show that our proposed method yields the best performance in terms of the mean squares error (MSE), the root mean squares error (RMSE), the mean absolute error (MAE) and the R squares value compared to the existing methods.