Journal of Healthcare Engineering / 2018 / Article / Tab 2

Research Article

A Novel Neural Network Model for Blood Pressure Estimation Using Photoplethesmography without Electrocardiogram

Table 2

Performance results of different number of the input parameters.

Number of the input parametersSBPDBP
e (mmHg)er (%)e (mmHg)er (%)

3 parameters (10 Hz interval)8.29 ± 6.606.00 ± 5.286.19 ± 6.639.14 ± 7.45
4 parameters (5 Hz interval)7.62 ± 6.085.49 ± 4.774.31 ± 3.833.84 ± 3.39
7 parameters (2 Hz interval)6.59 ± 5.254.72 ± 4.024.34 ± 5.234.72 ± 3.91
12 parameters (1 Hz interval)4.64 ± 3.633.42 ± 2.613.69 ± 2.743.12 ± 4.11
22 parameters (0.5 Hz interval)4.02 ± 2.792.84 ± 2.002.27 ± 1.824.39 ± 3.60