Modelling Arterial Pressure Waveforms Using Gaussian Functions and Two-Stage Particle Swarm Optimizer

Table 3

Statistical results of the computation time from 20 volunteers with different target MAEs for the four methods.

MAE (%)	CAPW (s)				RAPW (s)
	Nelder-Mead	MPSO	DMS-PSO	TSPSO	Nelder-Mead	MPSO	DMS-PSO	TSPSO
10	1.4±0.05^∗	0.9±0.04^∗	1.0±0.03^∗	0.6±0.02	1.4±0.04^∗	0.9±0.04^∗	1.0±0.04^∗	0.7±0.02
8	1.5±0.06^∗	1.0±0.05^∗	1.3±0.05^∗	0.8±0.03	1.5±0.06^∗	1.0±0.05^∗	1.2±0.04^∗	0.7±0.04
6	1.7±0.06^∗	1.5±0.07^∗	1.4±0.06^∗	1.1±0.03	1.7±0.07^∗	1.1±0.05^∗	1.4±0.05^∗	1.0±0.04
4	Inf.	4.2±0.12^∗	2.4±0.09^∗	1.3±0.06	Inf.	3.6±0.12^∗	2.4±0.10^∗	1.2±0.05
3	Inf.	5.5±0.19^∗	3.7±0.16^∗	1.4±0.07	Inf.	5.3±0.16^∗	3.4±0.16^∗	1.3±0.06
2	Inf.	7.4±0.23^∗	5.5±0.21^∗	1.5±0.07	Inf.	7.0±0.21^∗	5.3±0.23^∗	1.4±0.08
1.5	Inf.	Inf.	10.3±0.32^∗	2.0±0.10	Inf.	Inf.	9.7±0.29^∗	1.6±0.09

Data are expressed as mean ± standard deviation (SD) or as symbol “Inf.” “Inf.” indicates that the method could not achieve the target MAE even with the extremely large max_FEs (100000). MAE: mean absolute error; CAPW: carotid artery pressure waveforms; RAPW: radial artery pressure waveforms.
∗There is statistically significant difference when compared with TSPSO (𝑃<0.05).