Design of Adaptive Filter Using Jordan/Elman Neural Network in a Typical EMG Signal Noise Removal : Table 3

Table 3: For hidden layers = 4, input PE = 1, output PE = 1, transfer function = TanhAxon, learning rule = momentum, maximum epochs = 1000, and threshold MSE = 0.01.


Serial no.	Type of ANN	Hidden layer variation H1 H2 H3 H4	Correlation coefficient	MSE				(s)
Serial no.	Type of ANN	Hidden layer variation H1 H2 H3 H4	Correlation coefficient	Training	Cross validation	Testing		(s)

01	Jordan/Elman network	07,11,12,05	0.771 344 418	0.006 039 548	0.013 371 967	0.012 071 403	1.5151	25.6

02	Jordan/Elman network	07,11,12,05	0.78 827 949	0.008 325 698	0.017 142 589	0.013 621 869	1.5151	23.2

03	Jordan/Elman network	06,09,09,05	0.785 489 978	0.009 970 153	0.019 572 975	0.014 487 733	2.21 675	35.7