Table 2.

Mean values of MSE, R.² and computational time for artificial neural network models with different configurations

			5 nodes	15 nodes	30 nodes
Period	MSE	Train	0.0015	0.0008	0.0004
		Valid	0.0017	0.0011	0.0006
		Test	0.0016	0.001	0.0006
	R2		0.7124	0.8253	0.9057
	Comp. Time (s)		0.2263	0.3642	0.5568
Limb Lengths	MSE	Train	0.000142	0.000107	0.00007
		Valid	0.000145	0.000116	0.000084
		Test	0.000145	0.000116	0.000083
	R2		0.9923	0.994	0.9959
	Comp. Time (s)		1.4087	2.9573	8.66
Orientation	MSE	Train	0.0048	0.0038	0.0027
		Valid	0.0049	0.0041	0.0031
		Test	0.0049	0.0041	0.0032
	R2		0.9877	0.9901	0.9927
	Comp. Time (s)		0.9615	2.7845	11.1782
Ankle Joint Angle	MSE	Train	0.00257	0.00217	0.00168
		Valid	0.00265	0.00239	0.00204
		Test	0.00265	0.00237	0.002
	R2		0.8868	0.9019	0.9208
	Comp. Time (s)		21	398	1425
Knee Joint Angle	MSE	Train	0.0042	0.00342	0.00252
		Valid	0.00441	0.00378	0.00308
		Test	0.00437	0.00378	0.00303
	R2		0.9632	0.9692	0.9764
	Comp. Time (s)		31	262	1131
Hip Joint Angle	MSE	Train	0.00172	0.00137	0.00105
		Valid	0.00186	0.00154	0.00128
		Test	0.00179	0.00151	0.00126
	R2		0.9812	0.9847	0.9879
	Comp. Time (s)		46	219	1034
Pelvis-Body Joint Angle	MSE	Train	0.00044	0.0004	0.00035
		Valid	0.00044	0.00042	0.00038
		Test	0.00045	0.00041	0.00039
	R2		0.7164	0.7397	0.7666
	Comp. Time (s)		39	350	1510
Vertical GRF	MSE	Train	0.00355	0.00309	0.00257
		Valid	0.00354	0.0032	0.00292
		Test	0.00361	0.00332	0.00299
	R2		0.9832	0.985	0.9871
	Comp. Time (s)		33	588	1555
Horizontal GRF	MSE	Train	0.00032	0.00029	0.00024
		Valid	0.00034	0.00032	0.00028
		Test	0.00033	0.00031	0.00027
	R2		0.9643	0.9669	0.972
	Comp. Time (s)		27	209	1346