Intelligent Noise Removal from EMG Signal Using Focused Time-Lagged Recurrent Neural Network

Table 1

EMG signal datasets used for NN-based models. Total number of samples = 2000.

(a) Datasets based on forward and reverse tagging and % variation

S.N.	Dataset	Tagging	% variation
			(Train-CV-test)
1	Dataset1	Forward (training, cross validation, and testing)	10-15-75%
2	Dataset2		20-15-65%
3	Dataset3		30-15-55%
4	Dataset4		40-15-45%
5	Dataset5		50-15-35%
6	Dataset6		60-15-25%
7	Dataset7		70-15-15%
8	Dataset8		80-15-05%
9	Dataset9	Reverse (testing, cross validation, and training)	10-15-75%
10	Dataset10		20-15-65%
11	Dataset11		30-15-55%
12	Dataset12		40-15-45%
13	Dataset13		50-15-35%
14	Dataset14		60-15-25%
15	Dataset15		70-15-15%
16	Dataset16		80-15-05%

(b) Datasets based on multifold differential learning. Group I: (1–500 samples), Group II: (501–1000 samples), Group III: (1001–1500 samples), Group IV: (1501–2000 samples)

Dataset	Train on group	Test on group	Dataset	Train on group	Test on group
Dataset1	I	II	Dataset18	I + III	II + IV
Dataset2	I	III	Dataset19	I + IV	II
Dataset3	I	IV	Dataset20	I + IV	III
Dataset4	II	III	Dataset21	I + IV	II + III
Dataset5	II	IV	Dataset22	II + III	I
Dataset6	II	I	Dataset23	II + III	IV
Dataset7	III	IV	Dataset24	II + III	I + IV
Dataset8	III	I	Dataset25	II + IV	I
Dataset9	III	II	Dataset26	II + IV	III
Dataset10	IV	I	Dataset27	II + IV	I + III
Dataset11	IV	II	Dataset28	III + IV	I
Dataset12	IV	III	Dataset29	III + IV	II
Dataset13	I + II	III	Dataset30	III + IV	I + II
Dataset14	I + II	IV	Dataset31	I + II + III	IV
Dataset15	I + II	III + IV	Dataset32	II + III + IV	I
Dataset16	I + III	II	Dataset33	III + IV + I	II
Dataset17	I + III	IV	Dataset34	IV + I + II	III