Myoelectric Pattern Recognition Performance Enhancement Using Nonlinear Features

Table 4

The EMG pattern recognition performance enhancement of existing feature extraction methods by the LMAV and NSV when SVM classifier is used.

	Parameter	Group	FS1	FS2	FS3	FS4
Dataset 1	Accuracy	Group 1	95.73 ± 0.93	95.14 ± 1.06	95.53 ± 0.93	95.66 ± 1.02
		Group 2	96.74 ± 0.69	95.81 ± 0.84	95.84 ± 0.90	96.02 ± 1.00
	Sensitivity	Group 1	78.65 ± 4.67	75.70 ± 5.30	77.63 ± 4.64	78.30 ± 5.08
		Group 2	83.68 ± 3.43	79.06 ± 4.20	79.19 ± 4.50	80.10 ± 5.02
	Specificity	Group 1	97.63 ± 0.52	97.30 ± 0.59	97.51 ± 0.52	97.59 ± 0.56
		Group 2	97.97 ± 0.46	97.44 ± 0.54	97.44 ± 0.55	97.64 ± 0.65
	Precision	Group 1	79.42 ± 5.00	77.11 ± 5.13	78.83 ± 4.81	79.61 ± 5.35
		Group 2	84.59 ± 3.55	80.28 ± 3.91	80.21 ± 4.71	81.20 ± 5.11
	F1 score	Group 1	78.08 ± 4.84	75.06 ± 5.41	76.97 ± 4.83	77.71 ± 5.32
		Group 2	83.18 ± 3.62	78.43 ± 4.27	78.67 ± 4.69	79.49 ± 5.22
Dataset 2	Accuracy	Group 1	96.58 ± 0.97	96.00 ± 0.85	95.88 ± 1.00	95.32 ± 1.29
		Group 2	97.51 ± 0.76	96.63 ± 0.78	96.73 ± 0.71	95.95 ± 0.92
	Sensitivity	Group 1	82.88 ± 4.86	80.01 ± 4.24	79.42 ± 4.99	76.60 ± 6.44
		Group 2	87.53 ± 3.79	83.13 ± 3.90	83.63 ± 3.56	79.75 ± 4.59
	Specificity	Group 1	98.10 ± 0.54	97.78 ± 0.47	97.71 ± 0.55	97.40 ± 0.72
		Group 2	98.61 ± 0.42	98.13 ± 0.43	98.18 ± 0.40	97.75 ± 0.51
	Precision	Group 1	83.92 ± 4.92	81.51 ± 4.05	80.75 ± 5.01	78.12 ± 6.58
		Group 2	88.56 ± 3.49	84.38 ± 3.51	84.85 ± 3.38	81.33 ± 4.38
	F1 score	Group 1	82.63 ± 5.01	79.78 ± 4.34	79.00 ± 5.21	76.17 ± 6.70
		Group 2	87.40 ± 3.85	82.98 ± 3.98	83.39 ± 3.68	79.51 ± 4.67