Myoelectric Pattern Recognition Performance Enhancement Using Nonlinear Features

Table 5

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

	Parameter	Group	FS1	FS2	FS3	FS4
Dataset 1	Accuracy	Group 1	94.94 ± 1.04	94.39 ± 1.12	94.95 ± 0.98	95.25 ± 1.15
		Group 2	96.35 ± 0.82	95.39 ± 0.97	95.40 ± 0.99	95.76 ± 1.17
	Sensitivity	Group 1	74.71 ± 5.19	71.94 ± 5.60	74.74 ± 4.90	76.24 ± 5.77
		Group 2	81.76 ± 4.11	76.97 ± 4.84	76.98 ± 4.97	78.80 ± 5.86
	Specificity	Group 1	97.19 ± 0.58	96.88 ± 0.62	97.19 ± 0.54	97.36 ± 0.64
		Group 2	98.19 ± 0.38	97.67 ± 0.47	97.69 ± 0.50	97.79 ± 0.56
	Precision	Group 1	75.96 ± 5.19	73.59 ± 5.40	76.27 ± 4.69	77.74 ± 5.58
		Group 2	83.04 ± 3.97	78.21 ± 4.69	78.43 ± 4.71	80.10 ± 5.53
	F1 score	Group 1	74.22 ± 5.26	71.38 ± 5.64	74.26 ± 4.94	75.74 ± 5.91
		Group 2	81.25 ± 4.23	76.39 ± 4.93	76.52 ± 4.97	78.32 ± 5.85
Dataset 2	Accuracy	Group 1	96.00 ± 1.08	95.54 ± 0.96	95.47 ± 1.24	94.85 ± 1.41
		Group 2	97.15 ± 0.91	96.40 ± 0.88	96.36 ± 0.76	95.49 ± 1.04
	Sensitivity	Group 1	80.06 ± 5.38	77.69 ± 4.80	77.34 ± 6.20	74.25 ± 7.04
		Group 2	85.74 ± 4.54	82.00 ± 4.38	81.80 ± 3.79	77.47 ± 5.19
	Specificity	Group 1	97.78 ± 0.60	97.52 ± 0.53	97.48 ± 0.69	97.14 ± 0.78
		Group 2	98.42 ± 0.50	98.00 ± 0.49	97.98 ± 0.42	97.50 ± 0.58
	Precision	Group 1	81.62 ± 5.41	79.11 ± 4.71	78.68 ± 6.11	75.36 ± 7.28
		Group 2	86.81 ± 4.19	83.36 ± 4.24	83.17 ± 3.55	79.15 ± 4.54
	F1 score	Group 1	79.87 ± 5.58	77.30 ± 5.12	77.00 ± 6.40	73.71 ± 7.33
		Group 2	85.54 ± 4.76	81.73 ± 4.60	81.61 ± 3.88	77.15 ± 5.37