Research Article

Myoelectric Pattern Recognition Performance Enhancement Using Nonlinear Features

Table 1

The EMG pattern recognition performances of different feature extraction methods.

ParameterClassifierFS1FS2FS3FS4Proposed

Dataset 1AccuracyLDA96.79 ± 1.0096.27 ± 0.8896.21 ± 0.9296.03 ± 1.0597.79 ± 0.52
SVM95.73 ± 0.9395.14 ± 1.0695.53 ± 0.9395.66 ± 1.0297.52 ± 0.56
KNN94.94 ± 1.0494.39 ± 1.1294.95 ± 0.9895.25 ± 1.1597.23 ± 0.63
SensitivityLDA83.96 ± 4.9881.36 ± 4.3881.03 ± 4.6080.15 ± 5.2588.97 ± 2.58
SVM78.65 ± 4.6775.70 ± 5.3077.63 ± 4.6478.30 ± 5.0887.62 ± 2.80
KNN74.71 ± 5.1971.94 ± 5.6074.74 ± 4.9076.24 ± 5.7786.16 ± 3.13
SpecificityLDA98.22 ± 0.5597.93 ± 0.4997.89 ± 0.5197.79 ± 0.5898.77 ± 0.29
SVM97.63 ± 0.5297.30 ± 0.5997.51 ± 0.5297.59 ± 0.5698.46 ± 0.35
KNN97.19 ± 0.5896.88 ± 0.6297.19 ± 0.5497.36 ± 0.6498.62 ± 0.31
PrecisionLDA85.24 ± 4.7282.41 ± 4.2882.55 ± 4.4981.41 ± 4.3289.95 ± 2.53
SVM79.42 ± 5.0077.11 ± 5.1378.83 ± 4.8179.61 ± 5.3588.47 ± 2.77
KNN75.96 ± 5.1973.59 ± 5.4076.27 ± 4.6977.74 ± 5.5887.03 ± 3.11
F1 scoreLDA83.49 ± 5.1380.56 ± 4.5780.58 ± 4.6979.41 ± 5.5988.55 ± 2.72
SVM78.08 ± 4.8475.06 ± 5.4176.97 ± 4.8377.71 ± 5.3287.26 ± 2.88
KNN74.22 ± 5.2671.38 ± 5.6474.26 ± 4.9475.74 ± 5.9185.77 ± 3.24

Dataset 2AccuracyLDA97.18 ± 0.9496.73 ± 0.9396.69 ± 1.0795.78 ± 1.0898.36 ± 0.45
SVM96.58 ± 0.9796.00 ± 0.8595.88 ± 1.0095.32 ± 1.2998.07 ± 0.59
KNN96.00 ± 1.0895.54 ± 0.9695.47 ± 1.2494.85 ± 1.4197.97 ± 0.66
SensitivityLDA85.88 ± 4.7283.66 ± 4.6683.43 ± 5.3478.88 ± 5.4091.78 ± 2.25
SVM82.88 ± 4.8680.01 ± 4.2479.42 ± 4.9976.60 ± 6.4490.34 ± 2.96
KNN80.06 ± 5.3877.69 ± 4.8077.34 ± 6.2074.25 ± 7.0489.83 ± 3.28
SpecificityLDA98.43 ± 0.5298.18 ± 0.5298.16 ± 0.5997.65 ± 0.6099.09 ± 0.25
SVM98.10 ± 0.5497.78 ± 0.4797.71 ± 0.5597.40 ± 0.7298.93 ± 0.33
KNN97.78 ± 0.6097.52 ± 0.5397.48 ± 0.6997.14 ± 0.7898.87 ± 0.36
PrecisionLDA87.02 ± 4.6985.26 ± 4.3184.79 ± 4.8680.05 ± 4.9092.65 ± 1.99
SVM83.92 ± 4.9281.51 ± 4.0580.75 ± 5.0178.12 ± 6.5891.20 ± 2.68
KNN81.62 ± 5.4179.11 ± 4.7178.68 ± 6.1175.36 ± 7.2890.70 ± 2.95
F1 scoreLDA85.55 ± 4.8483.22 ± 4.6883.14 ± 5.3478.37 ± 5.5091.59 ± 2.29
SVM82.63 ± 5.0179.78 ± 4.3479.00 ± 5.2176.17 ± 6.7090.19 ± 3.02
KNN79.87 ± 5.5877.30 ± 5.1277.00 ± 6.4073.71 ± 7.3389.70 ± 3.39