Table 2.
Test accuracy of classifiers when given different levels of noise.
Noise level | Normal subjects | Patient after stroke | ||||
---|---|---|---|---|---|---|
EMG | Feature fusion | Decision fusion | EMG | Feature fusion | Decision fusion | |
1e−5 | 85.91% | 93.54% | 96.61% | 81.07% | 91.11% | 95.61% |
2e−5 | 81.87% | 92.37% | 95.57% | 77.21% | 89.89% | 95.39% |
1e−4 | 58.88% | 88.13% | 90.13% | 53.50% | 86.04% | 91.07% |