Skip to main content
. 2019 Jul 25;13:250. doi: 10.3389/fnhum.2019.00250

Table 3.

Detection performance when non-candidates for classification output were included in the learning.

Sub. Four commands Three commands Two commands
Accuracy (%) ITR (bits/min) Accuracy (%) ITR (bits/min) Accuracy (%) ITR (bits/min)
1 60.0 0.53 66.7 0.44 75.0 0.25
2 70.0 0.86 74.4 0.68 83.3 0.47
3 60.0 0.53 67.8 0.47 78.3 0.33
4 70.0 0.86 75.6 0.72 83.3 0.47
5 78.3 1.20 84.4 1.07 91.7 0.78
6 61.7 0.58 66.7 0.44 75.0 0.25
7 100 2.67 100 2.11 100 1.30
8 71.7 0.92 76.7 0.76 85.0 0.52
9 66.7 0.74 74.4 0.68 85.0 0.52
10 100 2.67 100 2.11 100 1.30
Mean 73.8 1.16 78.7 0.95 85.7 0.63
SD 14.2 0.78 11.8 0.61 8.6 0.38

The average results of all target combinations were presented for <5 commands.