Table 3.
Classification accuracies for top vs. bottom.
| Window position [%] | |||||
|---|---|---|---|---|---|
| F | M | B | FB | MB | |
| S1 | 70.99 (54.00) | 72.84 (46.34) | 74.07 (50.33) | 80.86 (54.11) | 88.27 (52.55) |
| S2 | 86.21 (52.03) | 83.91 (48.89) | 79.31 (46.24) | 93.10 (49.73) | 89.66 (49.12) |
| S3 | 75.33 (50.72) | 74.00 (52.42) | 71.33 (51.31) | 83.33 (53.19) | 86.00 (52.36) |
| S4 | 74.67 (52.03) | 74.67 (50.02) | 72.00 (50.24) | 86.67 (50.63) | 82.67 (52.14) |
| S5 | 77.78 (47.31) | 76.92 (49.44) | 80.34 (51.89) | 93.16 (49.82) | 90.60 (49.65) |
| S6 | 75.69 (54.41) | 73.61 (46.44) | 70.83 (52.62) | 81.25 (54.42) | 75.69 (49.55) |
| Mean | 76.78 ± 5.12 | 75.99 ± 4.12 | 74.65 ± 4.17 | 86.40 ± 5.61 | 85.48 ± 5.58 |
Values outside parenthesis are means of the three highest classification accuracies obtained among all independent component pairs. Values in parenthesis are means of accuracies when five inputs in shuffled conditions were fed to model which made values outside parenthesis. S indicates subject.