Table 2 The comparison of testing results with and without CAWR.
From: Emotion recognition with multiple physiological parameters based on ensemble learning
Dataset | Random seed | Accuracy (%) | ||
|---|---|---|---|---|
RI | RI + CAWR | Improvement | ||
SELF | 3 | 65.30 | 88.75 | 23.45 |
6 | 60.16 | 88.10 | 27.94 | |
9 | 69.51 | 89.53 | 20.02 | |
12 | 66.01 | 87.81 | 21.80 | |
15 | 48.25 | 86.99 | 38.74 | |
SWELL | 3 | 92.75 | 97.43 | 4.68 |
6 | 94.06 | 96.03 | 1.97 | |
9 | 95.53 | 97.82 | 2.29 | |
12 | 88.37 | 96.50 | 8.13 | |
15 | 57.61 | 97.35 | 39.74 | |
WESAD | 3 | 95.10 | 98.21 | 3.11 |
6 | 93.85 | 98.03 | 4.18 | |
9 | 96.06 | 98.21 | 2.15 | |
12 | 93.43 | 98.15 | 4.72 | |
15 | 94.92 | 97.79 | 2.87 | |
