TABLE 3.
The average accuracy and F1-measure of different algorithms in valence and arousal classification.
Method | Valence |
Arousal |
||
ACC | F1 | ACC | F1 | |
1NN | 0.484 | 0.529 | 0.504 | 0.555 |
TCA | 0.49 | 0.533 | 0.521 | 0.583 |
JDA | 0.496 | 0.535 | 0.515 | 0.578 |
BDA | 0.519 | 0.56 | 0.538 | 0.572 |
JPDA | 0.541 | 0.576 | 0.568 | 0.626 |
JPDA-BPM | 0.533 | 0.615 | 0.573 | 0.613 |
SSJPDA | 0.617 | 0.627 | 0.635 | 0.643 |
SSJPDA-BPM | 0.636 | 0.653 | 0.644 | 0.679 |
The numbers in bold indicate the highest value of the experimental results.