Table 2. AUC comparison of best performing features on all dataset using linear classifier.
Feature Name | AUC |
---|---|
MFCC-3 | 0.8919 |
LPC coef-1 | 0.8708 |
Tonality index | 0.8659 |
LPC coef-2 | 0.8557 |
Averaged PSD-17 | 0.8423 |
Spectral irregularity | 0.8354 |
WT coef STD-1 | 0.8327 |
WT coef energy-1 | 0.8323 |
Averaged PSD-16 | 0.8300 |
LPC coef-3 | 0.8211 |
Spectral brightness | 0.8153 |
95% freq roll-off | 0.8147 |