Table 5 Classification performance per feature per classifier (original Vgg16 with data augmentation) after 100 training epochs.
From: An ensemble learning approach to digital corona virus preliminary screening from cough sounds
Feature | Original Vgg16 with data augmentation | |||||
|---|---|---|---|---|---|---|
ACC | Prec | Recall | F1 | kappa | AUC | |
SPEC | 0.63 | 0.61 | 0.70 | 0.65 | 0.25 | 0.63 |
Chroma | 0.56 | 0.58 | 0.44 | 0.50 | 0.13 | 0.56 |
MFCC | 0.61 | 0.60 | 0.63 | 0.62 | 0.21 | 0.60 |
MelSpectrum | 0.62 | 0.63 | 0.56 | 0.60 | 0.23 | 0.61 |
PowerSPEC | 0.60 | 0.63 | 0.46 | 0.53 | 0.20 | 0.60 |
RAW | 0.57 | 0.63 | 0.37 | 0.46 | 0.15 | 0.57 |
Tonal | 0.54 | 0.60 | 0.22 | 0.32 | 0.07 | 0.54 |
All features | 0.53 | 0.52 | 0.88 | 0.65 | 0.05 | 0.53 |
