Table 9. F1 score of the best pairs of features for wheeze and normal sound classification using logistic regression on the Test-Set.
| First Feature | Second Feature | F1 Score |
|---|---|---|
| Tonality index | Spectral slope | 0.8319 |
| MFCC-3 | MFCC-1 | 0.8165 |
| MFCC-3 | Spectral slope | 0.8154 |
| Tonality index | MFCC-1 | 0.8152 |
| MFCC-3 | RMS | 0.8133 |
| MFCC-3 | Averaged PSD-2 | 0.8112 |
| MFCC-3 | Entropy mean | 0.8108 |
| Entropy difference | Spectral Irregularity | 0.8099 |
| MFCC-3 | Spectral Irregularity | 0.8089 |
| MFCC-10 | Spectral Irregularity | 0.8080 |
| MFCC-3 | Entropy ratio | 0.8078 |
| MFCC-3 | Spectral skewness | 0.8049 |
| Tonality index | - | 0.7863 |
| MFCC-3 | - | 0.7786 |