Research Article
Pattern Recognition Methods and Features Selection for Speech Emotion Recognition System
Table 5
Performance results of classifiers for MFCC (dynamic and acceleration coefficients too) features.
| MFCC | ANN | k-NN | GMM |
| Neutral state | 7445 22.1% | 1901 5.7% | 79.7% | 4919 14.6% | 2209 6.6% | 69.0% | 6305 18.8% | 3507 10.4% | 63.3% |
| Stress | 1684 5.0% | 22587 67.2% | 93.1% | 4210 12.5% | 22279 66.3% | 84.1% | 2824 8.4% | 20981 62.4% | 88.1% |
| | 81.6% | 92.2% | 89.3% | 53.9% | 91.0% | 80.9% | 69.1% | 85.7% | 81.2% | | Neutral state | Stress | | Neutral state | Stress | | Neutral state | Stress | |
|
|