Research Article
Pattern Recognition Methods and Features Selection for Speech Emotion Recognition System
Table 6
Gaussian mixture model classification accuracy for different combinations of emotions [%].
| Train 2/test | Train 1 | Anger | Boredom | Disgust | Fear | Happiness | Sadness | Neutral state |
| Anger | — | 91.7 | 85.7 | 83.4 | 70 | 96.4 | 90.8 | Boredom | 76.3 | — | 64.9 | 66.2 | 71.9 | 65.3 | 59 | Disgust | 59.1 | 64.3 | — | 62.5 | 56 | 78.5 | 64.7 | Fear | 52.4 | 72.6 | 59.7 | — | 47.2 | 81.8 | 70.4 | Happiness | 42.7 | 83.9 | 73.7 | 73.9 | — | 90.5 | 82.4 | Sadness | 87.1 | 67 | 75.5 | 73 | 85.5 | — | 75 | Neutral state | 78.6 | 53.3 | 69.1 | 63.9 | 74 | 65.7 | — |
|
|