Research Article
An Adaboost-Backpropagation Neural Network for Automated Image Sentiment Classification
Table 2
Comparison of the experimental results with the results obtained using the BP and Adaboost-BP methods.
| Emotion type | Recall rate of BP (%) | Recall rate of Adaboost-BP (%) | Precision rate of BP (%) | Precision rate of Adaboost-BP (%) |
| Sorrowful | 89.7% | 93.5% | 86.4% | 88.9% | Frightened | 88.3% | 90.8% | 82.6% | 86.4% | Disgusted | 82.9% | 87.2% | 73.1% | 79.8% | Relaxed | 91.6% | 93.2% | 88.5% | 90.2% | Angry | 90.1% | 93.1% | 84.9% | 89.5% | Frustrated | 79.3% | 86.2% | 71.6% | 77.8% | Scared | 90.9% | 93.6% | 85.7% | 90.1% | Happy | 93.6% | 95.8% | 89.1% | 92.2% | Proud | 83.0% | 87.1% | 78.3% | 84.3% | Hopeful | 91.4% | 93.8% | 85.7% | 89.6% |
|
|