Table 9 Impact of entity and concept knowledge on model performance for twitter16 datasets
From: KAGN:knowledge-powered attention and graph convolutional networks for social media rumor detection
Variants | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|
PTE | 0.8295 | 0.8359 | 0.8296 | 0.8315 |
PTE + KGE/E | 0.8352 | 0.8361 | 0.8347 | 0.8350 |
PTE + KGE/C | 0.8580 | 0.8628 | 0.8577 | 0.8594 |
PTE + KGE | 0.8672 | 0.8734 | 0.8676 | 0.8650 |
PTE + KAE/E | 0.8465 | 0.8478 | 0.8463 | 0.8460 |
PTE + KAE/C | 0.8594 | 0.8637 | 0.8582 | 0.8585 |
PTE + KAE | 0.8750 | 0.8749 | 0.8742 | 0.8744 |