Table 11 Impact of entity and concept knowledge on model performance for Politifact datasets
From: KAGN:knowledge-powered attention and graph convolutional networks for social media rumor detection
Variants | Accuracy | Precision | Recall | F1 |
|---|---|---|---|---|
PTE | 0.8203 | 0.8203 | 0.8210 | 0.8202 |
PTE + KGE/E | 0.8047 | 0.8074 | 0.8068 | 0.8047 |
PTE + KGE/C | 0.8359 | 0.8367 | 0.8345 | 0.8351 |
PTE + KGE | 0.8438 | 0.8434 | 0.8434 | 0.8438 |
PTE + KAE/E | 0.8516 | 0.8515 | 0.8509 | 0.8511 |
PTE + KAE/C | 0.8593 | 0.8634 | 0.8620 | 0.8593 |
PTE + KAE | 0.8672 | 0.8683 | 0.8672 | 0.8673 |