| Model,Precision,Recall,F1-Score,Accuracy,Training Time | |
| BiLSTM+PhoBert,88.47%,88.46%,88.44%,88.46%,192.62 seconds + 8057.01 seconds (feature extraction) | |
| Longformer-PhoBert,91.63%,91.64%,91.63%,91.64%,275155.18 seconds | |
| PhoBert-base,90.50%,90.51%,90.50%,90.51%,130717.66 seconds | |