Research Article
Early Rumor Detection Based on Deep Recurrent Q-Learning
Table 5
Experimental results of RDM.
| Model name | Accuracy | Precision | Recall | F1 score |
| CRF | 0.67 | 0.67 | 0.56 | 0.60 | GAN-GRU | 0.78 | 0.53 | 0.35 | 0.42 | RDM | 0.77 | 0.74 | 0.74 | 0.74 | LSTM | 0.81 | 0.77 | 0.69 | 0.72 | LSTM-Attention | 0.83 | 0.83 | 0.79 | 0.81 | SA-SE | 0.78 | 0.70 | 0.73 | 0.71 | SA-DE | 0.81 | 0.79 | 0.80 | 0.80 | DSA-DE | 0.84 | 0.81 | 0.82 | 0.82 |
|
|