TABLE 18. The Performance of Cross-Validation for Deep Neural Networks (The PolitiFact Dataset).
| DL Models | Measure Performance Methods | |||
|---|---|---|---|---|
| Accuracy | Precision | Recall | F1-score | |
| Modified LSTM one layer | 90.49± 2.35 | 90.55± 4.6 | 83.52± 6.75 | 86.4± 3.73 |
| Modified LSTM two layers | 94.21± 1.49 | 96.15± 2.84 | 88.0± 3.57 | 91.76± 2.15 |
| Modified LSTM three layers | 90.64± 1.38 | 90.96± 3.95 | 83.68± 4.32 | 86.85± 1.98 |
| Modified GRU one layer | 88.8± 2.54 | 87.85± 7.37 | 78.17± 7.76 | 81.85± 5.93 |
| Modified GRU two layers | 87.83± 2.87 | 87.45± 7.38 | 75.17± 7.06 | 80.26± 6.33 |
| Modified GRU three layers | 86.05± 1.83 | 82.97± 6.84 | 75.66± 6.06 | 78.32± 4.81 |