Table 3.
Comparison of different models.
| Modeling | ACC | Pre | Rec | F1 |
|---|---|---|---|---|
| Random forest | 85.9% | 91.75% | 75.50% | 84.36% |
| LightGBM | 87.82% | 91.83% | 84.87% | 87.76% |
| Bi-lstm | 89.71% | 92.50% | 90.36% | 89.36% |
| ResNet | 90.30% | 91.90% | 87.60% | 89.70% |
| XGboost | 93.51% | 95.02% | 92.02% | 93.53% |
| TPE_OptGBM | 95.08% | 94.80% | 97.51% | 96.14% |
The bold values represents the method we have proposed, and it is also the method that works best.