Table 3 Performance comparison between transfer learning and several data augmentation methods for plant secondary metabolic pathway prediction
From: Prediction of plant secondary metabolic pathways using deep transfer learning
Methods | Compounds | Accuracy (%) | Precision (%) | Recall (%) | F1_score (%) |
|---|---|---|---|---|---|
Transfer Learning | 2028 | 97.82 | 83.74 | 83.74 | 83.74 |
Data augmentation α | 7666 | 97.80 | 83.10 | 84.17 | 83.63 |
Data augmentation β | 52,530 | 97.70 | 81.84 | 84.32 | 83.06 |
Data augmentation γ | 9846 | 97.50 | 81.02 | 81.73 | 81.37 |
Data augmentation θ | 17,927 | 97.54 | 81.13 | 82.30 | 81.71 |