Research Article
BERT-m7G: A Transformer Architecture Based on BERT and Stacking Ensemble to Identify RNA N7-Methylguanosine Sites from Sequence Information
Table 1
The prediction results of different feature selection methods.
| Methods | SN (%) | SP (%) | ACC (%) | MCC | Optimal |
| EN | 95.68 | 95.01 | 95.34 | 0.9074 | 473 | XGBoost | 91.50 | 92.18 | 91.84 | 0.8375 | 753 | LightGBM | 90.15 | 90.42 | 90.28 | 0.8069 | 2206 | Boruta | 89.20 | 90.15 | 89.68 | 0.7953 | 1090 | SVD | 87.72 | 87.85 | 87.78 | 0.7570 | 473 | PCA | 87.45 | 87.59 | 87.52 | 0.7522 | 473 | SE | 86.24 | 86.78 | 86.51 | 0.7312 | 473 | LLE | 83.13 | 82.73 | 82.93 | 0.6603 | 473 | All | 88.66 | 87.73 | 88.19 | 0.7651 | 31488 |
|
|