Table 2 Comparison of different positive and negative sample ratios during MFERL training
From: Leveraging explainable multi-scale features for fine-grained circRNA-miRNA interaction prediction
Ratios | AUC | AUPR | F1_score | Accuracy | Recall | Precision |
|---|---|---|---|---|---|---|
1:1 | 0.9682 | 0.9667 | 0.9191 | 0.9182 | 0.9265 | 0.9119 |
1:2 | 0.9710 | 0.9445 | 0.8913 | 0.9265 | 0.9040 | 0.8790 |
1:5 | 0.9759 | 0.9117 | 0.8539 | 0.9512 | 0.8543 | 0.8536 |
1:10 | 0.9756 | 0.8771 | 0.8148 | 0.9666 | 0.8078 | 0.8219 |