Table 1 Relation between α and the fraction of the screening set contributing 80% of the BEDROC score
From: Estimation of the applicability domain of kernel-based machine learning models for virtual screening
| Data set percentage | α-value |
|---|---|
| 1.6% | 100.0 |
| 3.0% | 53.6 |
| 3.2% | 50.0 |
| 5.0% | 32.2 |
