Table 2. Cross-validation performance over eight folds and 160 rounds of boosting .
| Square annotation | All | Positive | Negative |
|---|---|---|---|
| Mean test-set size | 2684 | 1353 | 1331 |
| Mean train-set size | 18793 | 9470 | 9323 |
| Test-set error (%) | 6.6 | 6.2 | 7.0 |
| Mean ROC-AUC score | N/A | 0.856 | 0.905 |