Skip to main content
. 2018 Jan;24(1):87–94. doi: 10.3201/eid2401.171393

Table 3. Cross-predictability of the BRT models trained with the different epidemic waves of influenza A(H7N9), China, applied to the others, as measured by the area under the curve*.

Predictions Applied to
Wave 1 Wave 2 Wave 3 Wave 4 Wave 5
Wave 1 0.91 0.81 0.78 0.84 0.79
Wave 2 NA 0.85 0.78 0.83 0.76
Wave 3 NA NA 0.82 0.82 0.74
Wave 4 NA NA NA 0.83 0.75
Wave 5 NA NA NA NA 0.76

*BRT, boosted regression tree; NA, not applicable.