Skip to main content
. 2021 Jan 11;14:41. doi: 10.1186/s13071-020-04554-9

Table 3.

Percentage reduction in the out-of-sample mean absolute prediction error of the model compared with the baseline prediction

Leave-one-site-out Leave-one-planning-area-out
Ae. aegypti 3.6% (p = 0.477) 9.3% (p = 0.006)
Ae. albopictus 17.5% (p < 0.001) 14.1% (p < 0.001)

Under both types of spatial cross validation, the mean absolute prediction error was derived by averaging the absolute prediction errors across all sites. The p-values were obtained based on a two-sided paired t test