Skip to main content

Table 5 Prediction model results

From: A novel approach to modeling epidemic vulnerability, applied to Aedes aegypti-vectored diseases in Perú

 

Season

El Niño

AUC (95% CI)

Stage 1

S

0.901 (0.874, 0.929)

W

0.885 (0.853, 0.918)

S

Y

0.877 (0.842, 0.913)

S

N

0.859 (0.795, 0.922)

W

Y

0.914 (0.883, 0.944)

W

N

0.927 (0.904, 0.950)

Stage 2

S

0.867 (0.828, 0.905)

W

0.824 (0.767, 0.882)

S

Y

0.758 (0.689, 0.827)

S

N

0.905 (0.862, 0.949)

W

Y

0.889 (0.851, 0.927)

W

N

0.866 (0.822, 0.910)

Stage 3

S

0.673 (0.598, 0.747)

W

0.608 (0.526, 0.690)

S

Y

0.580 (0.490, 0.670)

S

N

0.679 (0.530, 0.828)

W

Y

0.658 (0.564, 0.753)

W

N

0.614 (0.521, 0.707)

  1. Prediction results from logistic regression models trained to two-thirds of the outcome data, and tested on remaining one-third. Outcome parameterized as one or more outbreaks in a given district over January 2016–September 2018. Stage 1: index case potential; stage 2: outbreak receptivity; stage 3: epidemic potential.AUC area under ROC curve,CI posterior credible interval, S summer; W winter, Y yes, N no