Table 1.
Candidate models for associations of remotely sensed variables and likelihood of a pixel being classified as elevated risk of exposure to Yersinia pestis in the West Nile region of Uganda*
| Model ID | Negative log-likelihood | K | AIC | ΔAIC | AUC | Sensitivity | Specificity | PPV | NPV | Independent model variables |
|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 56.91 | 5 | 66.91 | 3.21 | 80 | 81 | 69 | 57 | 88 | 1,300 meters, wetness, greenness, brightness |
| 2 | 55.73 | 6 | 67.73 | 4.03 | 80 | 86 | 74 | 62 | 91 | 1,300 meters, band 3, band 7, ST, brightness |
| 3 | 55.60 | 5 | 65.60 | 1.90 | 81 | 89 | 71 | 60 | 93 | 1,300 meters, band 3, band 6, brightness |
| 4 | 55.60 | 4 | 63.70 | 0 | 81 | 86 | 71 | 60 | 91 | 1,300 meters, band 3, band 6 |
*
K = number of estimated parameters included in the model; AIC = Akaike information criterion; AUC = area under the receiving operator characteristic curve; PPV = positive predictive value; NPV = negative predictive value; ST = surface temperature.