Advertisement

< Back to Article

Predicting Epidemic Risk from Past Temporal Contact Data

Fig 4

Validation of the risk assessment analysis.

(A), (B): Probability distributions of the risk ratio ν for the cattle trade network and the sexual contact network, respectively. Red lines are computed on training sets (2007–08 for cattle and s2-s3 for sexual contacts). The dashed lines peaking around 1 represent a null model based on reshuffling the infection statuses, i.e. randomly permuting the attribute “actually being infected” among the nodes for which risk assessment is performed. (C), (D): Probability distributions of the predictive power ω for the cattle trade network and the sexual contact network, respectively.

Fig 4

doi: https://doi.org/10.1371/journal.pcbi.1004152.g004