Skip to main content
Fig. 2 | Theoretical Biology and Medical Modelling

Fig. 2

From: Forecasting infectious disease emergence subject to seasonal forcing

Fig. 2

Performance of EWS over a moving window as measured by AUC for SIR systems approaching emergence in the presence of seasonal transmission. The heat plots show the relationship between AUC and seasonality without first removing seasonal trends (left), with one year differenced data (middle), and with seasonal decomposition (right). Data were not pre-processed prior to wavelet-based EWS. In each heat plot, seasonal intensity increases from left to right. The statistics are plotted with highest mean AUC to lowest from top to bottom. The bandwidth for all EWS is equal to 100 weeks (but the bandwidth is not pertinent for wavelet EWS). The AUC value indicates the area under the corresponding ROC curve. An AUC value of 1 indicates perfect detection ability while a value around 0.5 signifies that the statistic is no better than chance at distinguishing between data approaching emergence and data not approaching emergence

Back to article page