Fig. 3: The information loss from delays. | Nature Computational Science

Fig. 3: The information loss from delays.

From: Quantifying the information in noisy epidemic curves

Fig. 3

We compute \(\theta ({C}_{1}^{\tau })\) for delay distributions that are NB with various dispersions k. Smaller values of \(\theta ({C}_{1}^{\tau })\) indicate more substantial information losses, while smaller k indicates a more dispersed or variable reporting delay distribution. a, Influence of mean delay (\(\bar{\delta }\)) and k on \(\theta ({C}_{1}^{\tau })\), with colored curves representing different \(\bar{\delta }\) and the black dashed line as a reference value. Inset: variations in \(\theta ({C}_{1}^{\tau })\) at a given \(\bar{\delta }\) (matching colors) due to the dispersion k of the reporting delay distribution. b, Influence of the shape of the cumulative delay distributions, Fτ−t, at different k (increasing from blue to red) on our metric \(\theta ({C}_{1}^{\tau })\) (inset with corresponding colors).

Source data

Back to article page