Hamiltonian Analysis of Subcritical Stochastic Epidemic Dynamics

Figure 6

Changing trachoma prevalence at baseline and at 6-month intervals during the TANA trial of mass administration of azithromycin [23]. As the trial progresses, the prevalences become smaller and become more closely approximated by the exponential [5]. (Individual village prevalences are shown in tick marks on the horizontal axis. Curves result from beta distribution kernel density smoothing [24], with smoothing parameter determined from leave-one-out cross-validation [25].)