Figure 3
From: A Dirichlet process model for classifying and forecasting epidemic curves
A sample fit of a randomly selected epidemic curve to three parametric models. The x-axis represents the days and the y-axis is the normalized number of infected persons relative to the cumulative infected. The black curve is the true epidemic curve and the red is the fitted curve. Using the epidemic data, the maximum likelihood procedure is used in estimating the parameters for each of the models. The normalized epidemic data is plotted against samples from the distributions based on the estimated parameters.