Abstract
Brian Dennis and William Kemp (et al.) in ”Stochastic Model of Insect Phenology Estimation and Testing” use a Multinomial Maximum likelihood approach to estimate model parameters via ratios between variables. The model in this paper attempts to estimate deterministic epidemiological population parameters for the Kermack-McKendrick SIR model using simulated sample survey data. Ratios between populations in the SIR model are used as probabilities in the Multinomial distribution and resulting estimates of population paarameters are analyzed via Bootstrap confidence intervals, visual analysis, and mean squared error (MSE) estimates. The results show that certain surveys schemes perform better than others based on survey sample size, number of surveys, and survey spacing. All simulations generally produce unbiased estimates with some producing smaller variances than others. Applications include the estimating of infectious disease dynamics using small survey sampling in rural or small town universities and populations.