Groupe d’études et de recherche en analyse des décisions

Infectious diseases: Data analysis via continuous time population models

Edward L. Ionides

Analysis of population data on infectious diseases has the goals of increasing scientific understanding of the disease dynamics and making forecasts and policy decisions. Stochastic models used for data analysis purposes are typically set in discrete time. Continuous time models provide an attractive framework for formulating models and investigating theoretical properties. The gap between data analysis (discrete time) and theory (continuous time) arises because of challenges involved in carrying out inference directly for relevant continuous time processes. This work investigates the gap, and proposes new methods for inference from partially observed continuous time stochastic processes. Web: www.stat.lsa.umich.edu/~ionides