Carolina Ballroom
Statistical Modeling of Infectious Diseases: The Flu and the “Next” Disease (303437)
*Shannon K Gallagher, Carnegie Mellon UniversityModeling the spread of infectious diseases is an important problem in modern public health. We discuss a novel method for predicting influenza incidence in the United States, we generate flexible agents for use in agent based modeling for numerous diseases, and we create a tool for exploring classical diseases in the United States. First, we produce a new model for predicting influenza incidence in the United States. The new model is an extension on an Empirical Bayes model proposed by Brooks et al. Our model focuses on predicting the primary target of the peak incidence of influenza in 10 different regions in the United States. Our model performs twice as well as those generated by Empirical Bayes in predicting the peak week of influenza along with producing probabilistic forecasts. Next, we discuss agent based modeling the spread of recent infectious diseases such as Zika, Ebola, Chikungunya, and the “next” disease. We discuss our R package, spew, which generates a synthetic population of much of the world and how it can be used in agent based models. Finally, we build an interactive exploratory data analysis tool for visualizing disease incidence rates in the US.