A Bayesian Hierarchical Model for Estimating Influenza Epidemic Severity

We present a model for forecasting influenza severity which uses historic and current data from both ILINet and Google Flu Trends. The model takes advantage of the accuracy of ILINet data and the real-time updating of Google Flu Trends data, while also accounting for potential bias in Google Flu Trends data. Using both data sources allows the model to more accurately forecast important characteristics of influenza outbreaks than using ILINet data alone.