Hidden Markov Models (HMM) and Autoregressive Processes (AR)

Summary

Early detection of the influenza outbreaks is one of the biggest challenges of outbreak surveillance systems. In this paper, a finite, homogeneous two-state Hidden Markov Model (HMM) was developed to determine the epidemic and non-epidemic dynamics influenza-like illnesses (ILI) in a differenced time series. These dynamics were further modeled via a first-order auto-regressive process (AR1) based on the state, and parameters estimated via the Baum-Welch algorithm. The model was evaluated with US ILI data from 1998-2018.

Files

• *.html final document
• *.Rmd generates final document
• BW_fcns.R is a script containing the Baum-Welch functions to train the model
• viterbi.R is a script containing the decoding algorithm to determin epidemic vs non-epidemic weeks based on training data
• flu_paper_model.R is a JAGS model from a similar paper run as a comparison