https://github.com/ecoronado92/model_influeza_seasonal_dynamics
stochastic modeling | HMM
https://github.com/ecoronado92/model_influeza_seasonal_dynamics
baum-welch hmm influenza stochastic-processes viterb
Last synced: 3 months ago
JSON representation
stochastic modeling | HMM
- Host: GitHub
- URL: https://github.com/ecoronado92/model_influeza_seasonal_dynamics
- Owner: ecoronado92
- Created: 2019-07-01T14:44:56.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2019-12-19T01:58:23.000Z (over 5 years ago)
- Last Synced: 2025-03-14T00:26:54.861Z (3 months ago)
- Topics: baum-welch, hmm, influenza, stochastic-processes, viterb
- Language: HTML
- Homepage: https://ecoronado.github.io/
- Size: 2.39 MB
- Stars: 3
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Citation: citations.bib
Awesome Lists containing this project
README
# 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