9 Hidden Markov Modelling (HMM)

9.1 Prep the data

  1. create a single clean data frame containing data for every file

  2. Run a 2state HMMM, have to run the following for the HMM:

    prepData()

    define variables: stateNames <- dry, wet stepPar <- (inital mean for states based on prop_wet) anglePar <- (no angle parameter for our analysis)

    fitHMM() plotSates() - visualize the model

  3. test with different binning intervals (2hr, 5hr?)

  4. test covariates

  • wetness level (0-20) as a covariate.
  • maybe seasonality or a day/night indicator?

REFERENCES: https://cran.r-project.org/web/packages/momentuHMM/vignettes/momentuHMM.pdf

https://mjones029.github.io/Tutorials/HMM_tutorial.html