Fit the cholera model to incidence data using pMCMC (monty + dust2) — chlaa_fit_pmcmc • chlaa

Expects a data.frame with columns: - time: numeric times (days) - cases: integer case counts at those times

Usage

chlaa_fit_pmcmc(
  data,
  pars,
  n_particles = 200,
  n_steps = 2000,
  seed = 1,
  prior = NULL,
  packer = NULL,
  proposal_var = 0.02
)

Arguments

data: Data frame with columns time and cases.
pars: Starting parameter list.
n_particles: Number of particles for the dust2 filter likelihood.
n_steps: Number of MCMC steps.
seed: Random seed.
prior: Optional monty prior model. If NULL, uses `chlaa_default_prior()`.
packer: Optional monty packer. If NULL, uses `chlaa_default_packer()`.
proposal_var: Proposal variance (scalar) for diagonal random walk.

Value

A `chlaa_fit` object (also keeps monty class) with attributes: packer, prior, start_pars, data.

Details

Likelihood is defined inside `inst/odin/cholera_model_fit.R`: cases ~ NegativeBinomial(mu = reporting_rate * inc_symptoms, size = obs_size)