de_mcmc: de_mcmc

Description

Bayesian inference for a deterministic DE model (with models solved via an DE solver) with an observation model.

Usage

de_mcmc(
  N,
  data,
  de.model,
  obs.model,
  all.params,
  ref.params = NULL,
  ref.inits = NULL,
  Tmax,
  data.times,
  cnt = 10,
  plot = TRUE,
  sizestep = 0.01,
  solver = "ode",
  verbose.mcmc = TRUE,
  verbose = FALSE,
  ...
)

Value

a debinfer_result object containing input parameters, data and MCMC samples

Arguments

N: integer, number of MCMC iterations
data: data.frame of time course observations to fit the model to. The observations must be ordered ascending by time.
de.model: a function defining a DE model, compliant with the solvers in deSolve or PBSddesolve
obs.model: a function defining an observation model. Must be a function with arguments 'data', 'sim.data', 'samp'.
all.params: debinfer_parlist containing all model, MCMC, and observation
ref.params: an optional named vector containing a set of reference parameters, e.g. the true parameters underlying a simulated data set
ref.inits: an optional named vector containing a set of reference initial values, e.g. the true initial values underlying a simulated data set
Tmax: maximum timestep for solver
data.times: time points for which observations are available
cnt: integer interval at which to print and possibly plot information on the current state of the MCMC chain
plot: logical, plot traces for all parameters at the interval defined by cnt
sizestep: timestep for solver to return values at, only used if data.times is missing
solver: the solver to use. 1 or "ode" = deSolve::ode; 2 or "dde" = PBSddesolve::dde; 3 or "dede" = deSolve::dde
verbose.mcmc: logical display MCMC progress messages
verbose: logical display verbose solver output
...: further arguments to the solver