Simulations of a partially-observed Markov process
simulate generates simulations of the state and measurement processes.
methods
## S4 method for signature 'missing' simulate( object, nsim = 1, seed = NULL, ..., times, t0, params, rinit, rprocess, rmeasure, format = c("pomps", "arrays", "data.frame"), include.data = FALSE, verbose = getOption("verbose", FALSE) ) ## S4 method for signature 'data.frame' simulate( object, nsim = 1, seed = NULL, ..., times, t0, params, rinit, rprocess, rmeasure, format = c("pomps", "arrays", "data.frame"), include.data = FALSE, verbose = getOption("verbose", FALSE) ) ## S4 method for signature 'pomp' simulate( object, nsim = 1, seed = NULL, ..., format = c("pomps", "arrays", "data.frame"), include.data = FALSE, verbose = getOption("verbose", FALSE) ) ## S4 method for signature 'objfun' simulate(object, nsim = 1, seed = NULL, ...)
object: optional; if present, it should be a data frame or a pomp object.
nsim: The number of simulations to perform. Note that the number of replicates will be nsim times ncol(params).
seed: optional integer; if set, the pseudorandom number generator (RNG) will be initialized with seed. The RNG will be restored to its original state afterward.
...: additional arguments are passed to pomp.
times: the sequence of observation times. times must indicate the column of observation times by name or index. The time vector must be numeric and non-decreasing.
t0: The zero-time, i.e., the time of the initial state. This must be no later than the time of the first observation, i.e., t0 <= times[1].
params: a named numeric vector or a matrix with rownames containing the parameters at which the simulations are to be performed.
rinit: simulator of the initial-state distribution. This can be furnished either as a C snippet, an function, or the name of a pre-compiled native routine available in a dynamically loaded library. Setting rinit=NULL sets the initial-state simulator to its default. For more information, see rinit specification .
rprocess: simulator of the latent state process, specified using one of the rprocess plugins . Setting rprocess=NULL removes the latent-state simulator. For more information, see rprocess specification for the documentation on these plugins .
rmeasure: simulator of the measurement model, specified either as a C snippet, an function, or the name of a pre-compiled native routine available in a dynamically loaded library. Setting rmeasure=NULL removes the measurement model simulator. For more information, see rmeasure specification .
format: the format in which to return the results.
format = "pomps" causes the results to be returned as a single pomp object, if params is a vector, or a list of pomp objects, if params is a matrix with more than one column. Each of these will be identical to object except in that the latent states and observations will have been replaced by their simulated values.
format = "arrays" causes the results to be returned as a list of two arrays. The states element will contain the simulated state trajectories in a rank-3 array with dimensions nvar x (ncol(params)*nsim) x ntimes. Here, nvar is the number of state variables and ntimes the length of the argument times. The obs element will contain the simulated data, returned as a rank-3 array with dimensions nobs x (ncol(params)*nsim) x ntimes. Here, nobs is the number of observables.
format = "data.frame" causes the results to be returned as a single data frame containing the time, states, and observations. An ordered factor variable, .id , distinguishes one simulation from another.
include.data: if TRUE, the original data and covariates (if any) are included (with .id = "data"). This option is ignored unless format = "data.frame".
verbose: logical; if TRUE, diagnostic messages will be printed to the console.
A single pomp object, a pompList object, a named list of two arrays, or a data frame, according to the format option.
If params is a matrix, each column is treated as a distinct parameter set. In this case, if nsim=1, then simulate will return one simulation for each parameter set. If nsim>1, then simulate will yield nsim simulations for each parameter set. These will be ordered such that the first ncol(params) simulations represent one simulation from each of the distinct parameter sets, the second ncol(params) simulations represent a second simulation from each, and so on.
Adding column names to params can be helpful.
Some Windows users report problems when using C snippets in parallel computations. These appear to arise when the temporary files created during the C snippet compilation process are not handled properly by the operating system. To circumvent this problem, use the cdir and cfile options to cause the C snippets to be written to a file of your choice, thus avoiding the use of temporary files altogether.
More on pomp elementary algorithms: elementary_algorithms, kalman, pfilter(), pomp-package, probe(), spect(), trajectory(), wpfilter()
Aaron A. King