biips_pimh_init: Create a PIMH object.
In biips/rbiips: Bayesian Inference with Interacting Particle Systems

Description Usage Arguments Value See Also Examples

The function biips_pimh_init initializes the Particle Independent Metropolis-Hastings (PIMH) algorithm.

The PIMH algorithm provides MCMC samples of the variables in variable_names, using a SMC algorithm as proposal distribution in an independent Metropolis-Hastings (MH) algorithm.

1	biips_pimh_init(model, variable_names)

`model`	`biips` model object as returned by `biips_model`.
`variable_names`	character vector. The names of the unobserved variables to monitor. Names can contain subset indices which must define a valid subset of the variables of the model, e.g.: `c('var1',` `'var2[1]',` `'var3[1:10]',` `'var4[1, 5:10, 3]')`.

The function biips_pimh_init returns an object of class pimh which can be used to generate samples from the posterior distribution of the monitored variables in variable_names.

An object of class pimh is a list of functions that share a common environment. These functions are meant for internal purpose only. They are used to query information on the current state of the algorithm.

`model()`	Get the `biips` model object.
`variable_names()`	Get a character vector with the names of the monitored variables.
`sample(sample)`	Get and set the current sample.
`log_marg_like(log_marg_like)`	Get and set the current value of the log marginal likelihood.

biips_model, biips_pimh_update, biips_pimh_samples

modelfile <- system.file('extdata', 'hmm.bug', package = 'rbiips')
stopifnot(nchar(modelfile) > 0)
cat(readLines(modelfile), sep = '\n')

data <- list(tmax = 10, p = c(.5, .5), logtau_true = log(1), logtau = log(1))
model <- biips_model(modelfile, data)

n_part <- 50
obj_pimh <- biips_pimh_init(model, c('x', 'c[2:10]'))  # Initialize
is.pimh(obj_pimh)
out_pimh_burn <- biips_pimh_update(obj_pimh, 100, n_part)  # Burn-in
out_pimh <- biips_pimh_samples(obj_pimh, 100, n_part)  # Samples