sample_prior: Sample from prior distribution

View source: R/sample_prior.R

sample_priorR Documentation

Sample from prior distribution

Description

Function to obtain samples from the prior distributions of the Bayesian Mallows model. Intended to be given to update_mallows().

Usage

sample_prior(n, n_items, priors = set_priors())

Arguments

n

An integer specifying the number of samples to take.

n_items

An integer specifying the number of items to be ranked.

priors

An object of class "BayesMallowsPriors" returned from set_priors().

Value

An object of class "BayesMallowsPriorSample", containing n independent samples of \alpha and \rho.

See Also

Other modeling: burnin(), burnin<-(), compute_mallows(), compute_mallows_mixtures(), compute_mallows_sequentially(), update_mallows()

Examples

# We can use a collection of particles from the prior distribution as
# initial values for the sequential Monte Carlo algorithm.
# Here we start by drawing 1000 particles from the priors, using default
# parameters.
prior_samples <- sample_prior(1000, ncol(sushi_rankings))
# Next, we provide the prior samples to update_mallws(), together
# with the first five rows of the sushi dataset
model1 <- update_mallows(
  model = prior_samples,
  new_data = setup_rank_data(sushi_rankings[1:5, ]))
plot(model1)

# We keep adding more data
model2 <- update_mallows(
  model = model1,
  new_data = setup_rank_data(sushi_rankings[6:10, ]))
plot(model2)

model3 <- update_mallows(
  model = model2,
  new_data = setup_rank_data(sushi_rankings[11:15, ]))
plot(model3)

BayesMallows documentation built on Sept. 11, 2024, 5:31 p.m.