Sample from the posterior of an aggregate-downscale fit

Draws n samples from the per-component posterior mixture. Each draw picks a component by posterior_weights, then samples from N(posterior_components_means[[k]], posterior_components_covariances[[k]]).

Usage

posterior_sample_aggregate(object, n = 1000L, seed = NULL)

Arguments

object

An aggregate_downscale fit.

n

Integer. Number of posterior samples. Default 1000L.

seed

Integer or NULL. Random seed.

Value

An n x dim_x numeric matrix.

See also

Other downscaling and embeddings: aggregate_downscale(), dist_regression(), fit_cme(), kernel_downscale()

Examples

set.seed(1L)
A <- matrix(c(0.5, 0.5), nrow = 1L)
prior <- list(
  means = list(c(0, 0), c(2, 2)),
  covariances = list(diag(2L), diag(2L)),
  weights = c(0.5, 0.5)
)
fit <- aggregate_downscale(y = 1.0, aggregator = A,
                           latent_prior = prior, sigma_y = 0.2)
draws <- posterior_sample_aggregate(fit, n = 500L, seed = 1L)
colMeans(draws)
#> [1] 1.0163065 0.9701081