Gaussian sampling family with known covariance

Observations are single draws \(X \sim N(\theta, \Sigma)\) with \(\Sigma\) known, so the parameter is the mean.

Usage

gaussian_family(d, sigma = NULL)

Arguments

d

Integer dimension of the observation.

sigma

Known covariance matrix, or NULL for the identity.

Value

A gaussian_family.

Details

The sample space real_region() cannot be enumerated: use mc_engine() or gh_engine(), not exact_engine(). No certified gap bound is available at the time of writing.

Examples

gaussian_family(d = 2L)
#> <gaussian_family>
#>   parameters real_region, dimension 2
#>   outcomes   real_region, dimension 2
gaussian_family(d = 2L, sigma = diag(c(1, 4)))
#> <gaussian_family>
#>   parameters real_region, dimension 2
#>   outcomes   real_region, dimension 2