Compute variance-covariance from Fisher information

For a model_spec (not necessarily fitted), compute the asymptotic variance-covariance matrix from the inverse Fisher information.

Usage

model_vcov(spec, y, par, type = "observed")

Arguments

spec

A model_spec object

y

Numeric vector of data

par

Parameter values at which to evaluate

type

Type of Fisher information: "observed" or "expected"

Value

Variance-covariance matrix

Examples

spec <- model_spec_normal()
y <- rnorm(100, mean = 5, sd = 2)
model_vcov(spec, y, c(mu = 5, sigma = 2))