Construct a bivariate copula

Build a copula from a family name and its dependence parameter(s). Mirrors the C# BivariateCopula hierarchy of the Numerics library (ClaytonCopula, GumbelCopula, ...).

Usage

copula(family, theta, df = NULL, margin_x = NULL, margin_y = NULL)

Arguments

family

one of copula_names().

theta

the dependence parameter.

df

degrees of freedom, required for "StudentT" and ignored otherwise.

margin_x, margin_y

optional corehydro_dist marginals (see Details).

Value

a corehydro_copula.

Details

margin_x and margin_y are optional and are treated differently depending on their type, which matters for copula_random() (which needs marginals to draw on the data scale) and for the IFM/full log-likelihoods:

• a corehydro_dist (e.g. distribution("Normal", c(0, 1))) attaches that distribution exactly as given, with no re-fitting.

• this constructor does not accept a bare family-name string for the marginals; use copula_fit() to have a named marginal MLE-fitted to a sample.

See also

copula_fit() to estimate a copula (and optionally its marginals) from data.

Examples

copula_pdf(copula("Clayton", theta = 2), 0.3, 0.7)
#> [1] 0.6292895