fitbicop: Fit a Bivariate Copula Model

In vincenzocoia/copsupp: Vine Copulas made easy

Description

These functions choose and fit the best copula model for data with uniform margins. Intended for internal use. fitbicop_qcond not done yet.

Usage

fitbicop_lh(
  u,
  v,
  families = c("indepcop", "bvncop", "bvtcop", "mtcj", "gum", "frk", "joe", "bb1",
    "bb7", "bb8"),
  cpar = NULL,
  method = "aic"
)

Arguments

u	Vector of Unif(0,1) data.
v	Vector of Unif(0,1) data, same length as u.
families	Vector of copula families to try fitting. Rotations of each family are automatically considered.
method	Either "aic" or "bic" (case doesn't matter)

Value

A list with entries $cop (the best-fitting copula family), $cpar (the fitted copula parameters), and $aic and $bic.

Note

This function is not that smart – it assumes not all cpars are specified, and if some are specified, that the copula model is known.

I'm using BiCopSelect() just because it has smart starting points.

Examples

library(CopulaModel)
set.seed(13)
u <- runif(100)
v <- qcondbvtcop(runif(100), u, c(0.9, 1))
fitbicop_lh(u, v)
fitbicop_lh(u, v, families = "bvtcop", cpar = c(NA, NA))

v <- qcondgum(runif(100), u, 4)
fitbicop_lh(u, 1-v)

v <- runif(100)
fitbicop_lh(u, v)
fitbicop_lh(u, v, families = "indepcop")

vincenzocoia/copsupp documentation built on Aug. 23, 2020, 7:37 a.m.