gcop: The Gaussian Copula distribution.
In mlysy/GaussCop: Utilities for the Gaussian Copula Distribution

Description Usage Arguments Details Value See Also Examples

The Gaussian Copula distribution.

1
2
3

dgcop(X, gCop, log = FALSE, decomp = FALSE)

rgcop(n, gCop)

`X`	`n x p` values at which to evaluate the `p`-dimensional density.
`gCop`	An object of class `gaussCop` specifying the Gaussian Copula model.
`log`	Logical; whether or not to evaluate the density on the log scale.
`decomp`	Logical; if `TRUE` returns the normalized residuals `Z`, their log-density `zlpdf`, and the log-jacobian `zljac`, such that the total log-density is `xldens = zldens + zljac`. See Details.
`n`	Number of random samples to draw.

The density of Gaussian Copula distribution is

g(x) = ψ(z | R) ∏_{i=1}^d f_i(x_i)/φ(z_i),

z_i = Φ^(-1)(F_i(x_i)),

where ψ(z | R) is the PDF of a multivariate normal with mean 0 and variance R, f_i(x_i) and F_i(x_i) are the marginal PDF and CDF of variable i, and φ(z) and Φ(z) are the PDF and CDF of a standard normal.

dgcop provides the density of gCop, rgcop generates random values from gCop.

gcopFit for constructing gaussCop objects and fitting the Gaussian Copula model to observed data.

# simulate data and plot it
n = 5e4
dat = cbind(rnorm(n, mean = 1, sd = 3),
            rnorm(n, mean=4, sd = 0.5))
plot(dat, cex=0.5)
# fit Gaussian Copula
temp.cop = gcopFit(X = dat, fitXD = "kernel")
# simulate data from Copula model and add it to plot, should blend in
new.data = rgcop(100, temp.cop)
points(new.data, cex = 0.5, col="red")