dfcop: Discrete factor copula models
In tvatter/mdmd: Multivariate Discrete Mixture Distributions

Description Usage Arguments Details Value Examples

Fit discrete factor copula models to data.

1 2	dfcop(data, family_set = "all", selcrit = "bic", ngrid = 100, keep_data = TRUE)

`data`	a binary matrix or data.frame.
`family_set`	a character vector of families; as in `dfcop_dist()`, see Details for additional options.
`selcrit`	criterion for family selection, either `"loglik"`, `"aic"`, or `"bic"`.
`ngrid`	number of nodes and weights for the Gaussian quadrature.
`keep_data`	whether the data should be stored (necessary for computing fit statistics and using `fitted()`).

In addition to the families in dfcop_dist(), the following convenience definition from can be used (and combined):

"all": all families

.

"archimedean": archimedean families.
"elliptical": elliptical families.
"bbs": BB families.
"oneparametric ": one parameter families.
"twoparametric ": two parameter families.

Partial matching is activated. For example, you can write "arch" instead of the full name "archimedean".

An object inherting from dfcop and dfcop_dist.

# Sample size, dimension, correlation parameter
n <- 1e3
d <- 1e1
rho <- 0.7

# The model
prob <- runif(d)
dfcop <- dfcop_dist(prob, "gaussian", rho)

# Data
X <- rdfcop(n, dfcop)

# Fit
fit <- dfcop(X)
fit