pcondrvine: Conditional Distribution in a Regular Vine
In vincenzocoia/copsupp: Vine Copulas made easy

Description Usage Arguments Details Value Examples

Evaluates the conditional distribution of a variable in a regular vine out of all specified variables. As of now, will likely only work if the vine is regular, and truncated in the traditional sense.

1	pcondrvine(dat, rv, var, condset, maxint = 2, verbose = FALSE)

`dat`	vector or matrix of uniform observations (columns are variables).
`rv`	Regular vine object.
`var`	Integer; the variable you wish to condition on (i.e. the column number of `dat`, also present in `rv`).
`condset`	Vector of the conditioning variables (integers). Optional; leave blank to condition `var` on all other variables in `rv`.
`maxint`	Integer; maximum dimension of integration to tolerate. Put `Inf` if you don't want an upper limit.
`verbose`	Logical; should the function output how it goes about finding the conditional distribution?

To compute the conditional distribution, the vine is subsetted to the selected variables if possible. Then, if the conditioned variable is a leaf, the conditional distributon is directly computed. If it's not a leaf, the conditional distribution is computed by integrating the density.

If the subsetted vine does not exist, then the vine will be subsetted "as much as possible", and the remaining variables that cannot be removed will be integrated out to find the joint density of the selected variables, from which the conditional cdf will be found.

A vector of length = the number of observations in dat, representing the evaluated conditional distribution of variable var given the other variables in condset.

If integration beyond maxint dimensions is required to obtain the quantities, then an error is thrown.

## D-Vine example
G <- AtoG(CopulaModel::Dvinearray(5))[1:3, ]
rv <- rvine(G, "bvncop", makeuppertri(c(1:7/10), 2, 5, byRow = FALSE))
dat <- rrvine(10, rv)

## Compute 5|1:4. There's an algorithm for that.
pcondrvine(dat, rv, var=5, verbose=T)

## Compute 5|4. pcondrvine just uses 'pcondcop()'.
pcondrvine(dat, rv, var=5, condset=4, verbose=T)

## Compute 5|2:3. Two integrals takes ~13 min if maxint > 1.
pcondrvine(dat, rv, var=5, condset=c(2,3), maxint=1, verbose=T)

## Compute 4|(1,2,3,5). No algorithm for that.
pcondrvine(dat, rv, var=4, verbose=T)
pcondrvine(dat, rv, var=4, maxint=0, verbose=T) # No int. tolerance