Rosenblatt transformation

The Rosenblatt transformation is a method to transform uniform pseudo-observations into the original scale of the data, preserving the copula structure. Adapted from rvinecopulib::rosenblatt().

Usage

rosenblatt(x, model, cores = 1, randomize_discrete = TRUE)

Arguments

x

A numeric matrix or data.frame of observations. If not on \([0,1]^d\), they will be converted via rvinecopulib::pseudo_obs() unless already provided as pseudo-observations.

model

A vinecop object (from rvinecopulib) describing the dependence structure.

cores

Integer; number of threads to use. Default: 1.

randomize_discrete

Logical; if TRUE, apply jittering/randomized PIT for discrete/mixture margins (Rosenblatt for non-continuous margins). Default: FALSE.

Value

A numeric matrix of the same dimension as x, containing PIT-transformed values in \([0,1]\).