cmc_downstream: Add structure to a null model

In vincenzocoia/cmc: Copula-Marginal Composition Method

Description

Converts marginal Unif(0,1) distributions to have modelled distributions. CURRENTLY ONLY HANDLES PIT SCORES AS VALUES.

Usage

## S3 method for class 'cmc'
construct(object, newdata = NULL, from_col, to_col)

construct(object, newdata = NULL, from_col, to_col)

## S3 method for class 'cmc'
decompose(object, newdata = NULL, from_col, to_col)

decompose(object, newdata = NULL, from_col, to_col)

## S3 method for class 'cmc'
predict(object, ..., newdata = NULL, what = "pdist", at,
  to_col)

Arguments

object	A cmc-fitted model.
newdata	Data frame to operate on
from_col	Name of column containing the null values. For example, probabilities (PIT scores) to evaluate the quantile functions at.
to_col	Name of the column to append the output to. Leave blank if you want a vector output.
...	Not used
what	What to predict. Could be "pdist", "qdist", or "rdist".
at	Vector of values to evaluate the what function at.

Details

If from_col values are numbers, these are interpreted as PIT scores by applying the quantile function corresponding to each predictive distribution. If values are distributions, these are converted in such a way that the predictive distributions are obtained if the values are all Unif(0,1) distributions.

vincenzocoia/cmc documentation built on Nov. 18, 2019, 12:04 a.m.