predict.WAR: Predict distributions given a WAR object

In yqgchen/WR: Wasserstein regression for distributional data

Predict distributions given a WAR object

Description

Predict distributions given a WAR object

Usage

## S3 method for class 'WAR'
predict(object, n.ahead = 1, Ynew = NULL, ...)

Arguments

object	A WAR object, i.e., an output of WAR.
n.ahead	Number of steps ahead at which to predict. Default: 1.
Ynew	A q-by-m matrix holding the quantile functions of m distributions evaluated on object$qSup to which to apply the prediction, where m must be greater than or equal to object$order. Default: object$Y.
...	Not used.

Value

A list of the following elements:

Ypred	A q-by-l matrix holding the predicted quantile functions of l distributions ahead of Ynew, where l is given by n.ahead.
qSup	The input qSup
outOfLogSpace	A logical vector holding whether initial predicted log mapped distributions lie out of the log image space (TRUE) or not (FALSE).

yqgchen/WR documentation built on June 10, 2025, 6:04 p.m.