PIT.MultiQR: Probability integral transform: S3 Method for for MultiQR

In jbrowell/ProbCast: Flexible Probabilistic Forecasting

Probability integral transform: S3 Method for for MultiQR

Description

Transforms data (random variables) through their corresponding cumulative distribution function (specified by a MultiQR object and tail parameters) in order to produce variables with a standard uniform distribution.

Usage

## S3 method for class 'MultiQR'
PIT(qrdata, obs, tails, inverse = FALSE, ...)

Arguments

qrdata	A MultiQR object.
obs	A vector of observations corresponding to the rows of qrdata.
tails	A list of arguments passed to condCDF defining the tails of the CDFs. See ?contCDF for details. Tail parameters my have length 1 or length nrow(qrdata) if each row of qrdata has a corresponding parameter value.
inverse	A boolean. If true, the inverse transformation is appiled, i.e. uniform variable to random variable from original distribution.
...	Additional arguments passed to condCDF.

Details

Boundary imposed throwing a warning to ensure that output is in [0,1].

Value

The probability integral transform of obs through the predictive distribution defined by qrdata and interpolation scheme in contCDF with tails.

Author(s)

Jethro Browell, [e-mail]

jbrowell/ProbCast documentation built on July 20, 2024, 1:53 p.m.