predict.qreg_gam: Predict from model based on Generalised Additive Model and...
In jbrowell/ProbCast: Flexible Probabilistic Forecasting

predict.qreg_gam

R Documentation

Predict from model based on Generalised Additive Model and Linear Quantile Regression

Description

This function predicts from multiple conditional linear quantile regression models of the residuals of a generalised additive model fit using qreg_gam.

Usage

## S3 method for class 'qreg_gam'
predict(
  object,
  newdata = NULL,
  quantiles = NULL,
  model_name = NULL,
  sort = T,
  sort_limits = NULL
)

Arguments

`object`	An `qreg_gam` object containing the model to predict from.
`newdata`	A data frame or data table containing the values of the model covariates at which predictions are required.
`quantiles`	The probability levels at which quantile predictions should be produced.
`model_name`	The name of the model in `object` to be used for prediction. E.g. Specific cross-vlaidation fold, test model or some other version.
`sort`	`boolean` Sort quantiles using `SortQuantiles()`?
`sort_limits`	`Limits` argument to be passed to `SortQuantiles()`. Constrains quantiles to upper and lower limits given by `list(U=upperlim,L=lowerlim)`.

Details

Predict method for multiple quantile regression models of the class qreg_gam.

Value

A list with elements gam_pred, deterministic predictions (conditional expectation) from main GAM model, and mqr_pred, multiple predictive quantiles in a MultiQR object.