condlSample: Generate samples from conditional distribution.

In markwh/condSample: Generate random samples from the assumed error distribution of a regression model, conditional on predictor variable values.

Description

Generate samples from conditional distribution.

Generate samples from a conditional distribution obtained from a model prediction

Generate samples from a conditional distribution obtained from a model prediction

Usage

condlSample(object, ...)

## S3 method for class 'rcgam'
condlSample(object, newdata, flowcol = "flow",
  flow.units = "ft3/s", quantile, ...)

## S3 method for class 'rclm'
condlSample(object, newdata, flowcol = "flow",
  flow.units = "ft3/s", quantile, ...)

Arguments

object	a model object with a 'predict()' method
...	arguments to be passed to individual methods. These should always include : 'newdata' - to be passed to 'predict()' function ‘quantile' - either ’random' for stochastic sampling from conditional distribution or a numeric value on (0, 1) specifying the quantile to return
newdata	a data.frame containing precictor variables to use for prediction
object	An rclm or rcgam object to use for predicting
retransform	Should the predictions be returned as concentrations? (defaults to TRUE)
...	Arguments passed to 'predict.lm' or predict.gam function call
smear	Use Smearing estimator to correct transformation bias?
object	An rclm or rcgam object to use for predicting
newdata	a data.frame containing precictor variables to use for prediction
retransform	Should the predictions be returned as concentrations? (defaults to TRUE)
...	Arguments passed to 'predict.lm' or predict.gam function call
smear	Use Smearing estimator to correct transformation bias?

Value

Numeric vector containing conditional random sample (if ‘quantile = ’random'') or conditional quantiles from condition distribution defined by 'object' and 'newdata' seealso condlSample.lm

Methods (by class)

• rcgam: Method for rcgam objects

• rclm: Method for rclm objects

markwh/condSample documentation built on May 21, 2019, 12:25 p.m.