splitSampleTest: Conduct a differential split sample test on a regression...

In markwh/rcmodel: Calibrate and Validate Rating Curves for Riverine Load Estimation

Description

Returns prediction errors from a model calibrated for condition == FALSE and validated for condition == TRUE. Based on differential split sample test described in Klemes (1986)

Usage

splitSampleTest(object, condition, what = c("concentration", "load"),
  retransform = TRUE, scale = c("none", "gcv", "cv"), kfolds = 10,
  incl.data = FALSE, ...)

Arguments

object	an rcgam object
condition	a condition that returns logical when given a row of the default newdata for predict(object)
what	What to obtain errors for (concentration or load)
retransform	Perform retransformation before calculating errors?
scale	How to scale the residuals? (For comparing multiple models for different datasets) Defaults to "none".
kfolds	For scale == "cv" only. How many folds to use for crossvalidation?
incl.data	Return data and scale along with residuals? If so, a list is returned.
...	passed to predict.rcgam

markwh/rcmodel documentation built on May 21, 2019, 12:26 p.m.